CD ROM Paradise Collection 4

home *** CD-ROM | disk | FTP | other *** search

/ CD ROM Paradise Collection 4 / CD ROM Paradise Collection 4 1995 Nov.iso / database / bltp18.zip / CZ.HLP (.txt) < prev next >

Wrap

CZ Help | 1994-10-10 | 215KB | 3,815 lines

CZ_HELP! INDEX TUTORIAL_INDEX LICENSE_AGREEMEN LICENSE_A PRODUCT_SUPPORT STARTING_CZ USING_CZ USING_CZ_A USING_CZ_B ABOUT_CZ IS_BULLET IS_A_DATABASE IS_DBF IS_A_BTREE IS_A_NETWORK IS_FILE_LOCKING IS_NLS DESIGN_A_DB CREATE_A_DB ADD_TO_THE_DB QUERY_THE_DB UPDATE_THE_DB DELETE_A_RECORD PB_COMPILE_WITH CALL_BULLET SPECS_OVERALL SPECS_DBF SPECS_DBF_A SPECS_DBF_B SPECS_DBF_C SPECS_INDEX SPECS_INDEX_A SPECS_INDEX_B SPECS_INDEX_C SPECS_MEMORY SPECS_MEMORY_A SPECS_OS_CALLS SPECS_LANGUAGES SPECS_OSES SPECS_NETWORKS SPECS_PERFORMANC SPECS_PERF_A SPECS_PERF_B SPECS_PERF_C SPECS_PERF_D INITXB EXITXB ATEXITXB MEMORYXB BREAKXB BACKUPFILEXB STATHANDLEXB GETEXTERRORXB DVMONCXB CREATEDXB OPENDXB CLOSEDXB STATDXB READDHXB FLUSHDHXB COPYDHXB ZAPDHXB CREATEKXB CREATEKXB_A CREATEKXB_B CREATEKXB_C CREATEKXB_D CREATEKXB_E CREATEKXB_F OPENKXB CLOSEKXB STATKXB READKHXB FLUSHKHXB COPYKHXB ZAPKHXB GETDESCRIPTORXB GETRECORDXB ADDRECORDXB UPDATERECORDXB DELETERECORDXB UNDELETERECORDXBwc PACKRECORDSXB FIRSTKEYXB EQUALKEYXB NEXTKEYXB PREVKEYXB LASTKEYXB STOREKEYXB DELETEKEYXB BUILDKEYXB CURRENTKEYXB GETFIRSTXB GETEQUALXB GETNEXTXB GETPREVXB GETLASTXB INSERTXB UPDATEXB REINDEXXB LOCKXB UNLOCKXB LOCKKEYXB UNLOCKKEYXB LOCKDATAXB UNLOCKDATAXB DRIVEREMOTEXB FILEREMOTEXB SETRETRIESXB DELETEFILEDOS RENAMEFILEDOS CREATEFILEDOS ACCESSFILEDOS OPENFILEDOS SEEKFILEDOS READFILEDOS EXPANDFILEDOS WRITEFILEDOS CLOSEFILEDOS MAKEDIRDOS ACCESSPACK BREAKPACK COPYPACK CREATEDATAPACK CREATEKEYPACK DESCRIPTORPACK DOSFILEPACK DVMONPACK EXITPACK FIELDDESCTYPE HANDLEPACK INITPACK MEMORYPACK OPENPACK REMOTEPACK SETRETRIESPACK STATDATAPACK STATKEYPACK STATHANDLEPACK XERRORPACK ERRORS_BULLET ERRORS_BULLET_B ERRORS_BULLET_C ERRORS_BULLET_D ERRORS_BASIC ERRORS_BASIC_B ERRORS_DOS ERRORS_DOS_B ERRORS_DOS_C INITXBSRC EXITXBSRC ATEXITXBSRC MEMORYXBSRC BREAKXBSRC BACKUPFILEXBSRC STATHANDLEXBSRC GETEXTERRORXBSRC DVMONCXBSRC CREATEDXBSRC CREATEDXBSRC_A OPENDXBSRC CLOSEDXBSRC STATDXBSRC READDHXBSRC FLUSHDHXBSRC COPYDHXBSRC ZAPDHXBSRC CREATEKXBSRC CREATEKXBSRC_A OPENKXBSRC CLOSEKXBSRC STATKXBSRC READKHXBSRC FLUSHKHXBSRC COPYKHXBSRC ZAPKHXBSRC GETDESCRIPTORXBSz GETRECORDXBSRC ADDRECORDXBSRC UPDATERECORDXBSR DELETERECORDXBSR UNDELETERECORDSR_ PACKRECORDSXBSRC1 FIRSTKEYXBSRC EQUALKEYXBSRC NEXTKEYXBSRC PREVKEYXBSRC LASTKEYXBSRC STOREKEYXBSRC DELETEKEYXBSRC BUILDKEYXBSRC CURRENTKEYXBSRC GETFIRSTXBSRC GETEQUALXBSRC GETNEXTXBSRC GETPREVXBSRC GETLASTXBSRC INSERTXBSRC UPDATEXBSRC REINDEXXBSRC LOCKXBSRC UNLOCKXBSRC LOCKKEYXBSRC UNLOCKKEYXBSRC LOCKDATAXBSRC UNLOCKDATAXBSRC DRIVEREMOTEXBSRC FILEREMOTEXBSRC SETRETRIESXBSRC DELETEFILEDOSSRC RENAMEFILEDOSSRC CREATEFILEDOSSRC ACCESSFILEDOSSRC OPENFILEDOSSRC SEEKFILEDOSSRC READFILEDOSSRC EXPANDFILEDOSSRC^, WRITEFILEDOSSRC CLOSEFILEDOSSRC MAKEDIRDOSSRC ~INDEX CZ.HLP-BULLET for PowerBasic 3 System Mid-level Record/Key Access InitXB CreateDXB CreateKXB GetDescriptorXB FirstKeyXB ExitXB OpenDXB OpenKXB GetRecordXB EqualKeyXB AtExitXB CloseDXB CloseKXB AddRecordXB NextKeyXB MemoryXB StatDXB StatKXB UpdateRecordXB PrevKeyXB BreakXB ReadDHXB ReadKHXB DeleteRecordXB LastKeyXB BackupFileXB FlushDHXB FlushKHXB UndeleteRecordXB StoreKeyXB StatHandleXB CopyDHXB CopyKHXB PackRecordsXB DeleteKeyXB GetExtErrorXB ZapDHXB ZapKHXB BuildKeyXB DVmonCXB CurrentKeyXB High-level Access Network GetFirstXB InsertXB LockXB UnlockXB LockKeyXB GetEqualXB UpdateXB UnlockKeyXB LockDataXB UnlockDataXB GetNextXB ReindexXB DriveRemoteXB FileRemoteXB SetRetriesXB GetPrevXB GetLastXB Low-level DOS Access DeleteFileDOS OpenFileDOS WriteFileDOS RenameFileDOS SeekFileDOS CloseFileDOS Move cursor to index item CreateFileDOS ReadFileDOS MakeDirDOS and press <Enter>. AccessFileDOS ExpandFileDOS See: TUTORIAL_INDEX ~TUTORIAL_INDEX CZ.HLP-BULLET for PowerBasic 3 CZ.COM Using BULLET 1.08 Starting_CZ What: How to: Using_CZ is_BULLET design_a_DB PB_compile_with About_CZ is_a_database create_a_DB is_DBF add_to_the_DB is_a_Btree query_the_DB Error Codes is_a_network update_the_DB Errors_BULLET is_file_locking delete_a_record call_BULLET Errors_BASIC is_NLS Errors_DOS LICENSE_AGREEMENT Product_Support Structure Pack Types Specifications AccessPack DVmonPack RemotePack Specs_Overall BreakPack ExitPack SetRetriesPack Specs_DBF CopyPack FieldDescTYPE StatDataPack Specs_Index CreateDataPack HandlePack StatKeyPack Specs_Memory CreateKeyPack InitPack StatHandlePack Specs_OS_Calls DescriptorPack MemoryPack XErrorPack Specs_Languages DOSFilePack OpenPack Specs_OSes Specs_Networks Specs_Performance See: License_Agreement ~License_Agreement Before using this software you must agree to the following: 1. You are not allowed to operate more than one (1) copy of this software package at one time per license. This means that if you have 10 programmers that COULD possibly use the BULLET library at the same time, you must also have ten (10) BULLET licenses. 2. You are not allowed to distribute non-executable code containing BULLET code. This means that you are not allowed to redistribute BULLET code as another .LIB, for example. Also, if BULLET code is to be contained in a Dynamic Link Library (DLL) then it must be part of a stand-alone product. This means that you cannot provide a .DLL containing BULLET code if that .DLL is to be used as a programming library for other programmers. If you wish to distribute non-executable code containing BULLET code you must obtain written permission from the author. 3. This license grants you the right to use the BULLET library code on a royalty-free basis. See: License_a -MORE- ~License_a 4. BULLET is owned by the author, Cornel Huth, and is protected by United States copyright laws and international treaty provisions. You are not allowed to make copies of this software except for archival purposes. 5. You may not rent or lease BULLET. You may not transfer this license without the written permission of the author. If this software is an update or upgrade, you may not sell or give away previous versions. 6. You may not reverse engineer, decompile, or disassemble this software. 7. There are no expressed or implied warranties with this software. 8. All liabilities in the use of this software rest with the user. 9. U.S. Government Restricted Rights. This software is provided with restricted rights. Use, duplication, or disclosure by the Government is subject to restrictions as set forth in subparagraph (c)(1)(ii) of the Rights in Technical Data and Computer Software clause at 52.227-7013. Manufacturer is Cornel Huth/6402 Ingram Rd/San Antonio, TX 78238. This agreement is governed by the laws of the state of Texas. See: Product_Support ~Product_Support Support is available at my BBS, The 40th Floor, 7 days a week. Hours are 5pm to 9am, Central Time (USA). Weekend BBS hours are 24hrs, starting at 5pm Friday to 9am Monday. Hours other than during above periods are voice. Central Time is UTC time -5/6 hours, depending on season (UTC=London time). BBS tele#: 1(210)684-8065 (times listed above) N81, 300 to 14.4k bps Latest releases of BULLET are available for free download by registered users. Also available are other shareware products by me such as LP100, a linear programming optimizer, and Ruckus, a super-audio programmer's toolkit. My E-mail address: Fidonet: 1:387/800.8 (okay) Internet: cornel@ephsa.sa.tx.us (near term) cornel@LChance.sa.tx.us (okay) cornel@crl.com (tentative) Call the BBS for the latest info and e-mail address, or write. See: Starting_CZ ~Starting_CZ At CZ's initial load it looks into the current directory for CZ.HLP, then in the directory of CZ.COM, and last it looks for the pathname specified by the DOS variable CZH (SET CZH=C:\DOC\CZ.HLP). Use /f: for alternate locations or if CZ has any trouble locating its CZ.HLP file (C>cz /f:D:\BIN\CZ.HLP). Load CZ.COM from the DOS command line. Options are: /f:helpfile.ext Use other than default CZ.HLP help file /h# where n=1 to 4 Use alternate hot-key from default Alt-F1 #=1 Ctrl-h #=2 F12 #=3 left+right Shift #=4 Alt-left Shift /u uninstall CZ from memory (after initial load) /s temporarily put CZ to sleep by restoring all hooked vectors /r restore CZ from its sleep by rehooking vectors /? quick help E.g., C>cz /f:D:\PRG_C\CBULLET.HLP supply entire pathname when using /f: C>cz /h1 change hot key from Alt-F1 to Ctrl-H See: Using_CZ ~Using_CZ To activate CZ press the hot-key while the cursor is on the word you want to look up. Information on that word, if any, is displayed. If none is available, an index of help items in the dictionary is shown. Along the top-right of the index screens is the control bar. You can quickly move to the control bar by pressing <Home> or the See: line with <End>. Move the cursor to TUTORIAL_INDEX to select the second index screen or QUIT to return to whatever you were doing or over any index item. Then press <Enter> to move to that item's help entry. <F1> can be used as <Enter>. A mouse can also be used and is recommended. For example, to find out about CreateDXB, type it in your application and move the cursor on it. Press the hot-key. The CreateDXB screen is displayed. To see what pack type it uses, move to CreateDataPack (at Pack:) and press <Enter>. For a source example, move the cursor to Src: CreateDXBsrc. To go straight to the index from your application, press the hot-key with the cursor on a blank space. The <Esc> key returns you to your application. If there are more screens for the current topic the See: line has the same topic name plus a letter, and -MORE- at the end. Move the cursor (or mouse) to the topicname text (by using the <End> key) and press <Enter> (or click). See: Using_CZ_a -MORE- ~Using_CZ_a CZ.COM can be loaded high but it is ESSENTIAL that you have at least 15K of free UMB RAM available. It will load in as little as 4.5K but it will not operate correctly. Use MEM/c to see how much Upper Memory RAM is available. CZ opens the help file at installation. The help file is opened for Read-Only access with a Deny None sharing attribute. The file is closed when CZ is uninstalled (C>cz /u). CZ makes use of its own 256-byte stack. If you have several CZ help files, rename them to their particular application. For example: Rename the QuickBASIC BULLET CZ.HLP to QBULLET.HLP. Put QBULLET.HLP into your help files directory. Put SET CZH=C:\HELPFILES\QBULLET.HLP in your AUTOEXEC.BAT file. The next time CZ is installed it uses QBULLET.HLP from C:\HELPFILES. At anytime you can specify CZ.COM to use another help file. For example, if the current CZ help file is QBULLET.HLP but you want to use the CBULLET.HLP file, use C>cz /f:\helpfiles\cbullet.hlp. CBULLET.HLP then becomes the active file the next time you popup CZ. See: Using_CZ_b -MORE- ~Using_CZ_b Limitations (most apply to QB/QBX): 1) The QB environment may interpret the Alt-F1 keypress after CZ has popped down. If this is a problem then change the hotkey, e.g., C>cz /h2 to use F12. 2) In the QB (and QBX) environment the keypad <Enter> key is not recognized by CZ. Use the main <Enter>. This occurs only in then QB 4.5 and QBX editors. 3) If, after returning from CZ, the QB environment's menu bar does not respond to Alt-keys (like an Alt-F), click the left mouse button a few times, or press F6 to go to the immediate window and execute a SHELL. It's unlikely that you'll encounter this. 4) CZ is a stable TSR but like all TSRs in a DOS system, unforseen events can take place that could conceivably cause the computer to crash. Therefore, it's recommended that you save your work often (you should do so whether a TSR is installed or not). 5) CZ currently doesn't reset the mouse. If you're having mouse trouble, you'll need to reset your mouse driver (type MOUSE at the C>). Many problems come from other programs not restoring the mouse cursor to its original state. See: About_CZ ~About_CZ CZ HELP Context-sensitive Online Help Manager for the compiler libraries MS-DOS PowerBasic version (available for most DOS compilers) Copyright 1992-94 Cornel Huth Ver 1.08 INT2F MuxID=C2 10-Oct-94 See: Specs_Overall ~is_BULLET - What? BULLET is a program module that handles the details of putting information to and getting information from your hard disk using a standard data file format called the Xbase DBF format with very fast and efficient index file routines. It can be used as-is by most DOS compilers. BULLET is written in 100% assembly language. Why? Two reasons. First, control. There's no compiler code or run-time library code in between BULLET and your data. Second, efficiency. BULLET knows exactly what it requires from the operating system and when. Result: fast, small, and robust applications. See: is_a_database ~is_a_database - What? A database is a collection of data arranged so that the it can be accessed as useful information. For example, let's say we have two files. Each consists of two fields. The first file has codenumber and score. The second file has a codenumber and name. Separately, the files are merely a collection of data. Together, however, they tie the name to the score: score codenumber codenumber name 99 100 100 John 87 155 105 Paul 66 125 110 George : : : : Codenumber 100 is John, who scored 99. The other members scores are not in the abbreviated data file listing. A database can be a single data file but more often it is a group of related data files and usually these data files are indexed by keys (in the index file, also called key file) so that very fast, direct access is possible. See: is_DBF ~is_DBF - What? DBF is the file extension of dBASE III-compatible data files (filename.DBF). The file format is used by dBASE IV, FoxPro and many other database programs. Many programs can also use the file format to import/export data using it. The DBF format is the most common data file format used on PCs. A DBF-compatible data file consists 3 distinct areas. First is the data header. This contains information such as the number of records in the file. Second is the field descriptors. These descriptors define the makeup of each field in the record. The third is the record area. Each record is a logical unit of data. For example, a record, all of which are made up of the same fields but with different data, could (conceptually) look like this: field 1 field 2 field 3 field n record 1 Johnson Larry 465310555 ... record 2 Aberdeen Zara 465230555 ... record n See: is_a_Btree Specs_DBF ~is_a_Btree - What? A b-tree is a sorting method ideally suited to data structures maintained on a hard disk. It is very fast on retrieval and is inherently self-balancing during inserts and deletes. Self-balancing ensures performance remains consistent. The reason the b-tree is ideally suited to hard disks is that, when looking for a particular key, most of the time involved in accessing the key is spent by the hard drive moving to various locations on the disk. The task of a good access method is to reduce the number of seeks that the disk must perform. The b-tree accomplishes this by maintaining several keys (perhaps 50) on each node, with the necessary pointers to previous and following nodes. A b-tree of order 20 (19 keys per node) can find a key in a file of 1,000,000 keys in a MAXIMUM of 5 disk accesses, where each disk access visits a node. 'BASIC program to find *max* seeks needed/avg time Keys& = 1000000: KeysPerNode = 19: AvgSR = 25 Order = KeysPerNode + 1 max = (LOG((Keys& + 1) / 2) / LOG(Order / 2)) PRINT "Max nodes accessed for"; Keys; "keys & b-tree of order"; Order; PRINT "is"; max; "nodes" PRINT "Max disk time based on avg seek+read of"; AvgSR; PRINT "ms is"; AvgSR / 1000 * max; "seconds" See: is_a_network Specs_Index ~is_a_network - What? A network is a group of computers able to communicate with one another. Often called a LAN (local area network), a network allows resources to be shared. Sharing resources can lead to problems if steps are not taken to ensure that two computers don't try to share the same resource at the same time. For example, say two computers try to change the same record in a file on a network drive. Let's say both users are accessing the number of widgets in inventory. The first user gets there a micro-second before the second and allocates the last widget in stock. The second user comes in right after and, since the first user has not yet updated the inventory, allocates the very same widget. One widget, two users. When the first user updates the inventory, widgets in inventory is changed to 0 (previous - 1). The second updates the inventory in the same manner and sets widgets to 1 less what it was when it started, or 0 also. You see the problem. In order to successfully share a file on a network, the file must first be locked to a single user. Once that user has locked the file, he has sole access to the data within it and he will not experience the scenario above. When the user has completed the changes, he unlocks the file so that others may use it. See: is_file_locking ~is_file_locking - What? File locking is a means to obtain exclusive access to a file. This is needed in cases of multiple programs or users accessing a shared file at the same time. There are several methods to ensure only one process or user has access to a file. The first method is to open the file so that while the file is open only your program can access any part of it. This is simple to implement and the operating system handles the details of this. However, it requires your program to open/close files all the time since no other process may access the file while it is open. Another method is to use byte-level locks. Also managed by the OS, this method allows for restricting access to any particular region within the file. Which regions are to be locked is to be determined by your program, however, and it can be complex to perform multiple locks at the byte, field, or record level. Another use of the byte-level lock is to specify that all bytes within the file are to be locked. This greatly simplifies the process of obtaining a lock and has the advantage over a file access lock of not needing to open/close the file for each lock. It is very fast and easy to implement. BULLET offers all three lock types. See: is_NLS ~is_NLS - What? NLS stands for National Language Support. This feature is available in DOS 3.3 and later. BULLET makes use of NLS by getting from DOS the current DOS country collate-sequence table. The collate table is used to properly sort mixed-case character strings and also foreign (or non-USA) language character strings according to that country's alphabet. This is an option but is recommended. In addition, BULLET provides for a programmer-supplied collate-sequence table. See: design_a_DB ~design_a_DB - How to To design a database, above all else, know what information you require from it. Having established what you need to know, collect the data that lets you formulate this into useful information. For example, you want to track a class of students and determine how well they achieve on tests. The criterion you use is the test score. You determine that your data is 1) students, 2) tests, and 3) test scores. Too simplify, you use a single 20-character field for student, a 1-character field for test number ("1" to the "n" tests), and a numeric field for test scores (0 to 100). Since the objective is to track students' scores, arrange the data so that output consists of each student's score in test order. Do this by specifying an index file containing an index based on the student's name and test number: KeyExpression$ = "STUDENT + TEST" 'combine two character fields This is the reason that field 2, test number, is a character field. It can more easily be combined with other character fields than a numeric field. By using the routines of the database langauge, you can easily create the data and index files, add data, list student's scores, or make changes to the database. Note: these How_to examples are meant only to show the basis behind an operation. See: create_a_DB CreateKXB ~create_a_DB - How to Having defined the database, create it. First, create the datafile based on the 3 fields you defined in your design. To do this, DIM an array for the field descriptors for the number of fields (see also CreateDataPack): DIM FD(1 TO 3) AS FieldDescTYPE FD(1).FieldName = "STUDENT" + STRING$(10,0) 'must be zero-filled FD(1).FieldType = "C" FD(1).FieldLength = CHR$(20) FD(1).FieldDC = CHR$(0) The FD() is a structure element FD(2).FieldName = "TEST" + STRING$(10,0) in CDP (CreateDataPack) as in: FD(2).FieldType = "C" CDP.Func=CREATEDXB FD(2).FieldLength = CHR$(1) : : FD(2).FieldDC = CHR$(0) CDP.FieldListPtrOff=VARPTR(FD(1)) FD(3).FieldName = "SCORE" + STRING$(10,0) CDP.FieldListPtrSeg=VARSEG(FD(1)) FD(3).FieldType = "N" : : FD(3).FieldLength = CHR$(3) FD(3).FieldDC = CHR$(0) Call CreateDXB to create the data file. To create the index file, first open the data file just created, then call CreateKXB to create the index file. Open the index file so we can use it (data file is already open). See: add_to_the_DB CreateDXB ~add_to_the_DB - How to Once you have the database designed and the data and key files created and open you can start putting the student's test data into it. Note that the DBF-format requires that all data in a data file be in ASCII format. This means that we must convert the numeric test score into its ASCII form. BASIC has the STR$() function to do this. In addition, numbers generally should be right-justified in their field. BASIC has the RSET statement to do this: TYPE StudentRecordTYPE 'this structure is exactly how the record is on disk tag AS STRING * 1 'DBF delete tag used to identify deleted records sname AS STRING * 20 '--the tag is always present in any DBF record testn AS STRING * 1 ' don't forget to include it in all your TYPEs score AS STRING * 3 END TYPE '25 DIM SR AS StudentRecordTYPE : INPUT "Student name, test number, score:",n$,t$,s% SR.sname = n$ : SR.testn = t$ : RSET SR.score = STR$(s%) status = DoInsert(SR, keyhandle) LOOP UNTIL LEN(n$) = 0 OR (status <> 0) See: query_the_DB InsertXB ~query_the_DB - How to Now that you have data in the database you want to see what's in there. Since the index file is in "STUDENT + TEST" order, the information we'll be getting out of the database is in Student name order, with each student's scores in test number order. If we want to look at all the students, we can use GetFirstXB to retrieve the first student's score for the first test. GetNextXB retrieves the next record (the first student's score for the second test), and so on. When all records have been retrieve GetNextXB returns an End Of File error code. If we want to look at a particular student's score only, we can use GetEqualXB to go directly to a student's first test score. GetNextXB get his next and so on until GetNextXB retrieves the next student's first test score. You can stop at this point (student names no longer match). We might also want to find all students who scored less than 65 on any test. To do this we can GetFirstXB, check SR.score for < 65 and if so print that record. Continue by using GetNextXB, printing each record that has a score < 65. See: update_the_DB GetFirstXB ~update_the_DB - How to To update a particular record in the database we must first locate and identify it using one of the get routines such as GetEqualXB. The Get() routine return the record data, and also the physical record number of the record accessed, into the AccessPack RecNo. Having used one of the Get() routines to read the data record from disk to memory, you can make any changes to the data record in memory. E.g., if a student's score needs to be changed from a 69 to a 96, first find the record (and its RecNo), then update the score field: INPUT "Student name, test number",n$,t$ SR.sname = n$ : SR.testn = t$ status = DoGetStudentTestScore(SR, keyhandle, Recno&) DO UNTIL status <> 200 'student as entered not found, call search routine status = DoFindStudentTestScore(SR, keyhandle, RecNo&) LOOP IF status = 0 THEN PRINT "Student:" ; SR.sname ; " old score:" ; SR.score INPUT "Enter new score:"; s% : RSET SR.score = STR$(val(s%)) status = DoUpdateTestScore(SR, keyhandle, Recno&) ENDIF Any change to a key field will initiate a key file update automatically. See: delete_a_record UpdateXB ~delete_a_record - How to To delete a particular record in the database we must first locate it using one of the get routines such as GetEqualXB. These Get() routines return the actual record number of the data record accessed by Get() into the AccessPack RecNo. Having used one of the Get() routines to find the data record, make a call to the delete function: 'delete all of Student's records, 1 record for each test he's taken INPUT "Student name to delete",n$ SR.sname = n$ status = DoGetStudent(SR, keyhandle, RecNo&) DO UNTIL status <> 200 'student as entered not found, call search routine status = DoFindStudent(SR, keyhandle,RecNo&) LOOP DO status = DoDeleteRecord(SR, RecNo&) 'does not affect key files IF status = 0 THEN status = GetNextChkIfSameStudent(SR,keyhandle,RecNo&) LOOP UNTIL status <> 0 The DeleteRecordXB routine does not physically remove the record from the data file but instead tags it as being "deleted". See: PB_compile_with DeleteRecordXB ~PB_compile_with - How to To create a stand-alone EXE file, compile your BASIC source as required. No special compiler switches are required. Add the following to the top of your source files: DEFINT A-Z $LINK "PBULLET.PBL" $INCLUDE "PBULLET.BI" $LINK "NOATEXIT.OBJ" See: call_BULLET AtExitXB ~call_BULLET - How to? BULLET is called through a single entry point. The only argument passed to it is a segmented far pointer to the control pack. The first two entries in this pack are the function to be performed and the function return status. BULLET is a FUNCTION call returning an INTEGER status value. Each function (or routine) uses a prescribed pack format. For example, some routines need only know the handle of the file, along with the function number itself. So, to flush a data file, for example, you would do the following: DIM HP AS HandlePack 'could also DIM SHARED HP.Func = %FLUSHDHXB 'FLUSHDHXB is defined as a CONST in PBULLET.BI HP.Handle = File2FlushHandle 'the handle as returned from the Open() routine stat = BULLET(HP) 'do the actual call to BULLET The value of stat is set to the completion code as returned by the FlushDHXB routine. It is the same as the value returned in HP.Stat *IN ALL BUT A FEW* cases: InsertXB, UpdateXB, ReindexXB, and LockXB. These routines return not the actual error code, but rather a transaction index number of the access that failed. See those routines for more information. See: is_BULLET FlushDHXB ~Specs_Overall BULLET is dBASE III/III+/IV .DBF-compatible. This format is compatible with a large base of software programs including the latest database packages such as dBASE IV and FoxPro. Spreadsheet packages such as Excel and 1-2-3 can directly import BULLET DBF data files, too. And because of BULLET's versatility, it can also create very non-standard data files. This may be a useful feature if data secrecy is of concern. BULLET requires MS-DOS 3.30 or above. It uses 19K of code/static data space and requires at least 40K of workspace. 140K of workspace is ideal. Overall Specifications: DBF (per file) INDEX Max records: 16,777,215 Max nodes: 65,535 Record length: 2-4000 (8192) Max keys: 4,063,170 Max fields: 128 (255) Key length: 1-64 Field length: 1-254 (255) Max key fields: 16 Total open index plus data files can be up to 255. Numbers in () indicate extended specifications. See: Specs_DBF ~Specs_DBF To remain compatible with other dBASE III .DBF platforms you should restrict your data files to the following specifications: File ID byte: 3 (83hex if .DBF has memo field, not currently supported) Max record size: 4000 bytes Max fields/rec: 128 Max field size: 254 bytes Allowable field name characters: A-Z and the _ (upper-case) Allowable field types: C-character, 1-254 bytes D-date, 8 bytes, in the format YYYYMMDD (19920531) L-logical, 1 byte, either space, "T" or "Y", "F" or "N" M-memo, 10 bytes, used as pointer into .DBT file (currently not supported) N-numeric, 1-19 bytes, ASCII format, uses explicit decimal if needed... ...decimal places may be 0, or 2 to (field size - 3) but no more than 15 Restrict all data in .DBF fields to ASCII. This means you should convert binary data to the equivalent ASCII representation, e.g., if you have the binary value 22154, it must first be converted to the string "22154" before you can store it to the .DBF data file. So, while your in-program code deals with binary data, your I/O code must convert it to/from ASCII. This is a dBASE-compatibility issue only. If you can forgo these requirements you can use binary fields, any- character field names, record sizes to 8192 bytes, and up to 255 fields. See: Specs_DBF_a -MORE- ~Specs_DBF_a A dBASE III .DBF is composed of 3 sections: the header, the field descriptors, and the data area. The header structure (first 32 bytes of file): Name Type Offset Meaning FileID byte 0 data file type id, 03 standard (43,63,83,88h) LastYR byte 1 last update year, binary LastMo byte 2 last update month, binary LastDA byte 3 last update day, binary NoRecs long 4 number of records in file HdrLen word 8 length of header, including field descriptors, +1 RecLen word 10 length of data record including delete tag internal byte 12-31 reserved The last update values are updated to the current date whenever the .DBF file is flushed or closed. Likewise, the NoRecs value is updated whenever a record is added to the .DBF. The FileID is specified when you create the file, HdrLen and RecLen are computed and stored when the file is created, too. See: Specs_DBF_b -MORE- ~Specs_DBF_b The field descriptor format (follows header, one per field): Name Type Offset Meaning FieldName char 0 field name 10 ASCII characters, A-Z or _ (0-filled) 0T byte 10 field name zero-termintor (must be 0) FieldType char 11 field type (C D L M N) internal long 12 reserved FieldLen byte 16 length of this field FieldDC byte 17 decimal count internal byte 18-31 reserved The unused bytes in the FieldName must be set to zeroes (CHR$(0)). Each field is described by a 32-byte descriptor. The first field's descriptor starts right after the header proper, at offset +32. After the last field descriptor is data byte ASCII 13. (Note: the orginal dBASE III has a 0 byte following this ASCII 13.) Immediately following this is the actual record data. See: Specs_DBF_c -MORE- ~Specs_DBF_c The data record format: The first record is located at offset HdrLen (from the header). The first byte of each record is a delete tag. This tag is maintained by the BULLET routines. A space, ASCII 32, means the record is not deleted; an asterisk, ASCII 42, means the record has been deleted (marked as deleted, often this is used as a method to temporarily tag records, for whatever purpose). Following the tag is the data for each field, not delimited (i.e., the fields run together without anything separating them). The second record is at offset HdrLen+reclen. The start offset of any record in the file can be computed as (recordnumber - 1) * reclen + HdrLen. All data is in ASCII form. An EOF marker (ASCII 26) is placed at the end of the last record. See: Specs_Index ~Specs_Index BULLET uses a proprietary, modified b-tree index method to manage the index files. The supported key types are: Type Length Meaning Character 1-64 ASCII, NLS, or user-supplied sort table Integer 2 signed or unsigned 16-bit value Long Int 4 signed or unsigned 32-bit value In addition to the above types, BULLET allows for unique or duplicate keys in the index file. If duplicates are allowed, BULLET enumerates each key with an enumerator word (see FirstKeyXB). The key may be composed of up to 16 character fields or substrings within those fields. Numeric fields are considered character fields by BULLET unless the key is set to binary (see KeyFlags). Integer or LongInt binary keys can be composed of a single field only. The key expression is specified in text (e.g., "LNAME+ SUBSTR(FNAME,1,1)+MI") and is fully evaluated when the index file is created. A BULLET index file is composed of 3 sections: the header, the collate-sequence table, and the node/key entry area. See: Specs_Index_a -MORE- ~Specs_Index_a The header structure: Name Type Offset Meaning FileID byte 0 index file type id, 20 RootNode word 1 root node number Keys 24bit 3 number of keys in index file AvalNode word 6 node number available for reuse FreeNode word 8 next free node number KeyLen byte 10 key length NodeKeys byte 11 number of keys that fit on a node CodePage word 12 code page ID CtryCode word 14 country code internal byte 16-21 reserved KeyFlags word 22 key flags KeyExprn byte 24-159 key expression internal byte 160 reserved KeyXFlds byte 161 number of fields used by key (1-16) KeyXlate byte 162-225 translated key expression internal byte 226-253 reserved CTsize word 254 collate-sequence table size See: Specs_Index_b -MORE- ~Specs_Index_b The collate-sequence table structure: table byte 256-511 sort weight table of ASCII character 0-255 Node/key entry structure (first entry is in node #1, file offset 512): 2A 0A 00 KEY123 7B 00 00 12 00 KEY178 B2 00 00 0C 00 ... 1. 2. 3. 4. 5. 6. 7. 8. 9. 1. Key count for that node (first byte of each node) 2. 16-bit node back pointer (for non-leaf nodes, 0 if leaf node) 3. First key value, "KEY123" in this case 4. 24-bit data record pointer (low word/hi byte) 7Bh = DBF record number 123 5. 16-bit node forward ptr/back ptr (for non-leaf nodes, 0 if leaf node) --in this case, it indicates that the key following KEY123 is in node# 12h --and also that the key before KEY178 is in that node as well 6. Second key (here "KEY178") 7. 24-bit data pointer (record number in DBF) 8. 16-bit forward node pointer (for non-leaf nodes, 0 if leaf node) 9. Repeat 6 to 8 for each key on node. (node size is 512 bytes) See: Specs_Index_c -MORE- ~Specs_Index_c As in many b-tree implementations, BULLET's index files maintain an average load percentage of approximately 66%. This means that in any given node, 66% of the available space is in use. The free space in the node is attributable to the constant reshaping of the file as keys are inserted or deleted, causing the nodes to be split and merged. A split will occur when an insert needs to add a key to an already full node; a merge will occur when a neighboring node is small enough to be merged into a just split node. This constant prune-and-graft of the b-tree results in a node load of about 66% (50% in degenerate cases such as with already sorted data). It's this aspect of the b-tree that makes it a consistent performer and a widely-used method of managing index files. The following formula can be used to determine the number of keys that an index file can hold: MaxKeys = MaxNodes * MaxKeysPerNode * LoadFactor MaxKeys = 65535 * 509/(keylen+5) * .66 The load factor can be increased to ~95% by using the ReindexXB routine. This load factor results in superior retrieval speeds since there are more keys on each node. Insertion speed will be decreased, however, since splitting will occur more frequently, though perhaps not noticeably. See: Specs_Memory ~Specs_Memory BULLET allocates memory on an as-needed basis. When linked to an executable program, BULLET makes use of 17.5K of code space and about 1.5K of static DGROUP data space. To accomodate the wide variety of compilers, BULLET's API structure will have the linker included all of the library into your final EXE program. All runtime memory allocations are obtained from the operating system. This means that if your compiler generates code to allocate all the free memory at startup for its far heap, you need to instruct it to free up some memory for BULLET's use. In QuickBASIC you can use SETMEM(). You can also use the linker /CP:n option to limit the amount that the compiler-generated startup code allocates. The EXEHDR.EXE program can also be used to modify this amount. The linker /CP: option and the EXEHDR program both modify the EXE header. To determine if you need to release memory use the MemoryXB routine. It reports the operating system memory available for use by BULLET. The amount of memory that BULLET requires is based on which routines are used. See the next screen for a list of the routines that make malloc calls to the operating system and how much memory they require. (Most compilers don't grab all memory as QuickBASIC when it's using $DYNAMIC). See: Specs_Memory_a MemoryXB -MORE- ~Specs_Memory_a Routines making dynamic memory allocations and amount (within 16 bytes): Routine Bytes Basis InitXB 272 permanent, released when program ends (JFTmode=1) BackupFileXB 32K temp, released when routine exits CreateDXB 48+(NF*32) temp, released when routine exits (NF=NoFields) CreateKXB 544 temp, released when routine exits OpenDXB 144+((1+NF)*32) semi-permanent, released when file closed OpenKXB 1264 semi-permanent, released when file closed PackRecordsXB RL to 64K temp, released when routine exits (RL=RecLength) ReindexXB 32K to 128K temp, released when routine exits UpdateXB 2K+RL temp, released when routine exits (RL=RecLength) For example, when BackupFileXB is called it attempts to allocate 32K from the OS. If 32K is not available, BackupFileXB returns with an error code of 8 (DOS error #8, not enough memory). If you won't be using Backup or Reindex, BULLET can make do with much less memory (use table above). Needed stack space is 4K (max) for ReindexXB. Other routines can operate with less than 1K of stack space. In other words, stack use is minimal. See: Specs_OS_calls ~Specs_OS_calls BULLET makes use of the following operating system calls: INT21/25 DOS_setvector INT21/44/0B DOS_setsharingretrycount INT21/2A DOS_getdate INT21/48 DOS_malloc INT21/30 DOS_version INT21/49 DOS_free INT21/35 DOS_getvector INT21/51 DOS_getpsp INT21/39 DOS_makedir INT21/56 DOS_renamefile INT21/3D DOS_openfile INT21/59 DOS_getextendederror INT21/3E DOS_closefile INT21/5A DOS_createtempfile INT21/3F DOS_readfile INT21/5B DOS_createnewfile INT21/40 DOS_writefile INT21/5C DOS_lockunlockfile INT21/41 DOS_deletefile INT21/65/01 DOS_getextendedcountryinfo INT21/42 DOS_movefileptr INT21/65/06 DOS_getcollatesequencetable INT21/44/09 DOS_isdriveremote INT21/67 DOS_sethandlecount INT21/44/0A DOS_isfileremote INT2F/10/00 DOS_isshareinstalled No other operating system calls are made. No BIOS calls are made. See: Specs_Languages ~Specs_Languages BULLET is compatible with most DOS compilers. The only requirements are that your compiler allow you to: 1. Call a library routine via a FAR call using PASCAL calling convention 2. Pass a far pointer (of the parameter pack) on the stack, by value 3. Supply far pointers to the various pack parameters 4. Be able to return an integer value from the FAR call (this is optional but recommended for the transaction-based routines) These requirements can be met with most BASIC, C, and other-language DOS compilers. CZ online help is currently available in BASIC and C versions. Others are pending. You should be able to do well with either of these versions using other-language compilers since the only difference is the source code examples. See: Specs_OSes ~Specs_OSes BULLET is currently available only for MS-DOS and compatible operating systems. It requires DOS 3.3 or higher. To provide efficient memory use, BULLET uses a single-buffer cache per index file. The single-buffer cache also provides for very quick network access since a minimum amount of memory needs to be flushed when releasing control of BULLET files. For maximum speed, however, an external high-performance disk cache can be used. Hyperdisk is a good choice (shareware, $50+). A properly configured cache can increase BULLET's performance from 10 to 300%. The most improvement is with the InsertXB routine. The least is with ReindexXB and PackRecordsXB, which do most of their work in temporarily allocated memory. Hyperdisk is about the best designed disk cache available for PCs. SmartDRV 4.x is also good. If you do not use a disk cache then it's recommended that you set your BUFFERS= statement in CONFIG.SYS to at least 20 or 30. Even without a disk cache, BULLET is still very fast. Also, be sure to set your FILES= to the number of files that you'll be opening at any one time. If you set FILES=20 you can have BULLET open 14 files (CZ.COM uses 1 and DOS reserves 5 more). You can set FILES=255 allowing BULLET to open up to 249 files at one time. DO NOT set FILES= to a value greater than 255. See: Specs_Networks ~Specs_Networks BULLET currently operates on all DOS-compatible network platforms. Be sure to install SHARE.EXE (or compatible) on the server and, if you are mutlitasking, on your local machine. If you'll be opening many files you should extended the default SHARE file-sharing information space and the number of locks that can performed at one time. The DOS 5.0 default is /F:2048 and /L:20. This allocates 2K for file-sharing info space and allows 20 consecutive locks to be active. If the F: value is too low, error 5 (extended error 32) is returned on an open attempt. If you extend the JFT in InitXB and plan to use many files, say more than 50, be sure to extend /F: by 2K for every 50 additional files and set the /L: to the number of files you plan on having open. If L: is too low, error 1 (ext err 36) is returned on a lock attempt. As an example, if you'll be using 100 files, set FILES=106 in CONFIG.SYS, set SHARE /F:4096 /L:106, and IP.JFTmode=1 for InitXB. These values are a minimum. If you have more than one process active, you need to account for other apps. Note that Windows always returns a "SHARE is installed" using the DOS detection routines used by BULLET. To determine if SHARE is actually installed, attempt to perform a lock using one of the LockXB routines. An error code indicates that SHARE (or compatible) is not installed. See: Specs_Performance ~Specs_Performance Test: Reindex 1,000 to 1,000,000 records (BB_LAI10.BAS) DBF: extended DBF using binary sort field key: LONG+SIGNED+UNIQUE Machine: 486/33 SHO 1Meg * * Records Time Reindex Rate * ------- ---- ------------ 100k * 1000 < 1 1000+/sec 5000 2 2500 * 10000 4 2500 * 25000 7 3571 3500+ records indexed/second! 50000 14 3571 Times in table are in seconds 10k * 100000 28 3571 200000 81 2469 * 500000 355 1408 1000000 1124 890 1k time (secs) 100 200 300 400 18:00 20:00 (min) See: Specs_Perf_a -MORE- ~Specs_Perf_a Test: Add 1,000 to 1,000,000 records (BB_LAI10.BAS) DBF: extended DBF using binary sort field key: not indexed Machine: 486/33 SHO 1Meg * * * Records Time Add Rate ------- ---- ------------ 100k * 1000 < 1 1000+/sec 5000 2 2500 * 10000 4 2500 * 25000 12 2083 2000+ records added/second! 50000 24 2083 Times in table are in seconds 10k * 100000 50 2000 200000 103 1942 * 500000 262 1908 1000000 529 1890 1k time (secs) 100 200 300 400 8:00 10:00 (min) See: Specs_Perf_b -MORE- ~Specs_Perf_b Test: Access 1,000 to 1,000,000 keys(*), keys+records(+) (BB_LGK10.BAS) DBF: extended DBF using binary sort field key: LONG+SIGNED+UNIQUE Machine: 486/33 SHO 1Meg * + * + * + Keys Time Access Rate Keys+Recs Time Access Rate ------- ---- ----------- --------- ---- ------------ 100k * + 1000 < 1 1000+/sec 1000 < 1 1000+/sec 10000 1 10000 10000 6 1667 100000 22 4545 100000 68 1471 200000 49 4082 200000 144 1389 500000 147 3401 500000 427 1171 10k *+ 1000000 424 2358 1000000 948 1055 Times for forward/reverse (Next/Prev) were similar ( + Times in table are in seconds 1k time (secs) 100 200 300 400 15:00 17:00 (min) See: Specs_Perf_c -MORE- ~Specs_Perf_c Test: Insert 1,000 random records to existing data/key files (BB_LAA10.BAS) DBF: extended DBF using binary sort field key: LONG+SIGNED+UNIQUE Machine: 486/33 SHO Base records/keys Inserted Time / Rate ----------------- -------- ---- ---- 1000 1000 15 67 records+keys inserted/second 10000 1000 18 56 Times in table are in seconds 100000 1000 19 53 200000 1000 20 50 This tests the InsertXB speed. A data file containing the base records and an index file containing the same number of keys has inserted into it 1000 random keys and records. The records are added to the end of the data file and the keys are inserted into their appropriate place in the index file. Typically InsertXBs would be done online. The rate of 50 inserts/second at 200,000 records would indicate a time of 0.02 seconds per insert on even a very large database (0.015 secs/insert @ 1000 records). See: Specs_Perf_d -MORE- ~Specs_Perf_d Test: Update 1,000 records by changing the key field (BB_LAU10.BAS) DBF: extended DBF using binary sort field key: LONG+SIGNED+UNIQUE Machine: 486/33 SHO Base records/keys Updated Time / Rate ----------------- ------- ---- ---- 1000 1000 50 20 records+keys updated/second 10000 1000 59 17 Times in table are in seconds 100000 1000 108 9 200000 1000 126 8 This tests the UpdateXB speed. A data file containing the base records and an index file containing the same number of keys has the first 1000 keys and records changed. This forces the old data record to be replaced, the old index key to be deleted, and a new key for the record to be inserted in the index file. Typically UpdateXBs would be done online. The rate of 8 updates/second at 200,000 records would indicate a time of 0.125 seconds per update on even a very large database (0.05 secs/update @ 1000 records). See: Specs_Overall ~InitXB Pack: InitPack Src: InitXBsrc Func: 0/System Before using any routine you must initialize the BULLET file system. If you want more than the standard number of file handles, set InitPack.JFTmode to 1. This expands the current process's Job File Table to allow 255 open files maximum. On return the DOS version (INT21/30h) is in InitPack.DOSver. Major version in the high byte. Minor in the low. The BULLET version (*100) is returned as is the address of the ExitXB routine. You can use this address to register ExitXB with your own _atexit function if your runtime library does not provide _atexit already. Note: _atexit is a routine available in most DOS, OS/2, and ANSI runtime library code and is called just prior to the program ending. See AtExitXB for information on what to do if your library does not have _atexit. See: ExitXB ~ExitXB Pack: ExitPack Src: ExitXBsrc Func: 1/System Before ending your program you should call ExitXB to close any open BULLET files. This also will release any memory still allocated to those files. This restores the default keyboard break handlers if they were changed. In normal operation you would see to closing all files yourself. However, if your program fails to reach the programmed end, it's very possible that files may still be left open. It is essential that you properly close all BULLET files before ending. There are two methods to achieve this: 1. In BASIC use ON ERROR GOTO label, where label is in the main module. The code at label would then call ExitXB. 2. Use AtExitXB to automatically register ExitXB to be executed in the normal shut-down code of the compiler. This method is preferred. In the QB environment you can easily stop and restart the executing program without having reached the end of the program where you call ExitXB. Previously opened BULLET files remain open and memory used remains allocated. Eventually you will run out of files or out of memory. Use AtExitXB to prevent this. See: AtExitXB ~AtExitXB Pack: ExitPack Src: AtExitXBsrc Func: 2/System Used to automatically close all BULLET files, release allocated memory, and restore the default Break handlers when your program ends. Your compiler generates specific code to be executed in the course of ending your program. AtExitXB registers the ExitXB routine to be performed in this compiler- generated code. This routine is standard in most DOS, OS/2, and ANSI runtime libraries. If yours does not have _atexit, then you must link with the supplied NOATEXIT.OBJ file: C>link YOURPRG + NOATEXIT, ... You can tell if your compiler doesn't supply _atexit at link time. LINK reports '_atexit' : unresolved external. Add NOATEXIT.OBJ as described above. Be sure that your _atexit routine is for the medium, large, or huge memory models since BULLET uses multiple code segments and far calls. See: MemoryXB ExitXB BreakXB ~MemoryXB Pack: MemoryPack Src: MemoryXBsrc Func: 3/System This is the only BULLET routine that can be used before InitXB. It reports the largest free block of memory available from the OS. This memory does not include fragmented memory or UMB memory that BULLET can and will use. Some compilers provide their own memory management and thus will use most or all of the contiguous memory when its program loads. BULLET, however, allocates memory on an as-needed basis, directly from the OS, and releases it back to the OS as soon as it is no longer needed. If the memory amount returned by MemoryXB is low, your compiler's startup code has pre-allocated all OS memory. If the memory available from the OS is less than your application needs, you need to instruct your program to release some of the memory back to the OS so BULLET can use it. For QuickBASIC and BASIC PDS, see the BASIC function SETMEM(). With DOS able to use UMB memory, memory for BULLET requests may be provided from this region. You can use StatPack.HereSeg from StatXB to locate from which segment address the allocations are being made. Anything above C800h is UMB. See: BreakXB StatXB OpenDXB OpenKXB ~BreakXB Pack: BreakPack Src: BreakXBsrc Func: 4/System Disables system response to Control-C and Control-Break keys preventing users from inadvertently exiting the program without first doing a BULLET shutdown. It's REQUIRED that you reinstate the default break handlers with this routine before ending your program. ExitXB automatically reinstates the default break handlers. This routine will not disable Control-Alt-Delete (a warm-boot). If the user is at this point, he may prefer to exit via a warm-boot rather than reset the machine. Also, this routine does not prevent QuickBASIC and BASIC PDS from terminating a program at an INPUT statement. This is because these versions of BASIC are using their own method of INPUT aside from DOS. If AtExitXB has been issued, all BULLET files will be properly closed even if *User Break* occurs. This routine will not surpress the ^C displayed by DOS. If you don't want the ^C to be displayed move the cursor to a location off-screen, say, row 26. See: BackupFileXB ExitXB ~BackupFileXB Pack: CopyPack Src: BackupFileXBsrc Func: 5/System Copy an open BULLET key or data file. BULLET repacks and reindexes files in- place, requiring less disk space to perform the function. BackupFileXB allows you to safely copy a file before doing this. This function is recommended prior to packing a data file with PackRecordsXB since the data is very valuable. There is probably little need to do so when reindexing an index file since index files can be constructed very easily from the data file. See: StatHandleXB PackRecordsXB ReindexXB ~StatHandleXB Pack: StatHandlePack Src: StatHandleXBsrc Func: 6/System Get information on a DOS file handle number to determine if it is a BULLET file and if so, if that file is a BULLET key or data file. If the returned ID value is 0, the handle is to a BULLET index file. ID=1 then the handle is a BULLET .DBF file. ID= -1 then the handle is not a BULLET file. See: CreateDXB StatDXB StatKXB ~GetExtErrorXB Pack: XErrorPack Src: GetExtErrorXBsrc Func: 7/System Get the extended error information for the last operation. This information includes the extended error code, the error class, the recommended action, and the location of the error. See Errors_DOS for the extended error meaning an Errors_DOS_c for the class, action, and locus code meanings. Note that on fatal DOS errors, such as an open floppy drive door, the extended error code returned is 83 - fail on INT24. This indicates that the INT24 handler was invoked by DOS and that the INT24 handler told DOS to ignore the error. (BULLET invokes its own INT24 handler each time it accesses the DOS file system and restores it promptly after the access.) In such cases, this extended error code is less informative than the standard return code and, the other 'extended' information should be disregarded. (In fatal DOS errors the standard return code IS the extended error code.) This routine returns the extended error information for the LAST DOS system error. This information remains the same until the next DOS system error. See: CreateDXB Errors_DOS ~DVmonCXB Pack: DVmonPack Src: DVmonCXBsrc Func: 9/DEBUG Control BULLET debug monitor. This routine is available only in the debug engine. The monitor displays in realtime the state of a data file handle, or an index and data file handle pair if an index handle is specified. DVmonCXB is best used on dual-display systems in which the video output is sent to the secondary video monitor. In any case, a 4000-byte screen image is updated in real-time. To use the monitor, set mode=1, handle=file to monitor, and VideoSeg=segment address of 4000-byte area. The typical VideoSeg would be to video memory. If you have a color system as the main monitor and a mono as the secondary, set VideoSeg=&HB000. Detail system stats are continually updated to the secondary monitor. If you have a single monitor with at least 2 video pages, set VideoSeg to your base address plus the page size\16, typically &HB800+(4096\16). If you have only a single-page video system, you can allocate a 4000-byte memory area and update the video manually by moving it to your video display (80x25). See: CreateDXB StatDXB StatKXB ~CreateDXB Pack: CreateDataPack Src: CreateDXBsrc Func: 10/Mid-level Create a new BULLET .DBF data file. Before using this routine allocate a field description array of TYPE FieldDescTYPE for at least as many fields as are in the record. Conventional dBASE .DBF files have a FileID=3. Other possible FileIDs that you may come across are (in hex): 43h \__ are special-use Xbase IV DBF files, BULLET can process these file IDs 63h / since they are similar to ID type 3 83h --- this DBF file has an Xbase III/III+ memo field/file 88h --- this DBF file has an Xbase IV memo field/file In creating your .DBF files, specify FileID=3 to ensure compatibility across Xbase versions. BULLET makes no special use of the FileID byte. See: OpenDXB FieldDescTYPE CreateKXB ~OpenDXB Pack: OpenPack Src: OpenDXBsrc Func: 11/Mid-level Open an existing .DBF data file for use. You need to specify two things, the filename and the DOS file access mode. If the open succeeds, the DOS file handle is returned. Use this handle for all further access to this file. Each .DBF data file you open allocates 144+((1 + number of fields) * 32) bytes for internal use. This memory is not deallocated until you close the file with CloseDXB or execute ExitXB. You must open the data file before you can open (or create) any of its index files. See: CloseDXB OpenKXB ~CloseDXB Pack: HandlePack Src: CloseDXBsrc Func: 12/Mid-level Close an existing .DBF data file for use. Closing the file updates the file header and deallocates the memory used by this file. You MUST close all BULLET files before ending your program or file corruption may occur. To ensure that all files are closed in the event of an unscheduled program termination, use AtExitXB. See: StatDXB ExitXB CloseKXB ~StatDXB Pack: StatDataPack Src: StatDXBsrc Func: 13/Mid-level Get basic information on the BULLET .DBF data file handle specified. Information returned includes the number of records in the file, the record length, number of fields per record, and the date the file was last updated. Typically, your program will keep track of whether a particular handle belongs to a data file or a key file. In cases where this is not possible, call the StatHandleXB routine to determine what file type a handle is. Note that a just-created data file will have the LastUpdate date set to 0/0/0. See: ReadDHXB StatKXB StatHandleXB ~ReadDHXB Pack: HandlePack Src: ReadDHXBsrc Func: 14/Mid-level Reload the disk copy of the data header for the opened .DBF data file handle to the internal copy. In single-user, single-tasking systems this routine is not needed. However, in a multi-user or multi-tasking system it's possible, and desirable, for two or more programs to use the same data file. Consider this scenario: A data file has 100 records. Two programs access this data file, both opening it. Program 1 locks the file, adds a new record, then flushes and unlocks the file. Program 1 knows that there are now 101 records in the file. However, Program 2 is not aware of the changes that Program 1 made--it thinks that there are still 100 records in the file. This out-of-sync situation is easily remedied by having Program 2 reload the data header from the file on disk. How does Program 2 know that it needs to reload the header? It doesn't. Instead BULLET uses a simple yet effective approach when dealing with such problems. Whenever your program locks a file, BULLET automatically reloads the header. Whenever you unlock a file, BULLET automatically flushes the header. See: FlushDHXB ReadKHXB LockXB ~FlushDHXB Pack: HandlePack Src: FlushDHXBsrc Func: 15/Mid-level Write the internal copy of the data header for the opened .DBF data file handle to disk. The actual write occurs only if the header has been changed. This is to ensure that the data header on disk matches exactly the data header that is being maintained by BULLET. Also, this routine updates the operating system's directory entry for this file. Assume the following: A data file with 100 records. Your program opens the data file and adds 1 record. Physically, there are 101 records on disk. However, the header image of the data file on disk still reads 100 records. This isn't a problem, BULLET uses its internal copy of the data header and the internal copy does read 101 records. BUT, if there were a system failure now, the disk image would not get updated. After the system restart, BULLET opens the file, reads the header and thinks that there are 100 records. You lost a record. Now, if after that add above, your program issued a FlushDHXB, the header on disk is refreshed with the internal copy, keeping the two in-sync. Also, the routine updates the DOS directory entry, keeping things neat there as well. Still, it doesn't come without cost: flushing will take additional time, therefore, you may elect to flush periodically, or whenever the system is idle. See: CopyDHXB ReadDHXB FlushKHXB LockXB ~CopyDHXB Pack: CopyPack Src: CopyDHXBsrc Func: 16/Mid-level Copy the .DBF file structure of an open data file to another DOS file. This routine makes it easy for you to duplicate the structure of an existing .DBF file without having to specify all the information needed by CreateDXB. The resultant .DBF will be exactly like the source, including number of fields and field descriptions. It will contain 0 records. See: ZapDHXB CopyKHXB ~ZapDHXB Pack: HandlePack Src: ZapDHXBsrc Func: 17/Mid-level Delete all records for a .DBF data file. This routine is similar to CopyDHXB except for one major difference: ALL DATA RECORDS IN THE *SOURCE* FILE ARE PHYSICALLY DELETED, so be *careful*. If you have a .DBF file with 100 records and use ZapDHXB on it, all 100 records will be physically deleted and the file truncated to 0 records. There is no return from this routine. All data is gone. * C A U T I O N * See: CreateKXB CopyDHXB ZapKHXB ~CreateKXB Pack: CreateKeyPack Src: CreateKXBsrc Func: 20/Mid-level Create a new BULLET key file. Before you can create a key file, you must first have opened (and have created if necessary) the BULLET .DBF data file that it is to index. (BULLET couples index and data files tightly.) To create the key file, you need to provide the key expression, key flags, .DBF file link handle, and optionally, the code page ID, country code, and collate table. Key Expression The key expression is an ASCII character string composed of the elements that are to make up this index file's key. The key can be composed of any or all of the fields in the .DBF data record or sub-strings within any of those fields. Two functions are supported in evaluating a key expression. These are SUBSTR() and UPPER(). SUBSTR() is similar to BASIC's MID$() in that it extracts part of a string starting at a particular position for x number of characters. UPPER() is similar to UCASE$() in that it converts all lower-case letters to their upper-case equivalent. Since BULLET supports NLS, UPPER() conversion is not required for proper sorting of mixed-case text strings. See: CreateKXB_a CreateDXB -MORE- ~CreateKXB_a All names used in the key expression must be a valid field name in the DBF data file. Some sample key expressions given that the .DBF has the following fields: Fields... Valid key expressions FNAME C 25 0 kx = "LNAME" LNAME C 25 0 kx = "LNAME+FNAME" SSN C 9 0 kx = "SUBSTR(LNAME,1,4)+SUBSTR(FNAME,1,1)+SUBSTR(SSN,6,4) DEPT N 5 0 kx = "UPPER(LNAME+FNAME)" (for non-NLS index files) : : kx = "DEPT+SSN" (N- + C-type is valid for non-binary keys) Key Flags The key expression is used in conjunction with the key flags to determine the type of key generated. First, if your index file is to disallow duplicate keys, add 1 to KeyFlag. If you have a key composed of a character field(s) or portions thereof, you specify a KeyFlag = 2. This instructs BULLET that the sort order is left-to- right (proper mixed-case sorting is available, see code page ID). See: CreateKXB_b -MORE- ~CreateKXB_b If you have a key composed of a numeric field(s) or portions thereof, you can also specify a KeyFlag = 2. This instructs BULLET to treat the numeric field as a regular character field for sorting. To ensure proper sorting, you must decimal-align the +numeric strings in the .DBF data field, i.e., right-justify the numeric strings (dBASE .DBF numeric strings are stored as ASCII strings). These non-binary numeric fields are just like character fields to BULLET. In addition, if you have a key composed of a SINGLE numeric field (fld type N) and the field is an integer (NO DECIMAL POINT), you can specify a KeyFlag of 16 or 32. KeyFlag=16 is for a field known to be in word/integer range; KeyFlag=32 if the field is known to be in LongInt range. These KeyFlag values instruct BULLET to sort the key as a 16/32-bit BINARY value. It also stores the key as a 16- or 32-bit value (only 2 or 4 bytes) in the index, eventhough the data field is in ASCII (keyflag=16 or 32). Although not dBASE compatible, you may use BINARY FIELDS in your data records. dBASE always has ASCII data in the data fields, even if the field is numeric. For example, an N type field of 8.2 is stored as an ASCII text string in the data record, say, a string like " 1100.55". If you want dBASE compatibility your field data must also be ASCII. However, if you can forgo this requirement, you can use binary values in the fields. See: CreateKXB_c -MORE- ~CreateKXB_c To do this you must specify a field type of "B" (actually, anything but a "N") and, IF IT IS TO BE USED AS A KEY FIELD, also set the 16- or 32-bit KeyFlag. Unique and signed may also be flagged. The field length for a "B" field type is 2 or 4. Make sure the key flags match (2 if cINTEGER, 4 if cLONG). If you specify a binary key flag (for either N or B field types), you must also specify whether the field is to be treated as a signed or unsigned value. If values less than 0 are possible, add to KeyFlag the hex number &H8000. KeyFlag = %cUNIQUE + %cCHAR 'unique character key (NLS sort) KeyFlag = %cINTEGER + %cUNIQUE 'unique unsigned integer (binary sort) KeyFlag = %cUNIQUE + %cSIGNED + %cLONG 'unique signed long KeyFlag = %cCHAR 'character key with duplicates allowed KeyFlag = %cCHAR + %cINTEGER 'THIS IS AN INVALID KEY FLAGS! KeyFlag = %cLONG 'unsigned long (dups allowed) The following values are defined in PBULLET.BI: cUNIQUE=1, cCHAR=2, cINTEGER=&H10, cLONG=&H20, cNLS=&H4000, cSIGNED=&H8000 The NLS flag is assigned by BULLET. StatKXB is used to query KeyFlags. See: CreateKXB_d -MORE- ~CreateKXB_d The key expression you specify may be up to 136 characters, and evaluate out to 64 bytes (62 bytes if unique key is not specified). I.e, kx$ = "SUBSTR(..." can be up to 136 characters, and that the actual key built from this expression can be no longer that 64 bytes, or 62 if you did not specify UNIQUE. In general, shorter keys (the key itself, not the expression) offer better performance. DBF File Link Handle (XBlink) Since BULLET evaluates the key expression at CreateKXB, it must have access to the DBF file to verify that the key expression is valid. You must therefore supply CreateKXB with the OS file handle of the opened DBF data file. National Language Support (NLS) With DOS 3.3 and later, NLS is available. BULLET uses NLS to build the collate sequence table that it uses to ensure proper sorting of mixed-case keys as well as the sorting of foreign language alphabets. In order for BULLET to use the proper collate table, it must know what code page ID and coutry code to use. This table is made part of the index file so that all subsequent access to the index file maintains the original sort order, even if the MIS shop is moved to another location/computer system using another country code/code page. See: CreateKXB_e -MORE- ~CreateKXB_e Code Page ID To use the default code page ID of the computer in use, specify a code page ID of -1. This instructs BULLET to use the collate-sequence table as provided by MS-DOS running on the machine. You may also specify the code page ID for BULLET to use, but only if support for the code page ID is available on your machine. Look in your DOS manual under CUSTOMIZING FOR INTERNATIONAL USE for specific code page IDs and country codes. See also the COUNTRY and NLSFUNC commands. You may also specify a code page ID = 0 in which case no collate table is used. Country Code To use the default country code of the computer in use, specify a country code of -1. This instructs BULLET to use the collate-sequence table as provided by MS-DOS running on the machine. You may also specify the country code for BULLET to use, but only if support for the country code is available on your machine. Look in your DOS manual under CUSTOMIZING FOR INTERNATIONAL USE for specific code page IDs and country codes. See also the COUNTRY and NLSFUNC commands. You may also specify a country code = 0 in which case no collate table is used. Typically, you set CodePageID = -1, CoutryCode = -1 and CollatePtr = 0. See: CreateKXB_f -MORE- ~CreateKXB_f User-specified Collate Table If you are to use a MS-DOS supplied collate table (BOTH codepage ID and country codes are non-zero) then you do not need to specify a collate table--DOS will. The option to allow a user-specified collate table is to work around some DOS versions supplying incorrect collate tables. If you find that the DOS-supplied collate table is not valid (it's stored in the second sector of the file) for your country, you can supply the table to be used by pointing the CollatePtr variables to your in-memory version of a valid collate table. If you want to use the DOS-supplied collate table, you MUST set the CollatePtr variables = 0. Note: The collate table is a 256-byte table that contains the sort value of each character (0-255). For example, the first byte would be 0, second would be 1, and so on. Values for characters up to the lower-case letters (ASCII 97) are usually as you would expect: "A" has a value of 65. However, the lower-case letters have the same value as their upper-case counterparts: "a" also has a value of 65. BULLET uses this collate table to ensure proper sorting. If you specify EITHER code page ID OR country code = 0 then no collate table is used or built. Instead, sorting is done by standard ASCII sort. This is somewhat faster but less versatile. Use UPPER() for mixed-case sort if needed. See: OpenKXB CreateKXB ~OpenKXB Pack: OpenPack Src: OpenKXBsrc Func: 21/Mid-level Open an existing key file for use. Each key file that you open allocates 1264 bytes for internal use. This memory is not deallocated until you close the file with CloseKXB or execute ExitXB. You must open the data file before you can open its related index file because you must supply the handle of the data file that this index files indexes. See: CloseKXB OpenDXB ~CloseKXB Pack: HandlePack Src: CloseKXBsrc Func: 22/Mid-level Close an open key file. Closing the file updates the file header and deallocates the memory used by this file. You MUST close all BULLET files before ending your program or file corruption may occur. To ensure that all files are closed on the event of an unscheduled program termination, use AtExitXB. See: StatKXB ExitXB CloseDXB ~StatKXB Pack: StatKeyPack Src: StatKXBsrc Func: 23/Mid-level Get basic information on a BULLET key file handle specified. Information returned includes the number of keys in the file, the key length, the data file handle for this key, the last accessed record number of that data file, NLS information, and the key flags. Typically, your program will keep track of whether a particular handle belongs to a key file or a data file. In cases where this is not possible, call the StatHandleXB routine to determine what file type a handle is. See: ReadKHXB StatDXB StatHandleXB ~ReadKHXB Pack: HandlePack Src: ReadKHXBsrc Func: 24/Mid-level Reload the disk copy of the key header for the opened key file handle to the internal copy. In single-user, single-tasking systems this routine is not needed. However, in a multi-user or multi-tasking system it's possible, and desirable, for two or more programs to use the same data file. Consider this scenario: A key file has 100 keys. Two programs access this key file, both opening it. Program 1 locks the file, adds a new key, then flushes and unlocks the file. Program 1 knows that there are now 101 keys in the file. However, Program 2 is not aware of the changes that Program 1 made--it thinks that there are still 100 keys in the file. This out-of-sync situation is easily remedied by having Program 2 reload the key header from the file on disk. How does Program 2 know that it needs to reload the header? It doesn't. Instead BULLET uses a simple yet effective approach when dealing with such problems. Whenever your program locks a file, BULLET automatically reloads the header. Whenever you unlock a file, BULLET automatically flushes the header. See: FlushKHXB ReadDHXB FlushDHXB LockXB ~FlushKHXB Pack: HandlePack Src: FlushKHXBsrc Func: 25/Mid-level Write the internal copy of the key header for the opened key file handle to disk. The actual write occurs only if the header has been changed. This is to ensure that the key header on disk matches exactly the key header that is being maintained by BULLET. Also, this routine updates the operating system's directory entry for this file. Assume the following: A data file with 100 keys. Your program opens the key file and adds 1 key. Physically, there are 101 keys on disk. However, the header image of the data file on disk still reads 100 keys. This isn't a problem, BULLET uses its internal copy of the key header and the internal copy does read 101 keys. BUT, if there were a system failure now, the disk image would not get updated. After the system restart, BULLET opens the file, reads the header and thinks that there are 100 keys. You lost a key. Now, if after that add above, your program issued a FlushKHXB, the header on disk is refreshed with the internal copy, keeping the two in-sync. Also, the routine updates the DOS directory entry, keeping things neat there as well. Still, it doesn't come without cost: flushing will take additional time, therefore, you may elect to flush periodically, or whenever the system is idle. See: CopyKHXB ReadKHXB FlushDHXB LockXB ~CopyKHXB Pack: CopyPack Src: CopyKHXBsrc Func: 26/Mid-level Copy the key file structure of an open key file to another DOS file. This routine makes it easy for you to duplicate the structure of an existing key file without having to specify all the information needed by CreateKXB. The resultant key file will be exactly like the source, including key flags and key expression. It will contain 0 keys. See: ZapKHXB CopyKHXB ~ZapKHXB Pack: HandlePack Src: ZapKHXBsrc Func: 27/Mid-level Delete all keys for a key file. This routine is similar to CopyKHXB except for one major difference: ALL KEYS IN THE *SOURCE* FILE ARE PHYSICALLY DELETED, so be *careful*. If you have a key file with 100 keys and use ZapKHXB on it, all 100 keys will be physically deleted and the file truncated to 0 keys. There is no return from this routine. All data is gone. * C A U T I O N * See: GetDescriptorXB CopyKHXB ZapDHXB ~GetDescriptorXB Pack: DescriptorPack Src: GetDescriptorXBsrc Func: 30/Mid-level Get the field descriptor information for a field. You can specifiy either the fieldname or the field number (position of the field within the record where the first field is #1) to get info on. The field descriptor contains the following information: FIELDNAME 10 upper-case characters, A-Z and _ allowed, unused space is 0-filled and is 0-terminated (11 bytes, ASCII, byte 11 always=0) FIELDTYPE single ASCII character where C=character, N=numeric, D=date, L=logical, and M=memo field (1 byte, ASCII) FIELDLEN length of field: C=1-254, N=1-19, D=8 (yyyymmdd), L=1 (T/F/space), M=10, this is total field length (1 byte, binary) FIELDDC places right of decimal point if N field type, minimum if not 0 is 2, can be up to 6 or 8, non-N fields always 0 (1 byte, binary) See: GetRecordXB ~GetRecordXB Pack: AccessPack Src: GetRecordXBsrc Func: 31/Mid-level Get the physical record from the data file into a data buffer by record number. The data buffer is typically a TYPEd variable defined as the .DBF record itself is defined. For example, if the DBF record has 2 fields, LNAME and FNAME, then variable would be TYPEd as: TYPE RecBuffer tag AS STRING * 1 'the Xbase DBF delete flag (must be included) LastName AS STRING * 25 'same field length as the .DBF LNAME field FirstName AS STRING * 25 'same field length as the .DBF FNAME field END TYPE This method of accessing the data file does not use any indexing. Therefore, it typically is not used except for special purposes. The preferred method to access the data is by one of the keyed Get() routines. See: AddRecordXB GetEqualXB ~AddRecordXB Pack: AccessPack Src: AddRecordXBsrc Func: 32/Mid-level Append the record in the data buffer to the end of the DBF file. This method of adding a record does not involve any indexing. It is typically used to build a data file en masse and do the indexing after the entire .DBF file(s) has been built. If you have several thousand data records to be added at once, this method of building the DBF first and then using the ReindexXB routine is often faster than using the InsertXB routine for each record to add. The AddRecordXB is very fast. 400 recs/sec on an AT machine is typical. Over 2000 recs/sec can be added on a fast 486 machine--that's 120,000 records added per minute. The record number used is determined by BULLET and it is returned in AP.RecNo. See: UpdateRecordXB InsertXB ReindexXB ~UpdateRecordXB Pack: AccessPack Src: UpdateRecordXBsrc Func: 33/Mid-level Write the updated data record to the the physical record number. This method of writing the updated record must not be used if any field(s) in the record is used as a key field(s) and has been changed. This method of updating a record is very fast if you know that that update is not going to alter any field used as a key in any index file that uses it. You must, of course, first get the data record into the record buffer. Then you can change it and write the update out to disk with this routine. If you need to change a field(s) that is used as a key field or part of one, use the UpdateXB routine. UpdateXB not only dynamically updates all related index files if you change a key field, it also will undo any and all changes if an error occurs in the transaction. See: DeleteRecordXB GetRecordXB UpdateXB ~DeleteRecordXB Pack: AccessPack Src: DeleteRecordXBsrc Func: 34/Mid-level Tag the record at the physical record number as being deleted. This does not tag any in-memory copies of the record so be sure to mark any such copies as being deleted yourself. The first byte of every .DBF record is reserved for the delete tag. This tag is a space (ASCII 32) if the record is normal, or a * (ASCII 42) if it's marked as being deleted. This delete tag is a reserved field in the DBF record and as such is not defined as a formal field with a descriptor, etc. Make sure that you define your in-memory buffers to reserve the first byte for the delete tag. The Xbase DBF standard doesn't physically remove records marked as deleted from the data file. It doesn't mark them as available/reusable either. To physically remove records marked as deleted use PackRecordsXB. Records can be temporarily marked as deleted then recalled to normal status. The Key/Get routines (GetFirstXB, etc.) return the record number needed for this routine after each access in AP.RecNo. See: UndeleteRecordXB PackRecordsXB ~UndeleteRecordXB Pack: AccessPack Src: UndeleteRecordsrc Func: 35/Mid-level Tag the record at the physical record number as being normal (not deleted). This does not tag any in-memory copies of the record so be sure to mark any such copies as being normal yourself. The first byte of every .DBF record is reserved for the delete tag. This tag is a space (ASCII 32) if the record is normal, or a * (ASCII 42) if it's marked as being deleted. This delete tag is a reserved field in the DBF record and as such is not defined as a formal field with a descriptor, etc. Make sure that you define your in-memory buffers to reserve the first byte for the delete tag. The Xbase DBF standard does not physically remove records marked as deleted from the data file so you can "recall" then back to normal status as easily as you marked them deleted. See: PackRecordsXB DeleteRecordXB ~PackRecordsXB Pack: AccessPack Src: PackRecordsXBsrc Func: 36/Mid-level Rebuild the open DBF file by physically removing all records marked as deleted. Packing occurs in place using the existing file. It's recommended that you use BackupFileXB to copy the current DBF file before using this routine in case of a failure during the pack process. The newly packed file is truncated to reflect the current, actual size. If there are index files for this .DBF file, they MUST all be reindexed after the pack process by using ReindexXB. This routine dynamically allocates at least as many bytes as the length of the record. More if available. See: FirstKeyXB DeleteRecordXB BackupFileXB ReindexXB ~FirstKeyXB Pack: AccessPack Src: FirstKeyXBsrc Func: 40/Mid-level Retrieve the first key in index order from the index file. This routine does not access the .DBF file and so does not retrieve the data record. What it does do is locate the first key of the index, returning it, and also returning the record number within the .DBF that the key indexes. To retrieve the data record you can use the GetRecordXB routine. The preferred method, however, is to use the GetFirstXB. The key returned includes an enumerator if a non-unique index file is involved. The enumerator is a little-endian 16-bit value that serves to differentiate up to 65535 "identical", non-unique keys. It is attached to all keys of non- unique index files and occupies the last two bytes of the key, high/low order This routine is typically used to position the index file to the first key so as to allow forward in-order access to the keys by using NextKeyXB. See: EqualKeyXB GetFirstXB GetRecordXB ~EqualKeyXB Pack: AccessPack Src: EqualKeyXBsrc Func: 41/Mid-level Search for the exact key in the index file. This routine does not access the .DBF file and so does not retrieve the data record. What it does do is search for the key in the index, and if found, returns the record number within the .DBF that the key indexes. The key must be an exact match, including enumerator word if a non-unqiue index file. To retrieve the data record you can use the GetRecordXB routine. The preferred method, however, is to use the GetEqualXB. This routine will only find EXACT matches to the specified key (including the enumerator if applicable). However, even if the exact key is not found in the index file, the index file is positioned so that the next NextKeyXB retrieves the key that would have followed the unmatched specified key. For example, if the key to match was "KINGS" (a partial key in this case), EqualKeyXB would return a key not found error. If you were to now do a NextKeyXB, the next key would be returned, let's say it is "KINGSTON". See: NextKeyXB GetEqualXB GetRecordXB ~NextKeyXB Pack: AccessPack Src: NextKeyXBsrc Func: 42/Mid-level Retrieve the next key in index order from the index file. This routine does not access the .DBF file and so does not retrieve the data record. What it does do is retreive the next key of the index, returning it, and also returning the record number within the .DBF that the key indexes. To retrieve the data record you can use the GetRecordXB routine. The preferred method, however, is to use the GetNextXB. The key returned includes an enumerator if a non-unique index file is involved. This routine is typically called after the index file has first been positioned to a known key using either FirstKeyXB or EqualKeyXB, or after a previous NextKeyXB or even PrevKeyXB. What it basically does is get the key following the current key, and then make that key the new current key. See: PrevKeyXB GetNextXB GetRecordXB ~PrevKeyXB Pack: AccessPack Src: PrevKeyXBsrc Func: 43/Mid-level Retrieve the previous key in index order from the index file. This routine does not access the .DBF file and so does not retrieve the data record. What it does do is retreive the previous key of the index, returning it and also returning the record number within the .DBF that the key indexes. To retrieve the data record you can use the GetRecordXB routine. The preferred method, however, is to use the GetPrevXB. The key returned includes an enumerator if a non-unique index file is involved. This routine is typically called after the index file has first been positioned to a known key using either LastKeyXB or EqualKeyXB, or after a previous PrevKeyXB or even NextKeyXB. What it basically does is to get the key previous the current key, and then make that key the new current key. See: LastKeyXB GetPrevXB GetRecordXB ~LastKeyXB Pack: AccessPack Src: LastKeyXBsrc Func: 44/Mid-level Retrieve the last key in index order from the index file. This routine does not access the .DBF file and so does not retrieve the data record. What it does do is locate the last key of the index, returning it, and also returning the record number within the .DBF that the key indexes. To retrieve the data record you can use the GetRecordXB routine. The preferred method, however, is to use the GetLastXB. This routine is typically used to position the index file to the last key so as to allow reverse in-order access to the keys by using PrevKeyXB. See: StoreKeyXB GetLastXB GetRecordXB ~StoreKeyXB Pack: AccessPack Src: StoreKeyXBsrc Func: 45/Mid-level Insert the key into the index file in proper key order. This routine does not add the data record to the .DBF file. It only inserts the key and record number into the index file. Use InsertXB, instead. To do a complete data record and key insert, you could use AddRecordXB to add the data record to the .DBF, BuildKeyXB to construct the key, then StoreKeyXB to insert the key and record number information into the index file. If that key already exists and the file allows duplicate keys, you need to attach the proper enumerator word and retry StoreKeyXB. This is much too much to do. Instead, just use InsertXB. All these details including adding the data record and multi-key inserts are performed automatically with just the single call. See: DeleteKeyXB InsertXB ~DeleteKeyXB Pack: AccessPack Src: DeleteKeyXBsrc Func: 46/Mid-level Physically remove the specified key from the index file. This routine requires an EXACT key match for all bytes of the key, including the enumerator word if a non-unique index file is involved. This routine would seldom be used, typically, since deleted dBASE data records are only physically deleted during a PackRecordsXB and the index file is rebuilt afterward using ReindexXB. See: BuildKeyXB CurrentKeyXB ~BuildKeyXB Pack: AccessPack Src: BuildKeyXBsrc Func: 47/Mid-level Build the key for the specifed data record based on the key expression for the index file. If the index file is non-unique, a 0-value enumerator is attached. The enumerator is a little-endian 16-bit value that serves to differentiate up to 65535 "identical", non-unique keys. It is attached to all keys of non- unique index files and occupies the last two bytes of the key. This routine, like most of the mid-level routines, typically would not be used since the high-level access routines take care of this detail automatically. See: CurrentKeyXB StoreKeyXB ~CurrentKeyXB Pack: AccessPack Src: CurrentKeyXBsrc Func: 48/Mid-level Retrieve the current key value for the specified key file handle and also the data record number that it indexes. This routine is useful in that it retrieves on demand the actual key value of the last accessed key in the index file (and the data record number). Most often you don't need this information so it would be a waste of time and space for your program to explicitly track each current key for each index file that you have open. See: GetFirstXB ReindexXB DeleteKeyXB ~GetFirstXB Pack: AccessPack Src: GetFirstXBsrc Func: 60/High-level Retrieve the first indexed key and data record. The key returned includes an enumerator if a non-unique index file is involved. This routine is typically used to process a database in index order starting at the first ordered key (and its data record). After processing this first entry, subsequent in-order access of the database is achieved by using GetNextXB until the end of the database is reached. This routine, like all the high-level Get routines, fills in the AP.RecNo of the record accessed. In GetFirstXB's case, it fills AP.RecNo with the record number pointed to by the first key. Since this is so, the AP pack is primed for an UpdateXB after each high-level Get. Other methods to get the record number are to use CurrKeyXB or any of the Key routines (KeyFirstXB, etc.). See: GetEqualXB FirstKeyXB UpdateXB ~GetEqualXB Pack: AccessPack Src: GetEqualXBsrc Func: 61/High-level Search for the exact key in the index file and return its data record. This routine will only find EXACT matches to the specified key (including the enumerator if applicable). However, even if the exact key is not found in the index file, the index file is positioned so that the next GetNextXB retrieves the key that would have followed the unmatched specified key. For example, if the key to match was "KINGS" (a partial key in this case), GetEqualXB would return a key not found error. If you were to now do a GetNextXB, the next key and data record would be returned, let's say the key is "KINGSTON" and its data record is the data record for that key. Another GetNextXB would retrieve the key and record after that. (GetPrevXB can be used in this fashion too.) This routine, like all the high-level Get routines, fills in the AP.RecNo of the record accessed. In GetEqualXB's case, it fills AP.RecNo with the record number pointed to by the matched key. Since this is so, the AP pack is primed for an UpdateXB after each high-level Get. Other methods to get the record number are to use CurrKeyXB or any of the Key routines (KeyEqualXB, etc.). See: GetNextXB EqualKeyXB ~GetNextXB Pack: AccessPack Src: GetNextXBsrc Func: 62/High-level Retrieve the next indexed key and its data record. The key returned includes an enumerator if a non-unique index file is involved. This routine is typically calld after the index file has first been positioned to a known key using either GetFirstXB or GetEqualXB, or after a previous GetNextXB or even GetPrevXB. What it basically does is get the key and data record following the current key, and then make that key the new current key. This routine, like all the high-level Get routines, fills in the AP.RecNo of the record accessed. In GetNextXB's case, it fills AP.RecNo with the record number pointed to by the next key. Since this is so, the AP pack is primed for an UpdateXB after each high-level Get. Other methods to get the record number are to use CurrKeyXB or any of the Key routines (KeyNextXB, etc.). See: GetPrevXB NextKeyXB ~GetPrevXB Pack: AccessPack Src: GetPrevXBsrc Func: 63/High-level Retrieve the previous indexed key and record. The key returned includes an enumerator if a non-unique index file is involved. This routine is typically called after the index file has first been positioned to a known key using either GetLastXB or GetEqualXB, or after a previous GetPrevXB or even GetNextXB. What it basically does is to get the key and data record previous the current key, and then make that key the new current key. This routine, like all the high-level Get routines, fills in the AP.RecNo of the record accessed. In GetPrevXB's case, it fills AP.RecNo with the record number pointed to by the previous key. Since this is so, the AP pack is primed for an UpdateXB after each high-level Get. Other methods to get the record number are to use CurrKeyXB or any of the Key routines (KeyPrevXB, etc.). See: GetLastXB PrevKeyXB ~GetLastXB Pack: AccessPack Src: GetLastXBsrc Func: 64/High-level Retrieve the last indexed key and record. This routine is typically used to process a database in reverse index order starting at the last ordered key (and its data record). After processing this last entry, subsequent reverse-order access of the database is achieved by using GetPrevXB until the top of the database is reached. This routine, like all the high-level Get routines, fills in the AP.RecNo of the record accessed. In GetLastXB's case, it fills AP.RecNo with the record number pointed to by the last key. Since this is so, the AP pack is primed for an UpdateXB after each high-level Get. Other methods to get the record number are to use CurrKeyXB or any of the Key routines (KeyLastXB, etc.). See: InsertXB LastKeyXB ~InsertXB Pack: AccessPack Src: InsertXBsrc Func: 65/High-level Add the data record to data file and insert the related key(s) into the linked index file(s). This routine is used to add new entries into a database, one at a time. The data record is first added to the data file, then for each related index file, a key is inserted into the appropriate index file. Up to 32 index files can be automatically maintained for each data file. This and several other routines are transaction-based. If a failure occurs prior to the routine's completion, all changes made to the database by the routine will be backed-out and the database (data and related index file(s)) effectively restored to its original state. If the routine failed to complete, the function return value is the number of the pack that caused the failure. The pack's Stat is checked to determine the error code. If the function return value is 0, YOU MUST STILL check the first pack's Stat. If it's non-zero, then the failure occured with the data record. See: UpdateXB StoreKeyXB ~UpdateXB Pack: AccessPack Src: UpdateXBsrc Func: 66/High-level Modify an existing data record (identified by record number) and automatically perform any index file updates needed to keep the index file(s) in sync. If any key fields changed between the original record and the new one, this routine updates the appropriate index file(s) by replacing the original key(s) with new the key(s) based on the updated data record. Up to 32 index files can be automatically maintained for each data file. Get routines (GetFirstXB, etc.) set the AP.RecNo of the record that UpdateXB uses. This and several other routines are transaction-based. If a failure occurs prior to the routine's completion, all changes made to the database by the routine will be backed-out and the database (data and related index file(s)) effectively restored to its original state. If the routine failed to complete, the function return value is the number of the pack that caused the failure. The pack's Stat is checked to determine the error code. If the function return value is 0, YOU MUST STILL check the first pack's Stat. If it's non-zero, then the failure occured with the data record. See: ReindexXB UpdateRecordXB ~ReindexXB Pack: AccessPack Src: ReindexXBsrc Func: 67/High-level Reindex all related index files for a data file. The index file(s) must already exist and be open. Any existing key data is overwritten by the new key data. In other words, if you have a 10MByte index file, ReindexXB uses the same file space building the news keys over the old. This results in a less fragmented disk and also minimizes disk space needed. You can also create a new, empty index file and reindex to that. This would be useful, for instance, if you needed to create a temporary index file--something that you'd use for a report, say, then delete after the report. This routine creates a TEMPORARY work file in either the current directory or, if the DOS environment variable TMP is defined, in the TMP= directory. The size of this file is approx. bytes = (RECORDS * (KEYLEN+6)). ReindexXB can operate in as little as 32K of available memory and can use up to 128K. The resultant index file(s) are optimized for minimum size AND maximum retrieval speed. If the routine failed to complete, the function return value is the number of the pack that caused the failure. The pack's Stat is checked to determine the error code. A return value of zero indicates no error occured. See: LockXB PackRecordsXB ~LockXB Pack: AccessPack Src: LockXBsrc Func: 80/Network Lock all bytes in the index file handle(s) for exclusive use by the current process and reload the index file header(s) from disk. Also lock all bytes in the related data file and reload the data file header from disk. The files must have been opened with the appropriate share attribute and not in compatibility mode. SHARE.EXE MUST be installed or DOS error 1 is issued. This routine is transaction-based and will lock all index files specified in AccessPack and the data file. If any lock fails, all previous locks by this routine are released. The return value indicates which access pack failed, if any. This value is used as the index into the AccessPack group for you to identify the error code. See LockXBsrc for determining this exactly. Use the DriveRemoteXB and/or FileRemoteXB to determine if locking is necessary. If the files are on a remote drive then it is best to use locking. Locking may also be necessary on multitasking local machines accessing shared files. This routine is a combination of LockKeyXB and LockDataXB. See: UnlockXB LockKeyXB LockDataXB DriveRemoteXB FileRemoteXB ~UnlockXB Pack: AccessPack Src: UnlockXBsrc Func: 81/Network Unlock all bytes in the specified file handle(s) (previously locked) and flush the file header(s) to disk (flush done before lock(s) released). Also unlock all bytes in the related data file and flush the data file header to disk. The files must have been opened with the appropriate share attribute and not in compatibility mode. SHARE.EXE MUST be installed or DOS error 1 is issued. This routine is transaction-based and will unlock all index files specified in AccessPack and the data file. If an unlock fails the routine exits with a return value indicating which access pack failed. This value is used as the index into the AccessPack group for you to identify the error code. Note that this routine does not attempt to re-lock those files unlocked successfully if an error occurs in the transaction. If an error does occur (unlikely) you will need to provide for unlocking the remaining files manually with the UnlockKeyXB and UnlockDataXB routines. You should not rely on the operating system to automatically unlock files when they're closed. This routine is a combination of UnlockKeyXB and UnlockDataXB. See: LockKeyXB UnlockKeyXB UnlockDataXB ~LockKeyXB Pack: AccessPack Src: LockKeyXBsrc Func: 82/Network Lock all bytes in the index file handle(s) for exclusive use by the current process and reload the index file header(s) from disk. The files must have been opened with the appropriate share attribute and not in compatibility mode. SHARE.EXE MUST be installed or DOS error 1 is issued. This routine is transaction-based and will lock all index files specified in AccessPack. If any lock fails, all previous locks by this routine are released. The return value indicates which access pack failed, if any. This value is used as the index into the AccessPack group for you to identify the error code. The advantage of using region locks (LockKeyXB locks the entire file region) to control file access is that the file does not need to be opened/closed using the Deny Read/Write sharing attribute. Opening the file for Deny None, and controlling subsequent access with region locks, allows for faster processing since files do not need to be constantly opened and closed, as they would if access were controlled by opening with Deny Read/Write. See: UnlockKeyXB LockXB ~UnlockKeyXB Pack: AccessPack Src: UnlockKeyXBsrc Func: 83/Network Unlock all bytes in the specified file handle(s) (previously locked) and flush the file header(s) to disk (flush done before lock(s) released). The files must have been opened with the appropriate share attribute and not in compatibility mode. SHARE.EXE MUST be installed or DOS error 1 is issued. This routine is transaction-based and will unlock all index files specified in AccessPack. If an unlock fails the routine exits with a return value indicating which access pack failed. This value is used as the index into the AccessPack group for you to identify the error code. All file locks should be released when exclusive access in no longer needed. It is not recommended that you end your program without having released active file locks. This is especially a valid concern for DOS versions prior to 5.0. DOS 5 releases locks on files that are closed. See: LockDataXB UnlockXB ~LockDataXB Pack: AccessPack Src: LockDataXBsrc Func: 84/Network Lock all bytes in the file handle's data file for exclusive use by the current process and reload the data file header from disk. You must set AP.RecNo=0 to do this. To lock a single record, set AP.RecNo=record# to lock. The files must have been opened with the appropriate share attribute and not in compatibility mode. SHARE.EXE MUST be installed or DOS error 1 is issued. This routine locks the specified data file. If the handle specified is that of an index file, that index file's related data file handle is used. For single- record locks, AP.Handle must have a data file handle specified. Header loading is not performed if locking a single record. The advantage of using region locks (LockDataXB locks the entire file region) to control file access is that the file does not need to be opened/closed using the Deny Read/Write sharing attribute. Opening the file for Deny None, and controlling subsequent access with region locks, allows for faster processing since files do not need to be constantly opened and closed, as they would if access were controlled by opening with Deny Read/Write. See: UnlockDataXB ~UnlockDataXB Pack: AccessPack Src: UnlockDataXBsrc Func: 85/Network Unlock all bytes in the specified file handle (previously locked) and flush the data file header to disk (flush done before lock released). To do this you must set AP.RecNo=0. To unlock a single record, set AP.RecNo=record# to unlock. The files must have been opened with the appropriate share attribute and not in compatibility mode. SHARE.EXE MUST be installed or DOS error 1 is issued. This routine unlocks the specified data file. If the handle specified is that of an index file that index file's related datafile handle is used. For single- record unlocks, AP.Handle must have a data file handle specified. Flushing is not performed if unlocking a single record. All file locks should be released when exclusive access in no longer needed. It is not recommended that you end your program without having released active file locks. This is especially a valid concern for DOS versions prior to 5.0. DOS 5 releases locks on files that are closed. See: DriveRemoteXB ~DriveRemoteXB Pack: RemotePack Src: DriveRemoteXBsrc Func: 86/Network Determine if specified drive is remote (default drive=0, A:=1, B=2, C=3...). This routine uses INT21/44/sub function 09. In addition to returning the IsRemote state, this routine sends back the result of the DX register and also the install state of SHARE.EXE. The meaning of the bitflags in Flags are (where IsRemote=0): Bit Meaning drive... 1 1=uses 32-bit sectoring 6 1=accepts Generic IOCTL (for INT21/44/0D,0E,0Fh) 7 1=accepts Query IOCTL Device (INT21/44/11h) 9 1=is local but shared by other computers in the network 11 1=accepts Does-Device-Use-Removable-Media (INT21/44/08) 13 1=requires media descriptor in FAT 14 1=accepts Receive/Send Control Data from Block Device (INT21/44/04,05) 15 1=is Substitution drive (set by the DOS SUBST command) (all other bits=0) See: FileRemoteXB LockXB ~FileRemoteXB Pack: RemotePack Src: FileRemoteXBsrc Func: 87/Network Determine if specified handle of file or device is remote. This routine uses INT21/44/sub function 0Ah. In addition to returning the IsRemote state, this routine sends back the result of the DX register and also the install state of SHARE.EXE. Flags bit 7=1 then handle is device, =0 then handle is file. Bit Meaning DEVICE... Bit Meaning DEVICE...(cont) 0 1=is console input device 13 1=is named pipe 1 1=is console output device 15 1=is remote, 0=is local 2 1=is null device (all other bits=0) 3 1=is clock device Bit Meaning FILE... 4 1=is special device 0-5 xxxxxx=drive number (0=A...) 5 1=is in binary mode, 0=in ASCII 6 1=has not been written to 6 0=returns EOF if device is read 12 1=is NoInherit 11 1=is network spooler 14 1=date/time not set at close 12 1=is NoInherit 15 1=is remote, 0=is local (all other bits=0) See: SetRetriesXB DriveRemoteXB LockXB ~SetRetriesXB Pack: SetRetriesPack Src: SetRetriesXBsrc Func: 88/Network Set the number of times DOS retries disk operations after a failure due to file-sharing operations (locked file regions from LockXB routines). This routine uses INT21/44/sub function 0Bh. By default DOS retries an operation 3 times (without pausing between attempts) before returning an error to the application. If you change the default values it's recommended that the default state be restored before your application ends (Retries=3, Pause=1). These values are pretty much determined by trial-and-error. You may find that adding a delay between retries returns fewer access-denied errors. See: DeleteFileDOS LockXB ~DeleteFileDOS Pack: DOSFilePack Src: DeleteFileDOSsrc Func: 100/DOS Delete the specified file. This routine uses DOS INT21/41 (interrupt 21h function 41h). See: RenameFileDOS ~RenameFileDOS Pack: DOSFilePack Src: RenameFileDOSsrc Func: 101/DOS Rename a file. May also be used to move the file to a new directory within the partition. This routine uses DOS INT21/56. If the specified directory differs from the file's directory, the file's directory entry is moved to the new directory. For example, if the FilenamePtr filename is C:\LP100\PROJ93A.INF and the NewFilenamePtr filename is C:\ARCH\PROJ93A.INA, the file is essentially renamed and also moved to the \ARCH directory. See: CreateFileDOS ~CreateFileDOS Pack: DOSFilePack Src: CreateFileDOSsrc Func: 102/DOS Create a new file. This routine uses INT21/3C. The specified filename/pathname must NOT already exist. The file created is not left open. You must OpenFileDOS to use it. The attribute used during the create can be: ATTRIBUTE VALUE MEANING Normal 0 normal access permitted to file Read-Only 1 read-only access permitted to file Hidden 2 file does not appear in directory listing System 4 file is a system file Volume 8 FILENAME used as volume label if no current label Archive 20h file is marked for archiving See: AccessFileDOS OpenFileDOS ~AccessFileDOS Pack: DOSFilePack Src: AccessFileDOSsrc Func: 103/DOS Determine if the specified file can be accessed with the specified access/sharing mode. This routine uses INT21/3D and INT21/3E. Basically, a Does-File-Exist routine. It uses the specified access/sharing attributes when trying to open the file. For example, if you specify DFP.Attr = &H42 (R/W access + Deny None sharing) and use AccessFileDOS on a Read-Only DOS file, the return value would be DOS error 5, Access Denied. See: OpenFileDOS ~OpenFileDOS Pack: DOSFilePack Src: OpenFileDOSsrc Func: 104/DOS Open the specified file with the specified access/sharing mode. This routine uses INT21/3D. ACCESS VALUE MEANING Read-only 0 open for read-only access Write-only 1 open for write-only access Read/Write 2 open for read/write access SHARE Compatibility 0 any process may share file (not recommended) Deny Read/Write 10h no other process may share file Deny Write 20h no other process may share file for write Deny Read 30h no other process may share file for read Deny None 40h any process may share file except in Compatibilty INHERIT mode NoInheritFlag 80h if set child processes do not inherit file handles (child process cannot inherit handle > 20) The file access mode is a combination of ACCESS + SHARE + INHERIT. See: SeekFileDOS OpenPack ~SeekFileDOS Pack: DOSFilePack Src: SeekFileDOSsrc Func: 105/DOS Position the DOS file pointer of the specified file to the specified position. This routine uses INT21/42. The position is a 32-bit value and is relative to either the start of the file, the current file pointer position, or the end of the file. Method Meaning 0 start move from the start of file (offset is a 32-bit unsigned value) 1 start move at the current position (offset a signed value) 2 start move at the end of file (offset a signed value) For example, to move to the 511th byte of a file (byte 0 being the first), set the offset value to 511 and use Method 0. On return, the absolute offset value of the new position is returned. This is useful with Method 2 since you can specify an offset of 0 and have the file length returned. Never position the file pointer to before the start of file. See: ReadFileDOS ~ReadFileDOS Pack: DOSFilePack Src: ReadFileDOSsrc Func: 106/DOS Read from the file or device the specified number of bytes into a buffer. This routine uses INT21/3F. On block devices (such as disks) input starts at the current file position and the file pointer is repositioned to the last byte read +1. It is possible to read less than the bytes specified without an error being generated. Compare the bytes to read with the returned bytes read value. If less then end of file was reached during the read, if 0 then file was at EOF. By using DOS's predefined handles you can read from the keyboard (STDIN) by using the STDIN handle, 0. The input will terminate after all specified bytes have been read or after a CR (ASCII 0Dh). If more bytes are entered than were requested, the next read will retrieve those excess bytes. Therefore, it's suggested that you specify 129 bytes to input (DOS will process 127+CR/LF bytes maximum when reading the STDIN device). Post-process the entered data by scanning for the CR/LF. See: ExpandFileDOS ~ExpandFileDOS Pack: DOSFilePack Src: ExpandFileDOSsrc Func: 107/DOS Expands the specified file by the specified number of bytes. This routine uses INT21/42 and INT21/40. This routine is useful in pre-allocating disk space. By reserving disk space in advance you can guarantee that enough disk space will be available for a future operation (especially if more than 1 process is running). You'll also be able ensure that the disk space that a file does use is as contiguous as possible. Database systems are dynamic and their files typically allocate new space on an as-needed basis. This dynamic allocation can cause parts of a file to be located throughout the disk system, possibly affecting performance drastically. By pre-allocating the disk space you can be assured of consistent throughput performance since the file is contiguous. See: WriteFileDOS ~WriteFileDOS Pack: DOSFilePack Src: WriteFileDOSsrc Func: 108/DOS Write to the file or device the specified number of bytes from a buffer. This routine uses INT21/40. If the number of bytes written is less than the specified bytes, this routine returns a -2 error code (or 65554 unsigned). On block devices (such as disk) output starts at the current file position, and the file pointer is repositioned to the last byte written +1. If the specified bytes to write is 0, the file is truncated at the current file pointer position. By using DOS's predefined handles you can write to the screen (STDOUT) by using the STDOUT handle, 1. See: CloseFileDOS ~CloseFileDOS Pack: DOSFilePack Src: CloseFileDOSsrc Func: 109/DOS Close the file flushing any internal buffers, releasing any locked regions, and update the directory entry to the correct size, date, and time. This routine uses INT21/3E. If you have opened a file using the DOS open routine you should close it when you no longer need it. This routine can be used to close the predefined DOS handles (0-4) and make those handles available for reuse. Typically handles 0 and 1 should not be closed by an application since they are the STDIN and STDOUT that DOS uses for the current application (keyboard and screen). Since BULLET provides for up to 250 user file handles for your applications it isn't necessary for you to eek 3 more file handles by closing handles 2-4. See: MakeDirDOS ~MakeDirDOS Pack: DOSFilePack Src: MakeDirDOSsrc Func: 110/DOS Create a new subdirectory. This routine uses INT21/39. See: DeleteFileDOS ~AccessPack Src: InsertXBsrc Func: InsertXB and many more TYPE AccessPack 'AP (AP is recommended pack name for DIM, AP.Func=) Func AS INTEGER 'varies Stat AS INTEGER 'ret:completion status Handle AS INTEGER 'OS handle RecNo AS LONG ' in:rec number to get/delete/update (if applicable) ' in:set to single rec# to lock or set to 0 to lock all 'ret:record number of data record accessed RecPtrOff AS INTEGER 'far pointer to record storage buffer RecPtrSeg AS INTEGER KeyPtrOff AS INTEGER 'far pointer to search key buffer KeyPtrSeg AS INTEGER NextPtrOff AS INTEGER 'far pointer to next key access pack NextPtrSeg AS INTEGER 'or 0:0 if end of link or if N/A END TYPE '22 The NextPtr variables are only used by InsertXB, UpdateXB, ReindexXB, and the LockXB routines. NextPtr is used as a link to the next related access pack, if any. Not all entries are used by all routines. Generally, any routine that gets/puts user data to the database uses this pack. See: BreakPack ~BreakPack Src: BreakXBsrc Func: BreakXB TYPE BreakPack 'BP Func AS INTEGER '4 Stat AS INTEGER 'ret:completion status Mode AS INTEGER '=0 disable Ctrl-C/Ctrl-Break, 1=restore END TYPE '6 bytes A simple pack. See: CopyPack ~CopyPack Src: BackupFileXBsrc Func: BackupFileXB, CopyDHXB, CopyKHXB TYPE CopyPack 'CP Func AS INTEGER '5=BackupFileXB,16=CopyDHXB,26=CopyKHXB Stat AS INTEGER 'ret:completion status Handle AS INTEGER 'handle of BULLET file FilenamePtrOff AS INTEGER 'far pointer to filenameZ FilenamePtrSeg AS INTEGER '(filename must end with a CHR$(0))) END TYPE '10 See: CreateDataPack ~CreateDataPack Src: CreateDXBsrc Func: CreateDXB TYPE CreateDataPackTYPE 'CDP Func AS INTEGER '10 Stat AS INTEGER 'ret:completion status FilenamePtrOff AS INTEGER 'far pointer to filenameZ to create FilenamePtrSeg AS INTEGER '(filename must end with a CHR$(0)) NoFields AS INTEGER 'number of fields per record FieldListPtrOff AS INTEGER 'far pointer to field list FieldListPtrSeg AS INTEGER '(field list is of type FieldDescTYPE ) FileID AS INTEGER 'file signature byte, usually=3 END TYPE '16 The filename pointed at by the FilenamePtr variables should be a fixed-length string (e.g., FILEN AS STRING * 80) so that VARSEG/VARPTR can be used to get its memory address. The filename must end with a CHR$(0) immediately following the last character of the filename: FILEN = YourFilename$ + CHR$(0). The FieldListPtr variables point to an array of type FieldDescTYPE. This array is dimensioned for as many fields as there are in the record and contains the field descriptors, one for each field. See: CreateKeyPack FieldDescTYPE ~CreateKeyPack Src: CreateKXBsrc Func: CreateKXB TYPE CreateKeyPack 'CKP Func AS INTEGER '20 Stat AS INTEGER 'ret:completion status FilenamePtrOff AS INTEGER 'far pointer to filenameZ FilenamePtrSeg AS INTEGER '(filename must end with a CHR$(0)) KeyExpPtrOff AS INTEGER 'far pointer to key expressionZ KeyExpPtrSeg AS INTEGER '(key expression must end with a CHR$(0)) XBlink AS INTEGER 'BULLET XB data file handle this key file indexes KeyFlags AS INTEGER 'bit 0=unique,1=char,4=int,5=long,(E=NLS),F=signed CodePageID AS INTEGER 'code page ID for NLS, -1 to use system default CountryCode AS INTEGER 'country code number for NLS, -1 to use default CollatePtrOff AS INTEGER 'far pointer to programmer-supplied collate table CollatePtrSeg AS INTEGER 'or 0:0 if using system-determined NLS table END TYPE '24 Bit 14 in KeyFlags (0Eh) is set by BULLET during CreateKXB if a collate table is present. See: DescriptorPack is_NLS ~DescriptorPack Src: GetDescriptorXBsrc Func: GetDescriptorXB TYPE DescriptorPack 'DP Func AS INTEGER '30 Stat AS INTEGER 'ret:completion status Handle AS INTEGER 'BULLET data file handle to get information on FieldNumber AS INTEGER 'field number to get info on, or if 0 then... FD AS FieldDescTYPE '...search for DP.FD.FieldName END TYPE '40 GetDescriptorXB allows you to get the field descriptor info for a particular field number (as in the first field, or the 10th field, etc.) or, if you don't know the physical field number, the routine can also get the info for a field by field name. To get the info for field number, say 5, set DP.FieldNumber = 5. The DP.FD structure element is filled in with field 5's information. To get the info for a field by fieldname, say LASTNAME, set DP.FieldNumber=0 & DP.FD.FieldName = "LASTNAME" + STRING$(11,0)--the fieldname must be zero-filled and zero-terminated--adding 11 ASCII zeroes ensures this requirement. See: DOSFilePack FieldDescTYPE ~DOSFilePack Src: AccessFileDOSsrc Func: AccessFileDOS (all routines ending with DOS) TYPE DOSFilePack 'DFP Func AS INTEGER 'varies, see DeleteFileDOS for first of DOS routines Stat AS INTEGER 'ret:completion status FilenamePtrOff AS INTEGER 'far pointer to filenameZ FilenamePtrSeg AS INTEGER '(filename must end with a CHR$(0)) Handle AS INTEGER 'in: handle to access ret: handle opened ASmode AS INTEGER 'open access/sharing mode Bytes AS INTEGER 'in: bytes to read ret: bytes read SeekOffset AS LONG 'seek to file position Method AS INTEGER 'seek method BufferPtrOff AS INTEGER 'far pointer to read/write buffer BufferPtrSeg AS INTEGER Attr AS INTEGER 'file create directory entry attribute NewNamePtrOff AS INTEGER 'far pointer to new filenameZ for rename NewNamePtrSeg AS INTEGER '(filename must end with a CHR$(0)) END TYPE '30 All of the xDOS routines use this pack. Often only a few of the structure elements are used by any one of the routines. Set only those needed. See: DVmonPack ~DVmonPack Src: DVmonCXBsrc Func: DVmonCXB TYPE DVmonPackTYPE 'THIS ROUTINE IS AVAILABLE ONLY IN THE DEBUG ENGINE Func AS INTEGER '9 Stat AS INTEGER 'ret:completion status Mode AS INTEGER '=0 disable montitoring, =1 enable Handle AS INTEGER 'file handle to monitor VideoSeg AS INTEGER 'segment to write screen image (e.g., &HB800) END TYPE '10 bytes This routine is supplied only in the BULLET debug engine. It displays real-time monitoring information of a .DBF file or index and .DBF file pair including searches, seeks, hits, current record number, current key, key node contents, key node pointers, stack state, key and record counts, and other info. By using the HereSeg value returned from StatKXB you can locate the searches, seeks, and hits data at (DEF SEG=HereSeg: Seeks&=GetLongAt&(517): DEF SEG) +513 Searches AS LONG ;keys searched for since open +517 Seeks AS LONG ;nodes seeked since open +521 Hits AS LONG ;seeks satisfied without disk access See: ExitPack ~ExitPack Src: InitXBsrc Func: ExitXB, AtExitXB TYPE ExitPack 'EP Func AS INTEGER '1=ExitXB, 2=AtExitXB Stat AS INTEGER 'ret:completion status END TYPE '4 bytes See: FieldDescTYPE ~FieldDescTYPE Src: CreateDXBsrc Func: CreateDXB TYPE FieldDescTYPE 'this TYPE is used by CreateDataPack ONLY FieldName AS STRING * 11 'zero-filled field name (use only ASCII 65-90,95) FieldType AS STRING * 1 'C-har,N-umeric,D-ate,L-ogical,M-emo FieldDA AS LONG '=0,reserved FieldLength AS STRING * 1 'C=1-254,N=1-19(varies),D=8,L=1,M=10 FieldDC AS STRING * 1 'decimal places for FieldType=N (0,2-15) A1 AS INTEGER '=0,reserved A2 AS INTEGER '=0,reserved filler AS STRING * 10 '=0,reserved END TYPE '32 If you can can forgo dBASE compatility you can use the B field type. This type is for fields that contain binary data (all dBASE fields contain ASCII text or numeric strings). If you specify a FieldType = "B" for, say an integer field, use a FieldLen = 2. If the field is a long integer, use FieldLen = 4. You can also use this non-standard field type for indexing. See CreateKXB for more. See: HandlePack CreateDataPack CreateKXB ~HandlePack Src: CloseDXBsrc Func: CloseDXB, ReadDHXB, FlushDHXB, ZapDHXB CloseKXB, ReadKHXB, FlushKHXB, ZapKHXB 'HP TYPE HandlePack '12=CloseDXB,14=ReadDHXB,15=FlushDHXB,17=ZapDHXB Func AS INTEGER '22=CloseKXB,24=ReadKHXB,25=FlushKHXB,27=ZapKHXB Stat AS INTEGER 'ret:completion status Handle AS INTEGER 'handle of BULLET file END TYPE '6 See: InitPack ~InitPack Src: InitXBsrc Func: InitXB TYPE InitPack 'IP Func AS INTEGER '0 Stat AS INTEGER 'ret:completion status JFTmode AS INTEGER 'expand JFT if non-zero DOSver AS INTEGER 'ret:DOS version Version AS INTEGER 'ret:BULLET version * 100 ExitOff AS INTEGER 'ret:far pointer to ExitXB routine, offset ExitSeg AS INTEGER 'ret:segment of ExitXB END TYPE '12 bytes See: MemoryPack ~MemoryPack Src: MemoryXBsrc Func: MemoryXB TYPE MemoryPack 'MP Func AS INTEGER '3 Stat AS INTEGER 'ret:completion status Memory AS LONG 'ret:largest free OS memory block END TYPE '8 bytes See: OpenPack ~OpenPack Src: OpenDXBsrc Func: OpenDXB, OpenKXB TYPE OpenPack 'OP Func AS INTEGER '11=OpenDXB,21=OpenKXB Stat AS INTEGER 'ret:completion status Handle AS INTEGER 'ret:OS handle of file opened FilenamePtrOff AS INTEGER 'far pointer to filenameZ to open FilenamePtrSeg AS INTEGER '(filename must end with a CHR$(0)) ASmode AS INTEGER 'DOS access-sharing mode (see OpenFileDOS) xbHandle AS INTEGER 'if opening key file this is its related data file END TYPE '14 '(if opening data file xbHandle is not used) 'Note: you must supply xbHandle on index file opens See: RemotePack OpenFileDOS ~RemotePack Src: DriveRemoteXBsrc Func: DriveRemoteXB, FileRemoteXB TYPE RemotePack 'RP Func AS INTEGER '86=DriveRemoteXB,87=FileRemoteXB Stat AS INTEGER 'ret:completion status Handle AS INTEGER 'handle/drive depending on routine IsRemote AS INTEGER 'ret:0=local,1=remote Flags AS INTEGER 'ret:dx register as returned by DOS IsShare AS INTEGER 'ret:0=SHARE.EXE not loaded END TYPE '12 See: SetRetriesPack ~SetRetriesPack Src: SetRetriesXBsrc Func: SetRetriesXB TYPE SetRetriesPack 'SRP Func AS INTEGER '88 Stat AS INTEGER 'ret:completion status Mode AS INTEGER '0=set DOS default else use Pauses/Retries below Pause AS INTEGER '0-65535 loop counter between retries Retries As INTEGER '0-65535 retries to access locked file END TYPE '10 The default values for Retries is 3 and Pause is 1. The Pause value is used as a simple loop counter used to waste time. This loop IS dependent on CPU power so values are not portable across different machines. Do not use unrealistic values. For example, don't set Retries to 30,000 unless you really want to wait for DOS to try 30,000 times before returning an error! See: StatDataPack ~StatDataPack Src: StatDXBsrc Func: StatDXB TYPE StatDataPackTYPE 'SDP Func AS INTEGER '13 Stat AS INTEGER 'ret:completion status Handle AS INTEGER 'BULLET data file to get status on FileType AS STRING * 1 'ret:1=BULLET XB data file Dirty AS STRING * 1 'ret:0=not changed Recs AS LONG 'ret:records in file RecLen AS INTEGER 'ret:record length Fields AS INTEGER 'ret:fields per record () f1 AS STRING * 1 'reserved (1=update DVmon) LUyear AS STRING * 1 'ret:binary, year file last updated LUmonth AS STRING * 1 'ret:month --LUs are 0 if DBF newly created LUday AS STRING * 1 'ret:day HereSeg AS INTEGER 'ret:this file's control segment filler AS STRING * 10 'reserved END TYPE '32 See: StatKeyPack ~StatKeyPack Src: StatKXBsrc Func: StatKXB TYPE StatKeyPack 'SKP Func AS INTEGER '23 Stat AS INTEGER 'ret:completion status Handle AS INTEGER 'BULLET key file to get status on FileType AS STRING * 1 'ret:0=BULLET XB key file Dirty AS STRING * 1 'ret:0=not changed Keys AS LONG 'ret:keys in file KeyLen AS INTEGER 'ret:key length XBlink AS INTEGER 'ret:related BULLET XB data file handle XBrecno AS LONG 'ret:record number attached to current key HereSeg AS INTEGER 'ret:this file's control segment CodePageID AS INTEGER 'ret:code page ID number of key file sort CountryCode AS INTEGER 'ret:country code number of key file sort CollateTableSize AS INTEGER 'ret: size of collate table, 0 or 256 KeyFlags AS INTEGER 'ret:bit 0=unique,1=char,4=int,5=long,E=NLS,F=signed filler AS STRING * 2 END TYPE '32 See: StatHandlePack ~StatHandlePack Src: StatHandleXBsrc Func: StatHandleXB TYPE StatHandlePack 'SHP Func AS INTEGER '6 Stat AS INTEGER 'ret:completion status Handle AS INTEGER 'file handle to get information on ID AS INTEGER 'ret:0=XB index,1=XB data file,-1=not BULLET handle END TYPE '8 bytes See: XErrorPack ~XErrorPack Src: GetExtErrorXBsrc Func: GetExtErrorXB TYPE XErrorPack 'XEP Func AS INTEGER '7 Stat AS INTEGER 'ret:extended error Class AS INTEGER 'ret:error class Action AS INTEGER 'ret:suggested action Location AS INTEGER 'ret:error location END TYPE '10 bytes See: AccessPack Errors_DOS ~Errors_BULLET (200-209) 200 key not found - The search key for Equal was not matched exactly. Next/Prev routines can be used to continue search from point of mismatch. 201 key already exists - Attempted to add a key that already exists in the index file created to allow only unique keys. 202 end of file - A Next routine is past the last key of the index file. 203 top of file - A Prev routine is before the first key of the index file. 204 key file empty - A key access was attempted with no keys in the index file. 205 key type unknown - Generally indicates a corrupt index header (keyflags unknown at key insert). reserved,206-207 208 no more nodes - The index file has reached full capacity (32MB). ReindexXB can often shrink an index file by 30 to 50%. 209 key file corrupt - The index file is corrupt (write attempt to node 0). See: Errors_BULLET_b ~Errors_BULLET_b (210-232) 210 key file corrupt - The index file is corrupt (internal overflow). reserved,211-219 220 incorrect DOS version - BULLET requires DOS 3.3 or later. 221 invalid key length - The key is > 62 bytes (or 64 if unique specified). 222 file not open - The specified handle is not an open BULLET file. reserved,223 224 invalid record number - The specified record number is < 0, past the last record number in the .DBF, or is > 16,777,215. reserved,225-227 228 invalid filetype - The specified handle is not the correct type for the operation (i.e., specifying a data file handle for a key file operation). reserved,229-232 See: Errors_BULLET_c ~Errors_BULLET_c (233-243) 233 init not active - InitXB must be called before all others except MemoryXB. 234 init already active - InitXB has already been called. Use ExitXB first to call InitXB more than once per process. (Make sure the xxP.Func <> 0.) 235 too many indexes - BULLET can handle up to 32 index files per transaction record with the InsertXB and UpdateXB routines. Contact the author if you need to allow for more than 32 index files/transaction record. reserved,236-239 240 invalid key expression - The CreateKXB key expression could not be evaluated. reserved,241 242 field not found - The fieldname was not found in the descriptor area. 243 invalid field count - Too many fields were specified or the specified field number is past the last field. See: Errors_BULLET_d ~Errors_BULLET_d (244-255) reserved,244-249 250 invalid country info - The specifed country code or code page ID is not valid or not installed (according to DOS). 251 invalid collate table size - The specified country code/code page ID uses a collate-sequence table > 256 bytes (2-byte characters as with Kanji). 252 invalid keyflags - The specified keyflags are invalid. reserved,253-254 255 evaluation mode shutdown - BULLET evaluation period has completed. You can reinstall to continue evaluation, though you may want to consider your motives for reinstalling since the original evaluation period has expired. This error occurs only after the evaluation period has expired. It is not recommended that you continue to use BULLET after the evaluation period. It is possible for no 255 error to be generated for quite some time since it occurs only under certain load conditions and then only when certain routine sequences are performed. The specified evaluation period of 21 days should be adhered to. See: Errors_BASIC ~Errors_BASIC 1 NEXT without FOR 24 device timeout 2 syntax error 25 device fault 3 RETURN without GOSUB 26 FOR without NEXT 4 out of DATA 27 out of paper 5 illegal function call 29 WHILE without WEND 6 overflow 30 WEND without WHILE 7 out of memory reserved,31-32 8 label not defined 33 duplicate label 9 subscript out of range reserved,34 10 duplicate definition 35 subprogram not defined 11 division by zero reserved,36 12 illegal in direct mode 37 argument-count mismatch 13 type mismatch 38 array not defined 14 out of string space reserved,39 reserved,15 40 variable required 16 string formula too complex reserved,41-49 17 cannot continue 50 FIELD overflow 18 function not defined 51 internal error 19 no RESUME 52 bad file name or number 20 RESUME without error 53 file not found reserved,21-23 54 bad file mode See: Errors_BASIC_b ~Errors_BASIC_b 55 file already open reserved,77-79 56 FIELD statement active 80 feature removed 57 device I/O error 81 invalid name 58 file already exists 82 table not found 59 bad record length 83 index not found reserved,60 84 invalid column 61 disk full 85 no current record 62 input past end of file 86 duplicate value for unique index 63 bad record number 87 invalid operation on null index 64 bad file name 88 database needs repair reserved,65-66 89 insufficient ISAM buffers 67 too many files 68 device unavailable 69 communication-buffer overflow 70 permission denied 71 disk not ready 72 disk-media error 73 feature unavailable 74 rename across disks 75 path/File access error 76 path not found See: Errors_DOS ~Errors_DOS -2 disk full or unexpected end of file -1 bad filename 0 no error 1 function not supported 19 disk write protected 2 file not found 20 unknown unit 3 path not found 21 drive not ready 4 too many open files 22 unknown command 5 access denied (see Specs_Networks) 23 data error (CRC) 6 handle invalid 24 bad request structure length 7 MCBs destroyed 25 seek error 8 not enough memory 26 unknown medium type 9 memory block address invalid 27 sector not found 10 environment invalid 28 printer out of paper 11 format invalid 29 write fault 12 access code invalid 30 read fault 13 data invalid 31 general failure reserved-0Eh 32 sharing violation 15 disk drive invalid 33 lock violation 16 cannot remove current directory 34 disk change invalid/wrong disk 17 not same device 35 FCB unavailable 18 no more files 36 sharing buffer overflow See: Errors_DOS_b ~Errors_DOS_b 37 code page mismatched 58 incorrect response from network 38 handle EOF 59 unexpected network error 39 handle disk full 60 incompatible remote adapter reserved-28h 61 print queue full reserved-29h 62 no spool space reserved-2Ah 63 not enough space to print file reserved-2Bh 64 network name deleted reserved-2Ch 65 network access denied reserved-2Dh 66 network device type incorrect reserved-2Eh 67 network name not found reserved-2Fh 68 network name limit exceeded reserved-30h 69 NETBIOS session limit exceeded reserved-31h 70 sharing temporarily paused 50 network request not supported 71 network request not accepted 51 remote computer not listening 72 print/disk redirection paused 52 duplicate name on network reserved-49h 53 network pathname not found reserved-4Ah 54 network busy reserved-4Bh 55 network device no longer exists reserved-4Ch 56 NETBIOS command limit exceeded reserved-4Dh 57 network adapter hardware error reserved-4Eh See: Errors_DOS_c ~Errors_DOS_c reserved-4Fh DOS Class Codes 80 file exists 81 duplicate FCB 1 out of resources 7 application error 82 cannot make 2 temporary situation 8 not found 83 fail on INT24 3 authorization 9 bad format 84 out of structures 4 internal error 10 locked 85 already assigned 5 hardware failure 11 media failure 86 invalid password 6 system failure 12 already exists 87 invalid parameter 13 unknown 88 network write fault reserved-59h DOS Action Codes DOS Locus Codes 90 sys comp not loaded 1 retry immediately 1 unknown 2 delay and retry 2 block device 3 reenter input 3 network 4 abort ASAP 4 serial device 5 abort immediately 5 memory 6 ignore error 7 user intervention See: Errors_BULLET ~InitXBsrc Func: InitXB Pack: InitPack Func: 0/System DIM IP AS InitPack DIM EP AS ExitPack IP.Func = %InitXB 'InitXB defined in PBULLET.BI IP.JFTmode = 1 'expand JFT to 255 handles stat = BULLET(IP) IF stat = 0 THEN DOSmajor = IP.DOSver\256 DOSminor = IP.DOSver MOD 256 BULLETver = IP.Version SegExitXB = IP.ExitSeg OffExitXB = IP.ExitOff EP.Func = %AtExitXB 'register ExitXB with _atexit shutdown routine stat = BULLET(EP) ENDIF IF stat THEN 'error See: ExitXBsrc ~ExitXBsrc Func: ExitXB Pack: ExitPack Func: 1/System DIM EP AS ExitPack EP.Func = %ExitXB 'ExitXB defined in PBULLET.BI stat = BULLET(EP) The return value from ExitXB is currently always 0. See: AtExitXBsrc ~AtExitXBsrc Func: AtExitXB Pack: ExitPack Func: 2/System DIM IP AS InitPack DIM EP AS ExitPack IP.Func = %InitXB IP.JFTmode = 1 stat = BULLET(IP) IF stat = 0 THEN DOSmajor = IP.DOSver\256 DOSminor = IP.DOSver MOD 256 SegExitXB = IP.ExitSeg OffExitXB = IP.ExitOff 'register ExitXB with _atexit shutdown routine EP.Func = %AtExitXB 'AtExitXB defined in PBULLET.BI stat = BULLET(EP) 'the return value is that returned from the compiler '_atexit routine, 0=okay, anything else is an error ENDIF 'indicating, for example, the _atexit register is full IF stat THEN 'error See: MemoryXBsrc ~MemoryXBsrc Func: MemoryXB Pack: MemoryPack Func: 3/System DIM MP AS MemoryPack WorkSpace& = 50000 'a value at least 40K or so MemoryNeeded& = WorkSpace& + (144&+ ((1+NF)*32)) * TotalOpenDataFiles _ + (1264& * TotalOpenKeyFiles) MP.Func = %MemoryXB stat = BULLET(MP) IF MP.Memory < MemoryNeeded& THEN QBheap& = SETMEM(-MemoryNeeded& + 4096) 'release what we need+QB fudge MP.Func = %MemoryXB stat = BULLET(MP) IF MP.Memory < MemoryNeeded& THEN 'not enough memory END IF MP.Memory does not reflect memory available through DOS in the UMB area. It's possible that all memory requests can be satisfied by UMB RAM. Consult a DOS 5+ programmer reference for more information on this (see DOS INT21/58 for more). In the QuickBASIC/BASIC PDS environment do not call SETMEM(-x) more than once. See: BreakXBsrc ~BreakXBsrc Func: BreakXB Pack: BreakPack Func: 4/System DIM BP AS BreakPack BP.Func = %BreakXB 'BreakXB defined in PBULLET.BI BP.Mode = 0 'disable Ctrl-C/Ctrl-BREAK (do nothing on those keys) stat = BULLET(BP) 'stat=0 always If BreakXB is called multiple times with the same BP.mode each time, only the first is acted on. You can set BP.mode = 1 to restore the default handlers (those installed originally) and then again set BP.Mode = 0 to disable them. ExitXB calls this routine automatically as part of the BULLET shutdown to restore the original default break handlers. See: BackupFileXBsrc ~BackupFileXBsrc Func: BackupFileXB Pack: CopyPack Func: 5/System DIM AP AS AccessPack DIM CP AS CopyPack DIM TempStr AS STRING * 80 'used as fixed-length string for VARPTR() 'so same source QB AND BASIC7 /Fs compatible TempStr = NewFilename$ 'assign the var-len string to fixed-len str CP.Func = %BackupFileXB 'defined in PBULLET.BI CP.Handle = DataHandle 'handle of data file to backup CP.FilenamePtrOff = VARPTR(TempStr) 'ALWAYS use VARPTR() on fixed-len string CP.FilenamePtrSeg = VARSEG(TempStr) 'ALWAYS use VARSEG() on fixed-len string stat = BULLET(CP) 'copy the data file to the file NewFilename$ IF stat = 0 THEN AP.Func = %PackRecordsXB AP.Handle = DataHandle stat = BULLET(AP) ENDIF IF stat THEN 'error See: StatHandleXBsrc ~StatHandleXBsrc Func: StatHandleXB Pack: StatHandlePack Func: 6/System DIM SHP AS StatHandlePack DIM SKP AS StatKeyPack DIM SDP AS StatDataPack SHP.Func = %StatHandleXB 'defined in PBULLET.BI SHP.Handle = TheHandleNumber stat = BULLET(SHP) IF SHP.ID = 0 THEN 'handle belongs to an index file (index file/key file) SKP.Func = %StatKXB 'get key stats -- see StatKXB/StatDXB for more SKP.Handle = PassedHandleNumber ' on the SKP structure stat = BULLET(SKP) ELSEIF SHP.ID = 1 THEN '.DBF data file 'get DBF stats 'error not a BULLET file type ENDIF See: GetExtErrorXBsrc ~GetExtErrorXBsrc Func: GetExtErrorXB Pack: XErrorPack Func: 7/System 'an error just occured in the range 1 to 199 as returned in one of the 'pack.Stat variables (current max DOS error is 90 (5Ah)) 'remember, transaction-based routines return a bad pack index in the return 'stat value, which you use to check the appropriate pack.Stat variable DIM XEP AS XErrorPack XEP.Func = %GetExtErrorXB 'defined in PBULLET.BI stat = BULLET(XEP) IF stat <> 0 THEN PRINT "Extended Codes --" PRINT " error: "; XEP.Stat PRINT " error class: "; XEP.Class PRINT "recommened action: "; XEP.Action PRINT " location: "; XEP.Location PRINT "No error" ENDIF See: DVmonCXBsrc StatKXB ~DVmonCXBsrc Func: DVmonCXB Pack: DVmonPack Func: 9/DEBUG 'at this point a data file and a key file have been opened 'kf is that key file's DOS handle DIM DV AS DVmonPack DV.Func = %DVmonCXB 'defined in PBULLET.BI DV.Mode = 1 'enable monitoring DV.Handle = kf 'monitor key file handle, kf (and its XBlink data file) DV.VideoSeg = &HB800+(4096\16) 'output to color screen, page 1 (pages 0 to ?) stat = BULLET(DV) 'stat=0 always even if not DEBUG ENGINE For two-monitor systems (with a color monitor as the main system) output should be directed to &HB000, the mono monitor's video memory. DVmonCXB stands for Dual Video Monitor Control XB. See: CreateDXBsrc ~CreateDXBsrc Func: CreateDXB Pack: CreateDataPack Func: 10/Mid-level DIM CDP AS CreateDataPack DIM TempStr AS STRING * 80 'used as fixed-length string for VARPTR() REDIM FD(1 TO 2) AS FieldDescTYPE 'field descriptions for each of the fields... '...in the record (record has 2 fields) TempStr = filename$ + CHR$(0) 'assign filename to a fixed-len string... '...so we can use VARSEG/VARPTR 'build FD first for each of the fields in the record FD(1).FieldName = "STUDENT" + STRING$(10,0) 'must be zero-filled FD(1).FieldType = "C" 'a character field type FD(1).FieldLength = CHR$(20) 'up to 20 characters (a byte field--use CHR$()) FD(1).FieldDC = CHR$(0) 'non-numeric has no decimal places FD(2).FieldName = "SCORE" + STRING$(10,0) FD(2).FieldType = "N" FD(2).FieldLength = CHR$(3) 'dBASE numeric format, allow for "100" FD(2).FieldDC = CHR(0) 'no decimal places used for this field '(cont) 'for BINARY FieldType="B" see FieldDescTYPE See: CreateDXBsrc_a -MORE- ~CreateDXBsrc_a 'build the CDP CDP.Func = %CreateDXB 'defined in PBULLET.BI CDP.FilenamePtrOff = VARPTR(TempStr) 'point to filenameZ (Z=0-terminated str) CDP.FilenamePtrSeg = VARSEG(TempStr) CDP.NoFields = 2 'this example has 2 fields CDP.FieldListPtrOff = VARPTR(FD(1)) 'point to the first field decription... CDP.FieldListPtrSeg = VARPTR(FD(1)) '...defined in the previous screen CDP.FileID = 3 'standard dBASE file ID stat = BULLET(CDP) 'create the DBF data file IF stat THEN 'error Normally this code would be written as a generalized FUNCTION. The CDP could be a global allocation (DIM SHARED CDP AS CreateDataPack) and the FD() would also be (REDIM [SHARED] FD(1 TO 1) AS FieldDescTYPE) and REDIM'd to needed size when used. A possible header: DECLARE FUNCTION CreateDBF% (filename$, NoFields%, FD() AS FieldDescTYPE) To create a DBF you'd need just the filename, number of fields, REDIM the FD() array (if needed), fill in FD() array, then call the function. Look at the source code examples on the distribution disk for more. See: OpenDXBsrc CreateDXBsrc ~OpenDXBsrc Func: OpenDXB Pack: OpenPack Func: 11/Mid-level DIM OP AS OpenPack DIM TempStr AS STRING * 80 'used as fixed-length string for VARPTR() TempStr = filename$ + CHR$(0) 'assign filename to a fixed-len string... OP.Func = %OpenDXB 'defined in PBULLET.BI OP.FilenamePtrOff = VARPTR(TempStr) 'point to filenameZ (Z=0-terminated str) OP.FilenamePtrSeg = VARSEG(TempStr) OP.ASmode = ReadWrite + DenyNone 'defined in PBULLET.BI stat = BULLET(OP) IF stat THEN 'error The ASmode (access/sharing mode) determines how the operating system controls access to the file. See OpenFileDOS for the meanings of the various ASmodes. See: CloseDXBsrc OpenFileDOS ~CloseDXBsrc Func: CloseDXB Pack: HandlePack Func: 12/Mid-level DIM HP AS HandlePack HP.Func = %CloseDXB 'defined in PBULLET.BI HP.Handle = datahandle 'handle of the file to close stat = BULLET(HP) IF stat THEN 'error See: StatDXBsrc ~StatDXBsrc Func: StatDXB Pack: StatDataPack Func: 13/Mid-level DIM SDP AS StatDataPack SDP.Func = %StatDXB 'defined in PBULLET.BI SDP.Handle = datahandle 'data handle to get stats on stat = BULLET(SDP) 'must be a data handle, use StatHandleXB if you... IF stat = 0 THEN '...don't know the type of file a handle's for 'SDP.FileType is set to 1 'SDP.Dirty is set to 1 if the file has changed (0=not changed) 'SDP.Recs = number of records in the DBF file 'SDP.RecLen = record length 'SDP.Fields = number of fields in the record 'SDP.f1 is reserved 'SDP.LUyear = year file last updated, binary (year = ASC(SDP.LUyear)) 'SDP.LUmonth = month, binary 'SDP.LUday = day, binary 'SDP.HereSeg is set to this handle's control segment (location in memory) 'error ENDIF See: ReadDHXBsrc ~ReadDHXBsrc Func: ReadDHXB Pack: HandlePack Func: 14/Mid-level DIM HP AS HandlePack HP.Func = %ReadDHXB 'defined in PBULLET.BI HP.Handle = datahandle 'handle of file whose header you want to reload stat = BULLET(HP) IF stat THEN 'error This routine is automatically called by the network lock routines. See: FlushDHXBsrc LockDataXB ~FlushDHXBsrc Func: FlushDHXB Pack: HandlePack Func: 15/Mid-level DIM HP AS HandlePack HP.Func = %FlushDHXB 'defined in PBULLET.BI HP.Handle = datahandle 'handle of file you want to flush stat = BULLET(HP) IF stat THEN 'error Note that the physical write to disk is performed only if the file has changed since the open or last flush. This routine is automatically called by the network unlock routines. See: CopyDHXBsrc UnlockDataXB ~CopyDHXBsrc Func: CopyDHXBsrc Pack: CopyPack Func: 16/Mid-level DIM CP AS CopyPack DIM TempStr AS STRING * 80 'used as fixed-length string for VARPTR() TempStr = filename$ + CHR$(0) 'assign filename to a fixed-len string... CP.Func = %CopyDHXB 'defined in PBULLET.BI CP.Handle = datahandle 'handle of file to copy from (the source) CP.FilenamePtrOff = VARPTR(TempStr) 'far pointer to filenameZ for copy CP.FilenamePtrSeg = VARPTR(TempStr) '(the destination) stat = BULLET(CP) IF stat THEN 'error See: ZapDHXBsrc ~ZapDHXBsrc Func: ZapDHXB Pack: HandlePack Func: 17/Mid-level DIM HP AS HandlePack HP.Func = %ZapDHXB 'defined in PBULLET.BI HP.Handle = datahandle 'handle of file you want !ZAP! stat = BULLET(HP) IF stat THEN 'error Note that this removes ALL data records from the data file. See: CreateKXBsrc ~CreateKXBsrc Func: CreateKXB Pack: CreateKeyPack Func: 20/Mid-level 'This code assumes that the datafile was created as in CreateDXBsrc, and that 'the datafile was opened as in OpenDXBsrc. kx$ = "SUBSTR(STUDENT,1,5)" 'key expression, a non-unique, character key DIM CKP AS CreateKeyPack DIM TempStr AS STRING * 80 'used as fixed-length string for VARPTR() DIM TempStr2 AS STRING * 136 'used as fixed-length string for VARPTR() TempStr = filename$ + CHR$(0) 'assign filename to a fixed-len string TempStr2 = kx$ + CHR$(0) 'same for key expression string '(cont) See: CreateKXBsrc_a -MORE- ~CreateKXBsrc_a CKP.Func = %CreateKXB 'defined in PBULLET.BI CKP.FilenamePtrOff = VARPTR(TempStr) 'far pointer to filenameZ CKP.FilenamePtrSeg = VARSEG(TempStr) CKP.KeyExpPtrOff = VARPTR(TempStr2) 'far pointer to key expressionZ CKP.KeyExpPtrSeg = VARSEG(TempStr2) CKP.XBlink = datahandle 'the datafile handle returned from OpenDXB CKP.KeyFlags = cCHAR '-KEYFLAGS- are defined in PBULLET.BI CKP.CodePageID = -1 'use DOS default code page ID CKP.CountryCode = -1 'use DOS default country code CKP.CollatePtrOff = 0 'no user-supplied collate table... CKP.CollatePtrSeg = 0 '... stat = BULLET(CKP) IF stat THEN 'error Normally this code would be written as a generalized FUNCTION. The CKP could be a global allocation (DIM SHARED CKP AS CreateKeyPack). A possible header: DECLARE FUNCTION CreateNewIndex% (filename$, datahandle, kx$, KeyFlags%) To create an index file you'd need just filename, datahandle, key expression, and key flags (country code info if not using default), then call the function. See: OpenKXBsrc CreateKXBsrc ~OpenKXBsrc Func: OpenKXB Pack: OpenPack Func: 21/Mid-level DIM OP AS OpenPack DIM TempStr AS STRING * 80 'used as fixed-length string for VARPTR() TempStr = filename$ + CHR$(0) 'assign filename to a fixed-len string... OP.Func = %OpenKXB 'defined in PBULLET.BI OP.FilenamePtrOff = VARPTR(TempStr) 'point to filenameZ (Z=0-terminated str) OP.FilenamePtrSeg = VARSEG(TempStr) OP.ASmode = ReadWrite + DenyNone 'defined in PBULLET.BI OP.XBlink = datafilehandle 'OpenKXB needs to know the data file handle stat = BULLET(OP) IF stat THEN 'error The ASmode (access/sharing mode) determines how the operating system controls access to the file. See OpenFileDOS for the meanings of the various ASmodes. Before you can open an index file you must first open its associated data file. See: CloseKXBsrc OpenFileDOS ~CloseKXBsrc Func: CloseKXB Pack: HandlePack Func: 22/Mid-level DIM HP AS HandlePack HP.Func = %CloseDXB 'defined in PBULLET.BI HP.Handle = indexhandle 'handle of the file to close stat = BULLET(HP) IF stat THEN 'error See: StatKXBsrc ~StatKXBsrc Func: StatKXB Pack: StatKeyPack Func: 23/Mid-level DIM SKP AS StatKeyPack SKP.Func = %StatKXB 'defined in PBULLET.BI SKP.Handle = indexhandle 'handle to get stats on stat = BULLET(SKP) 'must be index handle, use StatHandleXB if you... IF stat = 0 THEN '...don't know the type of file a handle's for 'SKP.FileType is set to 0 'SKP.Dirty is set to 1 if the file has changed (0=not changed) 'SKP.Keys = number of key in the index file (index file=key file) 'SKP.KeyLen = physical key length (1-64 bytes) 'SKP.XBlink = datafile handle that this index file is associated with 'SKP.XBrecno is set to record number associated with last accessed key 'SKP.HereSeg is set to this handle's control segment (location in memory) 'SKP.CodePageID returns this index file's permanent code page ID 'SKP.CountryCode returns this index file's permanent country code 'SKP.CollateTableSize = 0 (no collate table present) or 256 (table present) 'SKP.KeyFlags = key flags specifed at CreateKXB (except NLS flag may be set) ELSE (NLS flag is bit 14, &H4000) 'error See: ReadKHXBsrc ~ReadKHXBsrc Func: ReadKHXB Pack: HandlePack Func: 24/Mid-level DIM HP AS HandlePack HP.Func = %ReadKHXB 'defined in PBULLET.BI HP.Handle = indexhandle 'handle of file whose header you want to reload stat = BULLET(HP) IF stat THEN 'error This routine is automatically called by the network lock routines. See: FlushKHXBsrc LockKeyXB ~FlushKHXBsrc Func: FlushKHXB Pack: HandlePack Func: 25/Mid-level DIM HP AS HandlePack HP.Func = %FlushKHXB 'defined in PBULLET.BI HP.Handle = indexhandle 'handle of file you want to flush stat = BULLET(HP) IF stat THEN 'error Note that the physical write to disk is performed only if the file has changed since the open or last flush. This routine is automatically called by the network unlock routines. See: CopyKHXBsrc UnlockKeyXB ~CopyKHXBsrc Func: CopyKHXBsrc Pack: CopyPack Func: 26/Mid-level DIM CP AS CopyPack DIM TempStr AS STRING * 80 'used as fixed-length string for VARPTR() TempStr = filename$ + CHR$(0) 'assign filename to a fixed-len string... CP.Func = %CopyKHXB 'defined in PBULLET.BI CP.Handle = indexhandle 'handle of file to copy from (the source) CP.FilenamePtrOff = VARPTR(TempStr) 'far pointer to filenameZ for copy CP.FilenamePtrSeg = VARPTR(TempStr) '(the destination) stat = BULLET(CP) IF stat THEN 'error See: ZapKHXBsrc ~ZapKHXBsrc Func: ZapKHXB Pack: HandlePack Func: 27/Mid-level DIM HP AS HandlePack HP.Func = %ZapKHXB 'defined in PBULLET.BI HP.Handle = indexhandle 'handle of file you want !ZAP! stat = BULLET(HP) IF stat THEN 'error Note that this removes ALL keys from the index file. See: GetDescriptorXBsrc ~GetDescriptorXBsrc Func: GetDescriptorXB Pack: DescriptorPack Func: 30/Mid-level DIM DP AS DescriptorPack DP.Func = %GetDescriptorXB 'defined in PBULLET.BI DP.Handle = datahandle 'handle of file IF FieldNumber > 0 THEN DP.FieldNumber = FieldNumber 'field number to get info on ELSE 'or, if 0, then DP.FieldNumber = 0 DP.FD.FieldName = fieldname$ + STRING$(10,0) 'fieldname$ to get info on ENDIF stat = BULLET(DP) IF stat = 0 THEN 'DP.FD.FieldName is set to field name 'DP.FD.FieldType is set to field type 'DP.FD.FieldLen is set to field length 'DP.FD.FieldDC is set to field DC ELSE 'error The DP.FD.FieldLen (and DP.FD.FieldDC) is a byte string so you need to use a FldLen=ASC(DP.FD.FieldLen) for its integer value See: GetRecordXBsrc ~GetRecordXBsrc Func: GetRecordXB Pack: AccessPack Func: 31/Mid-level TYPE RecordTYPE 'simple DBF record layout tag AS STRING * 1 < code AS STRING * 4 THE FIRST BYTE OF YOUR RECORD TYPES MUST BE TAG! bday AS STRING * 8 END TYPE DIM RecBuff AS RecordTYPE 'record has 2 fields, code/C/4.0, bday/D/8.0 DIM AP AS AccessPack AP.Func = %GetRecordXB 'defined in PBULLET.BI AP.Handle = datahandle 'handle to get record from AP.RecNo = RecnoToGet& 'record number to get AP.RecPtrOff = VARPTR(RecBuff) 'read record from disk into RecBuff AP.RecPtrSeg = VARSEG(RecBuff) stat = BULLET(AP) IF stat = 0 THEN PRINT RecBuff.code ; ", " ; RecBuff.bday 'sample output: 4321, 19331122 'error See: AddRecordXBsrc ~AddRecordXBsrc Func: AddRecordXB Pack: AccessPack Func: 32/Mid-level TYPE RecordTYPE 'simple DBF record layout tag AS STRING * 1 < code AS STRING * 4 THE FIRST BYTE OF YOUR RECORD TYPES MUST BE TAG! bday AS STRING * 8 END TYPE DIM RecBuff AS RecordTYPE 'record has 2 fields, code/C/4.0, bday/D/8.0 DIM AP AS AccessPack 'be sure to init the tag field to a space (ASCII 32) RecBuff.tag = " " : RecBuff.code = "1234" : RecBuff.bday = "19331122" AP.Func = %AddRecordXB 'defined in PBULLET.BI AP.Handle = datahandle 'handle to add record to AP.RecPtrOff = VARPTR(RecBuff) 'write record from RecBuff to disk... AP.RecPtrSeg = VARSEG(RecBuff) '...at next available record number stat = BULLET(AP) IF stat = 0 THEN PRINT "Record number used by AddRecordXB was "; AP.RecNo ELSE 'error See: UpdateRecordXBsrc ~UpdateRecordXBsrc Func: UpdateRecordXB Pack: AccessPack Func: 33/Mid-level 'see GetRecordXBsrc for this source example's preliminary code AP.Func = %GetRecordXB 'first get the record to update AP.Handle = datahandle AP.RecNo = RecnoToGet& ' AP.RecPtrOff = VARPTR(RecBuff) ' Do NOT use UpdateRecordXB to change AP.RecPtrSeg = VARSEG(RecBuff) ' any field(s) used in a key expression. stat = BULLET(AP) ' Instead use UpdateXB. IF stat = 0 THEN ' RecBuff.dbay = "19591122" 'change only non-key portions of record AP.Func = %UpdateRecordXB 'defined in PBULLET.BI stat = BULLET(AP) 'other AP. values are already set from the Get IF stat THEN 'though you should reassign AP.RecPtrOff/Seg if 'error 'you perform any BASIC string functions between 'BULLET calls (none were in this case). NOTE: QB/PDS will move strings '--by this I mean to set these again after a user-break/Ctrl-Break ' AP.RecPtrOff = VARPTR(RecBuff) in the QB/QBX environment !!! ' AP.RecPtrSeg = VARSEG(RecBuff) ' stat = BULLET(AP) 'do the update call See: DeleteRecordXBsrc UpdateXB ~DeleteRecordXBsrc Func: DeleteRecordXB Pack: AccessPack Func: 34/Mid-level DIM AP AS AccessPack AP.Func = %DeleteRecordXB 'defined in PBULLET.BI AP.Handle = datahandle 'handle of record to delete AP.RecNo = RecnoToDelete& 'to determine which record number any record stat = BULLET(AP) 'is, use one of the keyed access routines IF stat THEN 'error See: UndeleteRecordsrc (XB) ~UndeleteRecordsrc (XB) Func: UndeleteRecordXB Pack: AccessPack Func: 35/Mid-level DIM AP AS AccessPack AP.Func = %UndeleteRecordXB 'defined in PBULLET.BI AP.Handle = datahandle 'handle of record to undelete AP.RecNo = RecnoToUndelete& 'to determine which record number any record stat = BULLET(AP) 'is use one of the keyed access routines IF stat THEN 'error See: PackRecordsXBsrc ~PackRecordsXBsrc Func: PackRecordsXB Pack: AccessPack Func: 36/Mid-level DIM AP AS AccessPack AP.Func = %PackRecordsXB 'defined in PBULLET.BI AP.Handle = datahandle 'handle of data file to pack stat = BULLET(AP) IF stat THEN 'error See: FirstKeyXBsrc ~FirstKeyXBsrc Func: FirstKeyXB Pack: AccessPack Func: 40/Mid-level DIM AP AS AccessPack DIM TempStr AS STRING * 64 'used as fixed-length string for VARPTR() 'needs to be at least key length (MAXKEYLEN=64) AP.Func = %FirstKeyXB 'defined in PBULLET.BI AP.Handle = indexhandle 'handle to index file to access key from AP.KeyPtrOff = VARPTR(TempStr) 'far pointer to key buffer AP.KeyPtrSeg = VARSEG(TempStr) '(the first key is read from disk and put here) stat = BULLET(AP) IF stat = 0 THEN 'TempStr is filled in with the key (for as many bytes as the key length) 'so, if the key is a character type, you could PRINT LEFT$(TempStr,KeyLen) 'or, if key is a 32-bit LONG then PRINT CVL(LEFT$(TempStr,4)). When using 'the key returned, be aware that if UNIQUE was NOT specified then the 'enumerator word is attached to the end of the key (the right two bytes). 'Also, AP.RecNo is set to the record number of the first key in the index. 'error See: EqualKeyXBsrc ~EqualKeyXBsrc Func: EqualKeyXB Pack: AccessPack Func: 41/Mid-level DIM AP AS AccessPack DIM TempStr AS STRING * 64 'used as fixed-length string for VARPTR() 'passed key to find, start with enum=0 TempStr = FindKey$ + CHR$(0) + CHR$(0) + CHR$(0) '2 enumerator bytes + 0-term AP.Func = %EqualKeyXB 'defined in PBULLET.BI AP.Handle = indexfile 'handle to index file to find key from AP.KeyPtrOff = VARPTR(TempStr) 'far pointer to key buffer AP.KeyPtrSeg = VARSEG(TempStr) stat = BULLET(AP) IF stat = 0 THEN 'the key matched exactly (including enumerator, if present) 'TempStr is NOT ALTERED 'AP.RecNo is set to the record number of that key ELSEIF stat = 200 THEN AP.Func = %NextKeyXB 'if not found, get following key and check the stat = BULLET(AP) 'key proper (key less the enumerator bytes) IF stat = 0 THEN '(i.e., it may be proper key but enumerator=1) 'see NextKeyXBsrc for continuation See: NextKeyXBsrc ~NextKeyXBsrc Func: NextKeyXB Pack: AccessPack Func: 42/Mid-level 'see EqualKeyXBsrc for preliminary code AP.Func = %NextKeyXB 'defined in PBULLET.BI stat = BULLET(AP) 'KEYLEN assumed to equal actual key length... IF stat = 0 THEN '...as returned by StatKXB IF IndexFileIsNotUnique THEN IF LEFT$(TempStr,KeyLen-2) = FindKey$ THEN 'the next key matches! '(except for the enumerator) This code example follows up on the EqualKeyXBsrc example. See EqualKeyXB for more information on finding partial keys. See: PrevKeyXBsrc EqualKeyXBsrc ~PrevKeyXBsrc Func: PrevKeyXB Pack: AccessPack Func: 43/Mid-level DIM AP AS AccessPack DIM TempStr AS STRING * 64 'used as fixed-length string for VARPTR() 'assume code has already executed to locate a key--this code then gets the key 'before that one AP.Func = %PrevKeyXB 'defined in PBULLET.BI AP.Handle = indexfile 'handle to index file to access key from AP.KeyPtrOff = VARPTR(TempStr) 'far pointer to key buffer AP.KeyPtrSeg = VARSEG(TempStr) '(the prev key is read from disk and put here) stat = BULLET(AP) IF stat = 0 THEN 'TempStr is filled in with the key (for as many bytes as the key length). 'Also, AP.RecNo is set to the record number of the key. 'error See: LastKeyXBsrc ~LastKeyXBsrc Func: LastKeyXB Pack: AccessPack Func: 44/Mid-level DIM AP AS AccessPack DIM TempStr AS STRING * 64 'used as fixed-length string for VARPTR() AP.Func = %LastKeyXB 'defined in PBULLET.BI AP.Handle = indexfile 'handle to index file to access key from AP.KeyPtrOff = VARPTR(TempStr) 'far pointer to key buffer AP.KeyPtrSeg = VARSEG(TempStr) '(the last key is read from disk and put here) stat = BULLET(AP) IF stat = 0 THEN 'TempStr is filled in with the key (for as many bytes as the key length) 'AP.RecNo is set to the record number of the last key. 'error See: StoreKeyXBsrc ~StoreKeyXBsrc Func: StoreKeyXB Pack: AccessPack Func: 45/Mid-level DIM AP AS AccessPack DIM TempStr AS STRING * 64 'used as fixed-length string for VARPTR() 'Assume record has been added to data file (AddRecordXB, returning RecNo2Use) 'and key has been built (BuildKeyXB returning KeyToAdd$). TempStr = KeyToAdd$ + CHR$(0) 'add the key to a UNIQUE index file AP.Func = %StoreKeyXB 'defined in PBULLET.BI AP.Handle = indexfile 'handle to index file to insert key into AP.RecNo = RecNo2Use 'associate this record number with key AP.KeyPtrOff = VARPTR(TempStr) 'far pointer to key buffer AP.KeyPtrSeg = VARSEG(TempStr) '(the last key is read from disk and put here) stat = BULLET(AP) IF stat = 0 THEN 'key added ELSEIF stat = 201 THEN 'key already exists, which means you need to construct a unique enumerator-- 'provided the file wasn't created for UNIQUE keys...INSTEAD USE InsertXB! See: DeleteKeyXBsrc InsertXB ~DeleteKeyXBsrc Func: DeleteKeyXB Pack: AccessPack Func: 46/Mid-level DIM AP AS AccessPack DIM TempStr AS STRING * 64 'used as fixed-length string for VARPTR() TempStr = KeyToDelete$ + CHR$(0) 'delete the key from a UNIQUE index file '(else need to supply enumerator also) AP.Func = %DeleteKeyXB 'defined in PBULLET.BI AP.Handle = indexfile 'handle to index file of key to delete AP.KeyPtrOff = VARPTR(TempStr) 'far pointer to key buffer AP.KeyPtrSeg = VARSEG(TempStr) '(this key is searched for exactly) stat = BULLET(AP) 'if exact match found the key is deleted! IF stat = 0 THEN 'key deleted permanently ELSEIF stat = 200 THEN 'key as stated was not in the index file--if the index is not UNQIUE then 'you must supply the exact enumerator along with the key proper to delete '--you can use the CurrentKeyXB routine to obtain the exact current key 'other error See: BuildKeyXBsrc ~BuildKeyXBsrc Func: BuildKeyXB Pack: AccessPack Func: 47/Mid-level DIM AP AS AccessPack DIM TempStr AS STRING * 64 'used as fixed-length string for VARPTR() 'Assume record has been built and is ready to be added to the data file. The 'record is in the variable RecBuff (a fixed-length TYPE variable, typically). AP.Func = %BuildKeyXB 'defined in PBULLET.BI AP.Handle = indexfile 'handle to index file key is to be built for AP.RecPtrOff = VARPTR(RecBuff) 'far pointer to data record buffer AP.RecPtrSeg = VARSEG(RecBuff) AP.KeyPtrOff = VARPTR(TempStr) 'far pointer to key buffer AP.KeyPtrSeg = VARSEG(TempStr) '(the built key is put here) stat = BULLET(AP) IF stat = 0 THEN 'key built okay so can do a AddRecordXB followed by a StoreKeyXB 'but, again, InsertXB takes care of all this detail and then some--use it 'error See: CurrentKeyXBsrc ~CurrentKeyXBsrc Func: CurrentKeyXB Pack: AccessPack Func: 48/Mid-level DIM AP AS AccessPack DIM TempStr AS STRING * 64 'used as fixed-length string for VARPTR() AP.Func = %CurrentKeyXB 'defined in PBULLET.BI AP.Handle = indexfile 'handle to index file AP.KeyPtrOff = VARPTR(TempStr) 'far pointer to key buffer AP.KeyPtrSeg = VARSEG(TempStr) '(the current key is put here) stat = BULLET(AP) IF stat = 0 THEN 'TempStr set to current key (valid only for KeyLen bytes) 'Also, AP.RecNo is set to the record number of the key. 'error See: GetFirstXBsrc ~GetFirstXBsrc Func: GetFirstXB Pack: AccessPack Func: 60/High-level DIM AP AS AccessPack DIM TempStr AS STRING * 64 'used as fixed-length string for VARPTR() DIM RecBuff AS RecordTYPE 'see AddRecordXBsrc for record layout AP.Func = %GetFirstXB 'defined in PBULLET.BI AP.Handle = indexhandle 'handle to index file to access key from AP.RecPtrOff = VARPTR(RecBuff) 'far pointer to record buffer AP.RecPtrSeg = VARSEG(RecBuff) '(the record indexed by the key is put here) AP.KeyPtrOff = VARPTR(TempStr) 'far pointer to key buffer AP.KeyPtrSeg = VARSEG(TempStr) '(the first key is read from disk and put here) stat = BULLET(AP) IF stat = 0 THEN 'TempStr is filled in with the key (for as many bytes as the key length) 'RecBuff is filled in with the data record 'AP.RecNo is set to the record number of the first key in the index. 'error See: GetEqualXBsrc ~GetEqualXBsrc Func: GetEqualXB Pack: AccessPack Func: 61/High-level DIM AP AS AccessPack DIM TempStr AS STRING * 64 'used as fixed-length string for VARPTR() DIM RecBuff AS RecordTYPE 'see AddRecordXBsrc for record layout 'passed key to find, start with enum=0 TempStr = FindKey$ + CHR$(0) + CHR$(0) + CHR$(0) '2 enumerator bytes + 0-term AP.Func = %GetEqualXB 'defined in PBULLET.BI AP.Handle = indexfile 'handle to index file to find key from AP.RecPtrOff = VARPTR(RecBuff) 'far pointer to record buffer AP.RecPtrSeg = VARSEG(RecBuff) '(the record indexed by the key is put here) AP.KeyPtrOff = VARPTR(TempStr) 'far pointer to key buffer AP.KeyPtrSeg = VARSEG(TempStr) stat = BULLET(AP) IF stat = 0 THEN 'RecBuff, and AP.RecNo filled as expected (TempStr remains the same) ELSEIF stat = 200 THEN AP.Func = %GetNextXB 'if not found, can get following key--the next stat = BULLET(AP) 'key would logically follow the key not found '--this let's you search based on partial keys See: GetNextXBsrc ~GetNextXBsrc Func: GetNextXB Pack: AccessPack Func: 62/High-level DIM AP AS AccessPack DIM TempStr AS STRING * 64 'used as fixed-length string for VARPTR() DIM RecBuff AS RecordTYPE 'see AddRecordXBsrc for record layout AP.Func = %GetFirstXB 'defined in PBULLET.BI AP.Handle = indexhandle 'handle to index file to access key from AP.RecPtrOff = VARPTR(RecBuff) 'far pointer to record buffer AP.RecPtrSeg = VARSEG(RecBuff) '(the record indexed by the key is put here) AP.KeyPtrOff = VARPTR(TempStr) 'far pointer to key buffer AP.KeyPtrSeg = VARSEG(TempStr) '(the first key is read from disk and put here) stat = BULLET(AP) DO WHILE stat = 0 'print all records in key order PRINT AP.RecNo; RecBuff.code; ", "; RecBuff.bday AP.Func = %GetNextXB stat = BULLET(AP) IF stat <> 202 THEN 'error 202 means end of file (expected) 'error other than expected EOF See: GetPrevXBsrc ~GetPrevXBsrc Func: GetPrevXB Pack: AccessPack Func: 63/High-level DIM AP AS AccessPack DIM TempStr AS STRING * 64 'used as fixed-length string for VARPTR() DIM RecBuff AS RecordTYPE 'see AddRecordXBsrc for record layout AP.Func = %GetLastXB 'defined in PBULLET.BI AP.Handle = indexhandle 'handle to index file to access key from AP.RecPtrOff = VARPTR(RecBuff) 'far pointer to record buffer AP.RecPtrSeg = VARSEG(RecBuff) '(the record indexed by the key is put here) AP.KeyPtrOff = VARPTR(TempStr) 'far pointer to key buffer AP.KeyPtrSeg = VARSEG(TempStr) '(the last key is read from disk and put here) stat = BULLET(AP) DO WHILE stat = 0 'print all records in REVERSE key order PRINT AP.RecNo; RecBuff.code; ", "; RecBuff.bday AP.Func = %GetPrevXB stat = BULLET(AP) IF stat <> 203 THEN 'error 203 means top of file (expected) 'error other than expected TOF See: GetLastXBsrc ~GetLastXBsrc Func: GetLastXB Pack: AccessPack Func: 64/High-level DIM AP AS AccessPack DIM TempStr AS STRING * 64 'used as fixed-length string for VARPTR() DIM RecBuff AS RecordTYPE 'see AddRecordXBsrc for record layout AP.Func = %GetLastXB 'defined in PBULLET.BI AP.Handle = indexhandle 'handle to index file to access key from AP.RecPtrOff = VARPTR(RecBuff) 'far pointer to record buffer AP.RecPtrSeg = VARSEG(RecBuff) '(the record indexed by the key is put here) AP.KeyPtrOff = VARPTR(TempStr) 'far pointer to key buffer AP.KeyPtrSeg = VARSEG(TempStr) '(the last key is read from disk and put here) stat = BULLET(AP) IF stat = 0 THEN 'TempStr is filled in with the key (for as many bytes as the key length) 'RecBuff is filled in with the data record 'AP.RecNo is set to the record number of the last key in the index. 'error See: InsertXBsrc ~InsertXBsrc Func: InsertXB Pack: AccessPack Func: 65/High-level REDIM AP(1 TO 4) AS AccessPack 'array of 3 access packs, 1 for each index file 'TempStr and RecBuff previously defined FOR i = 1 TO 3 '3=number of related indexes to maintain AP(i).Func = %InsertXB AP(i).Handle = indexhandle(i) 'each index file's handle AP(i).RecPtrOff = VARPTR(RecBuff) AP(i).RecPtrSeg = VARSEG(RecBuff) see AddRecordXBsrc for RecBuff TYPE AP(i).KeyPtrOff = VARPTR(TempStr) (be sure you reserved the tag field) AP(i).KeyPtrSeg = VARSEG(TempStr) AP(i).NextPtrOff = VARPTR(AP(i + 1)) 'point to NEXT access pack AP(i).NextPtrSeg = VARSEG(AP(i + 1)) 'AP(1 TO 4) to avoid subscript error AP(3).NextPtrOff = 0 'reset last access pack to end-link value AP(3).NextPtrSeg = 0 stat = BULLET(AP(1)) IF stat = 0 THEN 'if stat=0 must still check AP(1).Stat IF AP(1).Stat <> 0 THEN 'error when adding data record TrueError = AP(stat).Stat 'the returned stat is array index of bad pack See: UpdateXBsrc ~UpdateXBsrc Func: UpdateXB Pack: AccessPack Func: 66/High-level REDIM AP(1 TO 4) AS AccessPack 'array of 3 access packs, 1 for each index file 'TempStr and RecBuff previously defined FOR i = 1 TO 3 '3=number of related indexes to maintain AP(i).Func = %UpdateXB AP(i).Handle = indexhandle(i) 'each index file's handle AP(i).RecNo = RecordNumberToUpdate& 'tell it which record to update AP(i).RecPtrOff = VARPTR(RecBuff) 'RecBuff has new, updated data AP(i).RecPtrSeg = VARSEG(RecBuff) AP(i).KeyPtrOff = VARPTR(TempStr) AP(i).KeyPtrSeg = VARSEG(TempStr) AP(i).NextPtrOff = VARPTR(AP(i + 1)) 'point to NEXT access pack AP(i).NextPtrSeg = VARSEG(AP(i + 1)) 'AP(1 TO 4) to avoid subscript error AP(3).NextPtrOff = 0 : AP(3).NextPtrSeg = 0 'reset last access pack to 0 stat = BULLET(AP(1)) IF stat = 0 THEN 'if stat=0 must still check AP(1).Stat IF AP(1).Stat <> 0 THEN 'error when writing data record TrueError = AP(stat).Stat 'the returned stat is array index of bad pack See: ReindexXBsrc ~ReindexXBsrc Func: ReindexXB Pack: AccessPack Func: 67/High-level REDIM AP(1 TO 4) AS AccessPack 'array of 3 access packs, 1 for each index file 'DIM'd to 4 to avoid bad subscript in loop FOR i = 1 TO 3 '3=number of related indexes to reindex AP(i).Func = %ReindexXB AP(i).Handle = indexhandle(i) 'each index file's handle AP(i).NextPtrOff = VARPTR(AP(i + 1)) 'point to NEXT access pack AP(i).NextPtrSeg = VARSEG(AP(i + 1)) 'AP(1 TO 4) to avoid subscript error AP(3).NextPtrOff = 0 AP(3).NextPtrSeg = 0 'reset last access pack to end-link value stat = BULLET(AP(1)) IF stat THEN 'if stat <> 0 then the... TrueError = AP(stat).Stat '...returned stat is array index of bad pack The reason AP() is REDIM AP(1 TO 4) is so that the AP(i + 1) in the code loop doesn't create an invalid subscript error. See: LockXBsrc ~LockXBsrc Func: LockXB Pack: AccessPack Func: 80/Network REDIM AP(1 TO 4) AS AccessPack 'array of 3 access packs, 1 for each index file 'DIM'd to 4 to avoid bad subscript in loop FOR i = 1 TO 3 '3=number of related indexes to Lock AP(i).Func = %LockXB AP(i).Handle = indexhandle(i) 'each index file's handle AP(i).NextPtrOff = VARPTR(AP(i + 1)) 'point to NEXT access pack AP(i).NextPtrSeg = VARSEG(AP(i + 1)) NEXT 'data file handle known internally by BULLET AP(3).NextPtrOff = 0 AP(3).NextPtrSeg = 0 'reset last access pack to end-link value stat = BULLET(AP(1)) IF stat > 3 THEN 'if stat > 3 (> number of packs) then the... TrueError = AP(3).Stat '...lock failed on the data file ELSEIF stat <> 0 THEN (|last ReadKHXB) TrueError = AP(stat).Stat '...lock failed on index file # stat The Lock routines use a different method to identify the bad pack when the failure was caused by the data file. See above. See: UnlockXBsrc ~UnlockXBsrc Func: UnlockXB Pack: AccessPack Func: 81/Network REDIM AP(1 TO 4) AS AccessPack 'array of 3 access packs, 1 for each index file 'DIM'd to 4 to avoid bad subscript in loop FOR i = 1 TO 3 '3=number of related indexes to Lock AP(i).Func = %UnlockXB AP(i).Handle = indexhandle(i) 'each index file's handle AP(i).NextPtrOff = VARPTR(AP(i + 1)) 'point to NEXT access pack AP(i).NextPtrSeg = VARSEG(AP(i + 1)) NEXT 'data file handle known internally by BULLET AP(3).NextPtrOff = 0 AP(3).NextPtrSeg = 0 'reset last access pack to end-link value stat = BULLET(AP(1)) IF stat > 3 THEN 'if stat > 3 (> number of packs) then the... TrueError = AP(3).Stat '...unlock failed on the data file ELSEIF stat <> 0 THEN TrueError = AP(stat).Stat '...unlock failed on index file # stat The Lock routines use a different method to identify the bad pack when the failure was caused by the data file. See above. See: LockKeyXBsrc ~LockKeyXBsrc Func: LockKeyXB Pack: AccessPack Func: 82/Network REDIM AP(1 TO 4) AS AccessPack 'array of 3 access packs, 1 for each index file 'DIM'd to 4 to avoid bad subscript in loop FOR i = 1 TO 3 '3=number of related indexes to Lock AP(i).Func = %LockXB AP(i).Handle = indexhandle(i) 'each index file's handle AP(i).NextPtrOff = VARPTR(AP(i + 1)) 'point to NEXT access pack AP(i).NextPtrSeg = VARSEG(AP(i + 1)) AP(3).NextPtrOff = 0 AP(3).NextPtrSeg = 0 'reset last access pack to end-link value stat = BULLET(AP(1)) IF stat <> 0 THEN TrueError = AP(stat).Stat 'lock failed on index file # stat '--if stat > 3 then failed on last internal '--ReadKHXB...This is EXTREMELY unlikely See: UnlockKeyXBsrc ~UnlockKeyXBsrc Func: UnlockKeyXB Pack: AccessPack Func: 83/Network REDIM AP(1 TO 4) AS AccessPack 'array of 3 access packs, 1 for each index file 'DIM'd to 4 to avoid bad subscript in loop FOR i = 1 TO 3 '3=number of related indexes to Lock AP(i).Func = %UnlockKeyXB AP(i).Handle = indexhandle(i) 'each index file's handle AP(i).NextPtrOff = VARPTR(AP(i + 1)) 'point to NEXT access pack AP(i).NextPtrSeg = VARSEG(AP(i + 1)) AP(3).NextPtrOff = 0 AP(3).NextPtrSeg = 0 'reset last access pack to end-link value stat = BULLET(AP(1)) IF stat <> 0 THEN TrueError = AP(stat).Stat 'unlock failed on index file # stat See: LockDataXBsrc ~LockDataXBsrc Func: LockDataXB Pack: AccessPack Func: 84/Network DIM AP AS AccessPack AP.Func = %LockDataXB 'defined in PBULLET.BI AP.Handle = datahandle 'handle of data file to lock AP.RecNo = 0& '=0 to lock all or, set to actual record number stat = BULLET(AP) ' to lock as in AP.RecNo = lockThisRec& IF stat THEN 'error See: UnlockDataXBsrc ~UnlockDataXBsrc Func: UnlockDataXB Pack: AccessPack Func: 85/Network DIM AP AS AccessPack AP.Func = %UnlockDataXB 'defined in PBULLET.BI AP.Handle = datahandle 'handle of data file to unlock AP.RecNo = 0& '=0 to unlock all or, set to actual record num stat = BULLET(AP) ' to unlock as in AP.RecNo = lockThisRec& IF stat THEN 'error 'note: you cannot unlock parts of a file with '1 single unlock (where AP.RecNo=0). Instead, 'you must unlock each record individually-- 'that is, if you made any single-record locks See: DriveRemoteXBsrc ~DriveRemoteXBsrc Func: DriveRemoteXB Pack: RemotePack Func: 86/Network DIM RP AS RemotePack RP.Func = %DriveRemoteXB 'defined in PBULLET.BI RP.Handle = drive2check 'drive to check (0=default, 1=A:,2=B:,3=C:...) stat = BULLET(RP) IF stat = 0 THEN 'RP.IsRemote set to 0 if drive local, 1 if remote 'RP.Flags set to DX register as returned by DOS 'RP.IsShare set to 0 if SHARE.EXE is not loaded, non-zero SHARE installed 'error (like invalid drive) See: FileRemoteXBsrc ~FileRemoteXBsrc Func: FileRemoteXB Pack: RemotePack Func: 87/Network DIM RP AS RemotePack RP.Func = %FileRemoteXB 'defined in PBULLET.BI RP.Handle = filehandle 'file handle to check stat = BULLET(RP) IF stat = 0 THEN 'RP.IsRemote set to 0 if file local, 1 if remote 'RP.Flags set to DX register as returned by DOS 'RP.IsShare set to 0 if SHARE.EXE is not loaded, non-zero SHARE installed 'error (like invalid handle) See: SetRetriesXBsrc ~SetRetriesXBsrc Func: SetRetriesXB Pack: SetRetriesPack Func: 88/Network DIM SRP AS SetRetriesPack SRP.Func = %SetRetriesXB SRP.Mode = 1 '1=set to user values, 0=set DOS default SRP.Pause = 5000 'do 5,000 loops between retries SRP.Retries = 5 'try 5 times before giving up with error stat = BULLET(SRP) IF stat THEN 'error 'it's unlikely an error occurs See: DeleteFileDOSsrc ~DeleteFileDOSsrc Func: DeleteFileDOS Pack: DOSFilePack Func: 100/DOS DIM DFP AS DOSFilePack DIM TempStr AS STRING * 80 TempStr = file2delete$ + CHR$(0) DFP.Func = %DeleteFileDOS 'defined in PBULLET.BI DFP.FilenamePtrOff = VARPTR(TempStr) DFP.FilenamePtrSeg = VARSEG(TempStr) stat = BULLET(DFP) IF stat THEN 'error See: RenameFileDOSsrc ~RenameFileDOSsrc Func: RenameFileDOS Pack: DOSFilePack Func: 101/DOS DIM DFP AS DOSFilePack DIM TempStr AS STRING * 80 DIM TempStr2 AS STRING * 80 TempStr = file2rename$ + CHR$(0) TempStr2 = newfilename$ + CHR$(0) DFP.Func = %RenameFileDOS 'defined in PBULLET.BI DFP.FilenamePtrOff = VARPTR(TempStr) DFP.FilenamePtrSeg = VARSEG(TempStr) DFP.NewNamePtrOff = VARPTR(TempStr2) DFP.NewNamePtrSeg = VARSEG(TempStr2) stat = BULLET(DFP) IF stat THEN 'error See: CreateFileDOSsrc ~CreateFileDOSsrc Func: CreateFileDOS Pack: DOSFilePack Func: 102/DOS DIM DFP AS DOSFilePack DIM TempStr AS STRING * 80 TempStr = file2create$ + CHR$(0) DFP.Func = %CreateFileDOS 'defined in PBULLET.BI DFP.FilenamePtrOff = VARPTR(TempStr) DFP.FilenamePtrSeg = VARSEG(TempStr) DFP.Attr = 0 'normal file directory attribute stat = BULLET(DFP) IF stat THEN 'error See: AccessFileDOSsrc ~AccessFileDOSsrc Func: AccessFileDOS Pack: DOSFilePack Func: 103/DOS DIM DFP AS DOSFilePack DIM TempStr AS STRING * 80 TempStr = file2access$ + CHR$(0) DFP.Func = %AccessFileDOS 'defined in PBULLET.BI DFP.FilenamePtrOff = VARPTR(TempStr) DFP.FilenamePtrSeg = VARSEG(TempStr) DFP.ASmode = &H42 'attempt R/W DENY NONE access stat = BULLET(DFP) IF stat THEN 'error See: OpenFileDOSsrc ~OpenFileDOSsrc Func: OpenFileDOS Pack: DOSFilePack Func: 104/DOS DIM DFP AS DOSFilePack DIM TempStr AS STRING * 80 TempStr = file2open$ + CHR$(0) DFP.Func = %OpenFileDOS 'defined in PBULLET.BI DFP.FilenamePtrOff = VARPTR(TempStr) DFP.FilenamePtrSeg = VARSEG(TempStr) DFP.ASmode = &H42 'open in R/W DENY NONE access stat = BULLET(DFP) IF stat = 0 THEN 'DFP.Handle set to handle of open file 'error See: SeekFileDOSsrc ~SeekFileDOSsrc Func: SeekFileDOS Pack: DOSFilePack Func: 105/DOS DIM DFP AS DOSFilePack DFP.Func = %SeekFileDOS 'defined in PBULLET.BI DFP.Handle = handle DFP.SeekOffset = 0& 'position 0 relative EOF (get length of file) DFP.Method = 2 'seek from END of file stat = BULLET(DFP) IF stat = 0 THEN 'DFP.SeekOffset set to absolute current offset '--in this case, the DFP.SeekOffset equals then length of the file 'error See: ReadFileDOSsrc ~ReadFileDOSsrc Func: ReadFileDOS Pack: DOSFilePack Func: 106/DOS DIM DFP AS DOSFilePack DIM ReadBuff AS STRING * 512 'sector buffer DFP.Func = %ReadFileDOS 'defined in PBULLET.BI DFP.Handle = handle DFP.Bytes = Bytes2Read '16-bit value, in this case 512 since that's DFP.BufferPtrOff = VARPTR(ReadBuff) 'the size of ReadBuff DFP.BufferPtrSeg = VARSEG(ReadBuff) stat = BULLET(DFP) IF stat = 0 THEN IF DFP.Bytes <> Bytes2Read THEN 'check if EOF processed 'hit EOF before reading all 512 bytes ELSE 'ReadBuff filled with 512 bytes of data read from the current disk pos 'disk position moved to the last byte read + 1 ENDIF 'error See: ExpandFileDOSsrc ~ExpandFileDOSsrc Func: ExpandFileDOS Pack: DOSFilePack Func: 107/DOS DIM DFP AS DOSFilePack DFP.Func = % ExpandFileDOS 'defined in PBULLET.BI DFP.Handle = handle DFP.SeekOffset = Bytes2ExpandBy& stat = BULLET(DFP) IF stat = 0 THEN 'file expanded by number of bytes specified 'error See: WriteFileDOSsrc ~WriteFileDOSsrc Func: WriteFileDOS Pack: DOSFilePack Func: 108/DOS DIM DFP AS DOSFilePack DIM WriteBuff AS STRING * 512 'sector buffer DFP.Func = %WriteFileDOS 'defined in PBULLET.BI DFP.Handle = handle DFP.Bytes = Bytes2Write '16-bit value, in this case 512 since that's DFP.BufferPtrOff = VARPTR(WriteBuff) 'the size of WriteBuff DFP.BufferPtrSeg = VARSEG(WriteBuff) stat = BULLET(DFP) IF stat = -2 THEN 'disk full ELSE IF stat THEN 'other error 'okay Unlike ReadFileDOS, if the number of bytes actually written does not equal Bytes2Write, the WriteFileDOS routine returns a DISK FULL error code (-2). See: CloseFileDOSsrc ~CloseFileDOSsrc Func: CloseFileDOS Pack: DOSFilePack Func: 109/DOS DIM DFP AS DOSFilePack DFP.Func = %CloseFileDOS 'defined in PBULLET.BI DFP.Handle =handle2close stat = BULLET(DFP) IF stat THEN 'error See: MakeDirDOSsrc ~MakeDirDOSsrc Func: MakeDirDOS Pack: DOSFilePack Func: 110/DOS DIM DFP AS DOSFilePack DIM TempStr AS STRING * 80 TempStr = subdir2make$ + CHR$(0) DFP.Func = %MakeDirDOS 'defined in PBULLET.BI DFP.FilenamePtrOff = VARPTR(TempStr) DFP.FilenamePtrSeg = VARSEG(TempStr) stat = BULLET(DFP) IF stat THEN 'error See: DeleteFileDOSsrc