Black Box 4

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

/ Black Box 4 / BlackBox.cdr / clipper / clbbs106.arj / CLBBS106.TXT

Wrap

Text File | 1991-03-04 | 86.2 KB | 1,983 lines

░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░ ░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░ ░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░ ░░░░░░░▓▓▓▓░░▓▓▓▓░░░░░▓▓▓▓░░░▓▓▓▓▓▓░░▓▓▓▓▓▓░░▓▓▓▓▓▓░░░▓▓▓▓░░░░░░░░░░░ ░░░░░░░░▓▓ ▓▓░░▓▓ ░░░░░▓▓ ░░░▓▓ ▓▓░░▓▓ ▓▓░░▓▓ ▓▓░▓▓ ▓▓░░░░░░░░░░░░ ░░░░░░░░▓▓ ░░ ░▓▓ ░░░░░░▓▓ ░░░░▓▓ ░▓▓ ░▓▓ ░▓▓ ░▓▓ ░▓▓ ▓▓▓░░ ░░░░░░░░░░░░ ░░░░░░░░░▓▓ ░░░░░░▓▓ ░░░░░░▓▓ ░░░░▓▓▓▓▓ ░▓▓▓▓▓ ░▓▓▓▓▓ ░▓▓▓░░░░░░░░░░░░░░░░ ░░░░░░░░░▓▓ ░░░░░░▓▓ ░░▓░░░▓▓ ░░░░▓▓ ░░▓▓ ▓▓░░▓▓ ▓▓░░░ ▓▓▓░░░░░░░░░░░░░░ ░░░░░░░░░▓▓░░▓▓░░▓▓ ░▓▓ ░░▓▓ ░░░░▓▓ ░░░░░▓▓ ░▓▓ ░▓▓ ░▓▓ ▓▓░░▓▓ ░░░░░░░░░░░░ ░░░░░░░░░▓▓▓▓ ▓▓▓▓▓▓▓ ░▓▓▓▓░░░▓▓▓▓░░░░▓▓▓▓▓▓ ▓▓▓▓▓▓ ░▓▓▓▓ ░░░░░░░░░░░ ░░░░░░░░ ░░ ░░ ░░░ ░░░░ ░░ ░░░ ░░░░░░░░░░ ░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░ ░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░ ░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░ Volume 1, Number 6 4 March 1991 (c) Daniel Doçekal, All Rights Reserved The BBS Clipper magazine, published PERIODICALLY, later WEEKLY...... Some of the material used comes from scanning CLIPPER echoes which are carried in various BBS throughout the World. These Echoes are very often the source of the most often asked Questions and Answers about Clipper. Other material, which is fully signed or abbreviated is the copyright of the appropriate persons. The publisher is not responsible for other authors submissions.... Published material is not necessarily the opinion of the publisher. Redaction: Publisher...................................Daniel Docekal Chief editor ...............................Daniel Docekal Language editor .................................Dave Wall Table of Contents 1. ARTICLES ............................................................. 1 DOSRCS - Revision Control System - part (4) .......................... 1 The Programmer's Guide to CLIPPER Linkers - part (3) ................. 5 Best way of making FOR cyclus, and what about WHILE cyclus ........... 10 WHICH packing/unpacking program to choose??????? ..................... 13 Will Clipper expand to graphics environment? ......................... 16 2. SOFTWARE ............................................................. 18 TIPS&TRICKS - how to found largest string in array ................... 18 DANIEL.LIB - NEW version of DANIEL.RMK ............................... 18 3. ANOMALIES ............................................................ 22 ANOMALIES reports and commets ........................................ 22 Clipper Debugger anomalies ........................................... 22 INCONSISTENCY ........................................................ 24 New RTLINK version, INCREMENTAL LINKING .............................. 24 4. ECHODUMP ............................................................. 26 CONFERENCE DUMP part (5) ............................................. 26 5. CLIPPER NET .......................................................... 28 Index of described files in Clipper BBS Magazine ..................... 28 CLIPBBS 1-06 Table of Contents (...) 4 Mar 1991 Routing-Plan for ClipperNet .......................................... 28 CLN - SOUND.ARJ ...................................................... 30 CLN - TBWHL4.ARJ ..................................................... 31 ClipNet - NFDESC2.ARJ ................................................ 31 ClipNet - NFLIB2.ARJ ................................................. 32 ClipNet - NFSRC2.ARJ ................................................. 32 ClipperNet - ACCESS.ARJ .............................................. 34 ClipperNet - ACHOO2.ARJ .............................................. 34 6. CLIPBBS .............................................................. 36 CLIPBBS distribution ................................................. 36 CLIPBBS, how to write an article!!! .................................. 37 - - - - - CLIPBBS 1-06 Page 1 4 Mar 1991 ============================================================================= ARTICLES ============================================================================= DOSRCS - Revision Control System - part (4) _________________ Check In - CI.EXE ═════════════════ CI.EXE is second most often used program from DOSRCS set of programs. It's used in simply form in our RCSPUT.BAT: │ @echo off │ echo Processing [%1] │ CI -q %1 One switch used in this case, -q, is used only for supressing some messages and questions given by CI.EXE. Complete syntax is: CI [options] file ... One important warning: Every RCS used and stored file is using user names for specifying who is able to access file and not. Originally every file is getting name of user who was storing first this file and this user has right to use this file. Other users not, until they are written inside of file as other possible users. It's done by RCS program, or by making access list empty. Superuser can of course access all files without limitation. ________ BRANCHES Branches are created automatically in most cases. Typicall example is this: TEST.PRG has initial revison 1.1 and next revisions 1.2 and 1.3 Owner of file will retrieve 1.1 from RCS and start changing (1.1 is locked!). After this will use CheckIn without specifying new number of revision. CheckIn will see that 1.1 is locked and therefore will assume that new revision number MUST be 1.1.1.1 (added 1.1 after 1.1) and will store it with this number. User then can continue in BRANCHING with number 1.1.1.2, 1.1.1.3... ________________________ UNLOCKING of LOCKED FILE CLIPBBS 1-06 Page 2 4 Mar 1991 Sometime is needed to unlock locked revision (user can for example delete .PRG file and forget to use PUT for storing it back!). When this will happend, nobody is able to get revision for locking again and must be used RCS.EXE for unlocking. See some next continuing about RCS.EXE _____________________________________ CHECKING of differences and -Q switch Normally, CI checks whether the revision to be deposited is different from the preceding one. If it is not different, CI either aboirts the deposit (if -Q is given) or asks whether to abort (if -Q is omitted). Deposit can be forced with -F option. ____________ LOG MESSAGES Log message is used for every deposited revision. CI is asking for LOG message every time when new revision (or forced revision) is coming. User/programmer can use as many lines wants and end them with "." or Ctrl Z. All lines are then part of $Log$ keyword. If CI is used for more files in one (CI *.PRG for example), is possible use the same message for all this files - CI is asking for reuse old message, but it's not working in case how i'm using CI, because PUT is calling RCSPUT with only one file at one time. But mostly changes in sources are different for everyone different. ___________________ INCLUDING NEW FILES New files of RCS are created when RCS file of give source/text is not exist. RCS is then creating new RCS file with initial revision (1.1 by default), access list is initialized to empty and CI is asking for DESCRIPTIVE TEXT instead of LOG message. This text is description of purpose of file and is NOT part of $Log$ definition. __________ CI options -r[<rev>] assigns revision number <rev> to checked in revision, releases corresponding lock and deletes working file. This -r option is automatically given if not specified. Non specified <rev> is request for automatic counting of new revision number. <rev> if specified must be higher than the latest one on the branch to which rev belongs, or must start a new branch. -f[<rev>] CLIPBBS 1-06 Page 3 4 Mar 1991 Forces a deposit. New revision is deposited even it is not different from preceding one. -k[<rev>] Searches working file for keyword values to determine its revision number, creating date, state and author (see article about KEYWORDS) and assigns these values to deposited revision rather than computing them. ║ This is quite nice switch with main interest in sending a ║ complete latest revision to another place (developers, ║ check out sites, beta test sites). Then receiver will use ║ CI -k *.* on all files and if correct keywords are written ║ in EVERY file (no problem when used one kind of header), CI ║ will take all correct number from files and will create ║ correct stored new revision in another set of RCS files on ║ another place! -l[<rev>] Works like -r, only it perform addition CO -L for just deposited revision. Therefore, deposited revision is immediately checked out again and locked. ║ This one is usefull in case that one wants to save ║ revision, but continue in workign and editing. -u[<rev>] Works like -l, except that deposited revision is NOT locked. ║ This one is useful in case that one wants to (for example) ║ copile revision immediately after checking, but is not ║ planning another editing. -q[<rev>] Diagnostics output is not printed, revision which is not different from preceding one is NOT deposited, unless -f is given. -d<date> Uses <date> for the checkin date and time. Date may be specified in free format as was explained in article about CO.EXE program. ║ Useful in case of need to say different date of creating ║ revision than reality, or for "-k" option where date is not ║ available. CLIPBBS 1-06 Page 4 4 Mar 1991 -m<msg> Uses <msg> as LOG message for ALL revisions checked in. If this switch is not used, CI will ask for filling in all LOG messages. -n<name> Assigns the symbolic name <name> to the number of the checked-in revision. CI printer an error if name is already assigned to another number. Later is possible to use this symbolic name instead of revision number in CO command. This is giving nice possibility to NAMED some from revisions which have key value. -N<name> Same as -n, but it's overriding previsou assignment of name. RCS is of course still keeping number style of revisions, only assigns additional name, therefore name can be reasigned to another revision. Please NOTE THAT this is "-N" and previous one was "-n". -s<state> Sets the state of the checked-in revision to the identifier <state>. Default is <Exp>. Can be used for division of revisions or writing down some special flags or comments. -t[<txtfile>] Writes DESCRIPTIVE text into the RCS file (deleted the existing text of course). If <txtfile> is omitted, CI prompts the user for text, otherwise descriptive text is copied from file <txtfile>. -w<login> Uses <login> for Author field of deposited revision. Can be used for change of author or for set author for "-k" option All flags are described, interesting are marked with double line, together with tip of use. See you later... .DD. ----------------------------------------------------------------------------- CLIPBBS 1-06 Page 5 4 Mar 1991 The Programmer's Guide to CLIPPER Linkers - part (3) - OVERLAY RELOADING WITH RTLINK/CLIPPER-5.0 - This is a very important "undocumented" feature of the RTLINK linker supplied with Clipper-5.0. I have been RELOADING objects from the Summer 87 Clipper libraries for years, so when I received my copy of Clipper-5.0, the first thing I attempted to do was to overlay some of the larger modules in the Clipper libraries using the RELOAD command in my linker script file. The Nantucket development team had the foresight to insure that their C and ASM compiler assigned a UNIQUE name to each Clipper C/ASM object in the Clipper libraries, thus allowing, you, the Clipper programmer to use both the "undocumented" RELOAD command and MODULE command in your .LNK script files. Refer to the script file on the next page for an example of how to use these features to reduce memory usage in your Clipper-5.0 applications. This link file creates an overlay area for overlaying the larger modules in the Clipper libraries which are not likely to call each other recursively, therefore you will probably notice very difference in speed performance yet you will get up to 40K more memory overhead depending on how much of the Clipper libraries your application uses. You may get a "warning" message during link time if your application does not call one of the modules referenced in an overlay area. In the event this happens, simply remove that module from the link file. For example, if you are not using any TBROWSE objects, remove the following line: SECTION MODULE D:\S89\EDIT\TBROWSE.C # Example of using RELOADABLE overlays with the # Clipper-5.0 version of RTLINK FI <my files> LIB <my libs> LIB \CLIPPER5\LIB\extend LIB \CLIPPER5\LIB\clipper LIB \CLIPPER5\LIB\terminal LIB \CLIPPER5\LIB\dbfntx OUTPUT <my .EXE> VERBOSE RELOAD FAR 200 BEGINAREA SECTION MODULE D:\S89\EDIT\MEMOEDIT.C MODULE D:\S89\MEMO\MEMOTRAN.C SECTION MODULE D:\S89\MEMO\MEMOREAD.C MODULE D:\S89\MEMO\MEMOWRIT.C MODULE D:\S89\MEMO\MEMOLINE.C MODULE D:\S89\MEMO\MLCOUNT.C MODULE D:\S89\MEMO\MLPOS.C SECTION MODULE D:\S89\EDIT\TBROWSE.C CLIPBBS 1-06 Page 6 4 Mar 1991 SECTION MODULE D:\S89\EDIT\ACHOICE.C MODULE D:\S89\ARRAY\ASCAN.C MODULE D:\S89\ARRAY\ASORT.C MODULE D:\S89\ARRAY\ARRAY.C MODULE D:\S89\ARRAY\DIRECTRY.C SECTION MODULE D:\S89\DBCMD\DBSTRUCT.C MODULE D:\S89\DBCMD\DBCREATE.C MODULE D:\S89\DBCMD\JOINLIST.C MODULE D:\S89\DBF\SORTOF.C SECTION MODULE D:\S89\TERM\GETHELP.C ENDAREA = DYNAMIC OVERLAYS = "Dynamic Overlays" are segments of code which are loaded into a "memory pool" at runtime. Dynamic overlay segments usually look like this in your link file: # only 1 area needed BEGINAREA SECTION FILE A SECTION FILE B SECTION FILE C SECTION FILE D SECTION FILE E SECTION FILE F ENDAREA In the above example, if a function in FILE A calls a function in FILE B, then FILE B is loaded into the memory pool along with FILE A and the memory pool is "dynamically" managed to insure that modules which call each other repetitively will remain in memory to improve speed performance. Some dynamic overlay linkers also provide automatic "reloading" to insure that the application will always have sufficient memory in the pool. Modules which don't need to be in the memory pool will be removed to allow room for others at runtime. Dynamic overlay linkers (ALINK, BLINKER, WARPLINK, RTLINK- CLIPPER 5.0) offer the advantage of simpler overlay design and usually better memory management. Some also have a tendency to "fragment" the Clipper free memory pool because they share the heap with the Clipper application rather than allocating a separate dos memory area. This can cause applications to run out of memory after prolonged running of the application if there is not sufficient memory overhead at the start of the application. Some dynamic overlay linkers also support overlay management by "frequency of usage". If a module is called often during the running of an application, it will remain in the overlay pool rather than being "swapped out" to load other overlays. This insures optimum runtime performance. Dynamic-overlay linkers such as WARPLINK and BLINKER both CLIPBBS 1-06 Page 7 4 Mar 1991 make a disclaimer in their documentation requiring that the objects being overlayed are "well-behaved". After much experimentation with our own libraries and many third- party libraries, it is now more clear what constitutes well-behaved modules: a. Well-behaved code uses the Clipper EXTEND interface, registers or the stack for parameter passing and does not use undocumented features of Clipper. b. All Clipper-compiled .OBJects are dynamically- overlayable. Modules which CANNOT be dynamically-overlaid are: a. Routines which handle interrupts. b. Modules in which the DATA area is changed during runtime rather than allocating memory in the root area. Dynamic overlaying of third-party libraries can be an aggravating learning experience if you don't get support from the third party vendor in trying to accomplish this task. From my experience, the entire concept of dynamic-overlaying of C/ASM code is not very well-defined. BLINKER and WARPLINK have chosen to accept this challenge while the more conservative RTLINK has chosen instead to dynamic-overlay Clipper-compiled objects only and handle all the C/ASM code using static or reloable overlays - which, in my opinion, is a much more reliable concept. In the long run, I believe that dynamic-overlay linkers will survive only if they work flawlessly with all third-party libraries or very clearly define the rules for compatability with each third-party product and develop a following of add-on support. - PROCEDURE-LEVEL OVERLAY LOADING - For a dynamic-overlay manager to work effectively, it is important that it load overlays at the smallest possible code- segment level, i.e., the procedure/function rather than the entire object. This requires managing the symbol table seperately from the code and data, therefore, these linkers place the symbol table into the root memory area and each function into a separate overlay segment. To keep the root memory symbol table as small as possible it is recommended that you compile your application into a few large objects rather than many small objects when using dynamic-overlay linkers. I am not exactly sure at this time how RTLINK/CLIPPER-5.0 accomplishes the symbol management, but compiling into large objects does not appear to be necessary with Clipper-5.0 to reduce the executable memory model. RTLINK/CLIPPER-5.0 does not appear to be dependent on the method that your source code CLIPBBS 1-06 Page 8 4 Mar 1991 is compiled with the exception of line numbers, which WILL increase your memory usage. = DYNAMIC LINKING = True "Dynamic linkers" should not be confused with "Dynamic Overlay" linkers. Dynamic overlay linkers perform all the linking and fix-ups for the entire application at one time and then produce an .EXEcutable program. True "dynamic linkers" perform all the linking and fix-ups at "runtime" by linking from your .OBJ files or "dynamic-libraries" during the running of the application. Dynamic linkers (dCLIP, dCLIPRUN, NOLINK) provide the best memory-management for very large applications. All other forms of linkers still require that each symbol reference in the application be resolved at link time and assigned an address in the "root" memory area. Even the most sophisticated of overlay management schemes cannot eliminate the need to create this "symbol table", therefore the memory usage of a program will always increase proportionally with the number of symbols in the .OBJects regardless of the overlay technique. Dynamic linkers solve this problem by adding to the "root- memory" symbol table only if a function or procedure is called at runtime. Dynamic linkers will load the executable portion of the code into an "object pool" and perform the symbol fixups in the main heap area. Once a symbol has been added to the heap, it will stay in memory even after returning from the called function, however the executable code may be removed from the object pool to make space for new objects which may be called by the application. Dynamic linkers also allow linking from sets of dynamic libraries during the running of an application therefore large applications such as accounting systems can be distributed with one .EXE (the "dynamic-link engine") and multiple "dynamic- libraries". There are at least two (2) pre-requisites for a language to be able to support true dynamic linking: 1. The compiled code must be fully relocatable and not contain any code or data which must remain at a fixed memory location. 2. The error-handling system must allow for reconciling "missing externals" at runtime. Clipper SUMMER-87 and 5.0 compile code into P-code (or Pseudo- code) objects rather than native code. This code is "interpreted" rather than "executed" at runtime, therefore it is fully relocatable. In addition, any called function or procedure which is not in memory or in an overlay will cause a runtime error CLIPBBS 1-06 Page 9 4 Mar 1991 which is passed through the UNDEF_ERROR() function (Summer 87) or the ErrorSys() code block (Clipper-5.0). dCLIP and NOLINK take advantage of this feature of Clipper by "dynamically-linking" the missing code from .OBJect files on the disk or from dynamic libraries to resolve the "missing external" error then returning a .TRUE. value from the error handler to continue program execution. The error is completely transparent to the operator because it is resolved by the dynamic-link system as follows: * Clipper's ERRORSYS.PRG module FUNCTION undef_error PARAM _proc,_line,_info IF SUBSTR(_info,1,16)="missing EXTERNAL" * get name of .OBJect from .MAP file or .DLB dynamic lib. _obj=DC_OBJFIND(_proc) * link object into memory and verify proc is loaded IF DC_OBJLOAD(_obj) .AND. TYPE(_proc+'()')='UI' RETURN .t. ENDIF ENDIF DO _disperror WITH _proc,_line,_info BREAK Some Dynamic-Linkers also provide a "dynamic-link library", so you can actually link a "linker" into your application, thereby taking advantage of multiple link technologies in one application. dCLIP.EXE is an example of an Clipper application which uses three (3) overlaying techniques all at the same time. DCLIP.EXE is built using the dCLIP "dynamic-link" library, the Clipper Libraries, some Clipper-compile objects and RTLINK. Much of the C/ASM code in the Clipper libraries and the DCLIPNL.LIB "dynamic-link" library are overlayed using the "RELOADABLE" overlay capability of RTLINK. The Clipper-compiled part of dCLIP is "DYNAMICALLY" overlayed with the RTLINK dynamic overlay manager, and the user "application" code is "DYNAMICALLY- LINKED" at runtime by the dCLIP dynamic link functions. This dynamic-link technology is easily included in any Clipper application by simply defining the DCLIPNL.LIB ahead of CLIPPER.LIB in your application link statement, then using a function call in your application to reference the object maps or dynamic libraries to use during program execution. ----------------------------------------------------------------------------- CLIPBBS 1-06 Page 10 4 Mar 1991 Best way of making FOR cyclus, and what about WHILE cyclus Next from serie of speed tests of CLipper is here. At this time we will be looking how is working FOR command and if is not better to make replacement with WHILE. TEST programs used for this purpose are simple: TEST1 pure FOR loop with 10000 of iterations │ local start := 0 ,; ─┐ │ n,n1 ├─the same for ALL │ ─┘ tests │ ? "test no 1, FOR 0->9999" │ start = seconds() │ for n:=0 to 9999 │ ?? n,chr(13) │ next │ ? │ ? " elapsed ",seconds()-start TEST2 pure FOR loop with 100,000 of iterations │ ? "test no 2, FOR 0->99999" │ start = seconds() │ for n:=0 to 99999 │ ?? n,chr(13) │ next │ ? │ ? " elapsed ",seconds()-start TEST3 combined two FOR loops with 100,000 of iterations │ ? "test no 3, FOR 0->9 ( FOR 0->9999 )" │ start = seconds() │ for n1=0 to 9 │ for n=0 to 9999 │ ?? n,chr(13) │ next │ next │ ? │ ? " elapsed ",seconds()-start TEST4 combined two FOR loops with 100,000 of iterations │ ? "test no 4, FOR 0->9999 (FOR 0->9)" │ start = seconds() │ for n1=0 to 9999 │ for n=0 to 9 │ ?? n,chr(13) │ next │ next │ ? │ ? " elapsed ",seconds()-start Comments: difference of T3 and T4 is only in way of nesting FOR loops. It was just for fun with possible different results. Functionally T2, T3 and T4 are doing EXACTLY the CLIPBBS 1-06 Page 11 4 Mar 1991 same job, only with different way. T1 is for counting theoreticall time for T2,T3 and T4 │ T1 │ T2 │ T3 │ T4 ────────┼───────────────┼───────────────┼───────────────┼────────── result│ 19.93 │ 213.55 │ 199.82 │ 200.31 theory│ │ 199.93 │ 199.93 │ 199.93 ────────┼───────────────┼───────────────┼───────────────┼────────── diff │ │ 13.62 │- 0.11 │ 0.38 Lets take a look to results. - First look is telling, that T3 is FASTEST from all other test with 100,000 of iterations. Real FOR with only one FOR command is much slower than TWO NESTED FOR LOOPS. Why is this? It's very simple and it can be seen on screen when tests are running. At moment when T2 is running over iteration with number 65535 is coming SLOWNESS of displaying with comparation to previous speed. Why is this happening is again very simple and it's NOT any "bug" or "gotcha" of Nantucket. Registers of your mikroprocessor in your PC are 16bits, it means they are able to keep value 0-65535. Nantucket seems made good work, because is OPTIMIZING small loops into ONLY one 16bit register and only when needed is taking another 16bit register. Then it's probably working with 32bits for FOR command and is of course slower (must operate with two instead of one register). - Above mentioned register explanations is also explaining why T3 and T4 are faster than T2. They are using TWO times only 16bit registers and therefore must be faster because of saved time. Time used for switching between two FORs is very small. - Still seems better to make 0->9 (0->9999) than T4, because it went (only little) bit faster. For lot of fan, i did another two tests with WHILE. Both are simulating T1 and T2. TEST11 WHILE loop for 10,000 iteration │ local start := 0 ,; │ n │ │ ? "test1, WHILE 0->9999" │ n = 0 │ start = seconds() │ while n++<10000 │ ?? n,chr(13) │ end │ ? │ ? " elapsed ",seconds()-start CLIPBBS 1-06 Page 12 4 Mar 1991 TEST12 WHILE loop for 100,000 iterations │ ? │ ? "test2, WHILE 0->99999" │ n = 0 │ start = seconds() │ while n++<100000 │ ?? n,chr(13) │ end │ ? │ ? " elapsed ",seconds()-start And now we can give FINAL table connected with previous one. │ T1 │ T2 │ T3 │ T4 ────────┼───────────────┼───────────────┼───────────────┼────────── result│ 19.93 │ 213.55 │ 199.82 │ 200.31 theory│ │ 199.93 │ 199.93 │ 199.93 ────────┼───────────────┼───────────────┼───────────────┼────────── diff │ │ 13.62 │- 0.11 │ 0.38 ────────┴───────────────┴───────────────┴───────────────┴────────── │ T11 │ T12 ────────┼───────────────┼──────────────── result│ 19.94 │ 213.33 ────────┴───────────────┴──────────────── In long run (T12) seems WHILE loop a bit faster. Interesting again. In normal run (with 16bit counting) it's bit slower, or the same speed as FOR loop. .DD. ----------------------------------------------------------------------------- CLIPBBS 1-06 Page 13 4 Mar 1991 WHICH packing/unpacking program to choose??????? Maybe you know, there exists war called Archivers War. Everyone is trying to explain why one archiver is better than other. My small addition to this "WAR" is only pure view to times and sizes created with using of several kind of archive programs. First, i took a database of size about 1MB because it can be very interesting for us, Clipper programmers. Files ORDER.DBF size 1121395 bytes ┌──────────┬────────┬───────┬────────────┬────────┬────────────────┐ │Archivers │ Make │Restore│ Final Size │ Percent│ Comment │ │ │ time │ time │ │ │ │ ├──────────┼────────┼───────┼────────────┼────────┼────────────────┤ │ZIP1.10 │ 4.78 │ 0.13 │ 119,038 │ 10.61 │ │ │ARJ2.20 │ 1.08 │ 0.27 │ 106,473 │ 9.49 │ (no options) │ │ARJ2.20 │ 1.01 │ │ 106,730 │ 9.51 │ -m1 │ │ARJ2.20 │ 0.88 │ │ 110,324 │ 9.83 │ -m2 │ │ARJ2.20 │ 0.62 │ │ 114,288 │ 10.19 │ -m3 │ │ARJ2.20 │ 0.48 │ 0.30 │ 142,626 │ 12.71 │ -m4 │ │ARJ2.20 │ 4.35 │ 0.23 │ 101,028 │ 9.00 │ -jm │ │ARJ2.20 │ 1.08 │ │ 106,202 │ 9.47 │ -jh65500 │ │PAK2.51 │ 0.90 │ 0.30 │ 134,845 │ 12.02 │ │ │PKPAK3.61 │ 0.22 │ 0.20 │ 123,691 │ 11.03 │ │ │LHA2.12 │ 1.56 │ 0.22 │ 108,286 │ 9.65 │ │ └──────────┴────────┴───────┴────────────┴────────┴────────────────┘ And then, i took RTLINK.EXE (everybody has this 300KBs big monster) and runned again the same compressions. Files RTLINK.EXE size 354701 bytes ┌──────────┬────────┬───────┬────────────┬────────┬────────────────┐ │Archivers │ Make │Restore│ Final Size │ Percent│ Comment │ │ │ time │ time │ │ │ │ ├──────────┼────────┼───────┼────────────┼────────┼────────────────┤ │ZIP1.10 │ 0.47 │ 0.08 │ 179,045 │ 50.47 │ │ │ARJ2.20 │ 0.63 │ 0.18 │ 166,526 │ 46.94 │ -m1 and no opt.│ │ARJ2.20 │ 0.48 │ │ 169,080 │ 47.66 │ -m2 │ │ARJ2.20 │ 0.40 │ │ 171,559 │ 48.37 │ -m3 │ │ARJ2.20 │ 0.28 │ │ 187,824 │ 59.95 │ -m4 │ │ARJ2.20 │ 0.67 │ │ 163,283 │ 46.03 │ -jm │ │ARJ2.20 │ 0.63 │ │ 166,509 │ 46.94 │ -jh65500 │ │PAK2.51 │ 0.50 │ 0.17 │ 176,890 │ 49.87 │ │ │PKPAK3.61 │ 0.17 │ 0.12 │ 241,809 │ 68.17 │ │ │LHA2.12 │ 0.60 │ 0.18 │ 170,431 │ 48.05 │ │ └──────────┴────────┴───────┴────────────┴────────┴────────────────┘ And now, we can do some sorts - editor will do this. Fastest on .DBF T1 points CLIPBBS 1-06 Page 14 4 Mar 1991 PKPAK3.61 │ 0.22 11 ARJ2.20 │ 0.48 -m4 10 ARJ2.20 │ 0.62 -m3 9 ARJ2.20 │ 0.88 -m2 8 PAK2.51 │ 0.90 7 ARJ2.20 │ 1.01 -m1 6 ARJ2.20 │ 1.08 (no options) 5 ARJ2.20 │ 1.08 -jh65500 4 LHA2.12 │ 1.56 3 ARJ2.20 │ 4.35 -jm 2 ZIP1.10 │ 4.78 1 Fastest on .EXE T2 points PKPAK3.61 │ 0.17 10 ARJ2.20 │ 0.28 -m4 9 ARJ2.20 │ 0.40 -m3 8 ZIP1.10 │ 0.47 7 ARJ2.20 │ 0.48 -m2 6 PAK2.51 │ 0.50 5 LHA2.12 │ 0.60 4 ARJ2.20 │ 0.63 (no options) 3 ARJ2.20 │ 0.63 -jh65500 2 ARJ2.20 │ 0.67 -jm 1 Creating smallest result on .DBF T3 points ARJ2.20 │ 101,028 -jm 11 ARJ2.20 │ 106,202 -jh65500 10 ARJ2.20 │ 106,473 (no options) 9 ARJ2.20 │ 106,730 -m1 8 LHA2.12 │ 108,286 7 ARJ2.20 │ 110,324 -m2 6 ARJ2.20 │ 114,288 -m3 5 ZIP1.10 │ 119,038 4 PKPAK3.61 │ 123,691 3 PAK2.51 │ 134,845 2 ARJ2.20 │ 142,626 -m4 1 Creating smallest result on .EXE T4 points ARJ2.20 │ 163,283 -jm 10 ARJ2.20 │ 166,509 -jh65500 9 ARJ2.20 │ 166,526 -m1 and no opt. 8 ARJ2.20 │ 169,080 -m2 7 LHA2.12 │ 170,431 6 ARJ2.20 │ 171,559 -m3 5 PAK2.51 │ 176,890 4 ZIP1.10 │ 179,045 3 ARJ2.20 │ 187,824 -m4 2 PKPAK3.61 │ 241,809 1 POINT scales (first way of view results): Archiver │ switches │ T1 T2 T3 T4 ║ SUM ──────────┼────────────────┼────────────────╫─────── ARJ2.20 │ -m3 │ 9 8 5 5 ║ 27 ARJ2.20 │ -m2 │ 8 6 6 7 ║ 27 PKPAK3.61 │ │ 11 10 3 1 ║ 25 CLIPBBS 1-06 Page 15 4 Mar 1991 ARJ2.20 │ -m1 │ 6 3 8 8 ║ 25 ARJ2.20 │ (no options)│ 5 3 9 8 ║ 25 ARJ2.20 │ -jh65500 │ 4 2 10 9 ║ 25 ARJ2.20 │ -jm │ 2 1 11 10 ║ 24 ARJ2.20 │ -m4 │ 10 9 1 2 ║ 22 LHA2.12 │ │ 3 4 7 6 ║ 20 PAK2.51 │ │ 7 5 2 4 ║ 18 ZIP1.10 │ │ 1 7 4 3 ║ 15 This first way of viewing result is giving very funny result, it's moving PKZIP at LATEST place because of BIG slownes on .DBF files and still worse results of compression than most from ARJ methods. Shocking is place of PKPAK3.161 on third place because if you will look back, then it's giving most worse result of size of result archives, but it's fastest from all arhives tested (in MAKING because extracting is not so much interesting in this competition). LOGICAL SCALES (second way of viewing results) Logical way of looking to results is very simple. PKPAK is not good for use because it's producing worse results of compressing. Logically i will then want forget about nice times of packing (actually, they aren't so much different, because machine used for testing was only 286/16Mhz LAPTOP). Then personal/logical scale is: ARJ (dependent from switches) LHA ZIP all others ready for drop out .DD. ----------------------------------------------------------------------------- CLIPBBS 1-06 Page 16 4 Mar 1991 Will Clipper expand to graphics environment? alias Theoretical article about FUTURE OF CLIPPER and NANTUCKET :-) What's a Clipper at this moment? I'm sure about one thing, at this moment it's character based compiler with many difficulities of implementing it inside of graphics based environment. With this is connected also question when we will see a Clipper for Window.... Clipper needs great change to be able fulfill all needs for graphically oriented application of this time. Windows and any graphics based environment is fixed to fact, that application or compiler used for creating or programs are enough fast in moment of displaying. If not, user is getting complete nice system with not nice background of slow speed. Way of clippering at this moment is single process task. Windows is event driven system with many facilities of multitasking system, including DDE (dynamic data exchange) system. Nantucket then has two ways to choose: - slowly adapt Clipper to new environment with producting several changed versions until it will fit into needs of Windows programming - drastically change Clipper to something different and incompatible in many cases with original xBASE language ideas. Problem of second alternative is, that it will leave many programmers in "old" fashioned Clipper which can come into dead end. Will not be changed and upgraded because will be here new "windows Clipper" with main interest of Nantucket (passing perfect inside of WINDOWS hysteria coming into world). In this drasctic case will come something like is coming in case of S87 -> 5.0x jump. Many old programmers have problems to recognize significant changes and additions of new Clipper and are not able to use it in all directions. But it's still very small and non problematic at all. What can come in case of Windows Clipper will be creating of new compiler. In case of Nantucket which seems to jumping from left to right with difficulties to maintain correctly and enough fast world market it can mean only one. If step to create Windows Clipper will come into reality, Nantucket will have to drop idea of Clipper in old fashion and create new language, or replace completely philosophy of market. Of course is here possibility to follow first slow way, which can be more easy for Nantucket, but it can only mean loosing of Windows market whcih is coming to be bigger and bigger and probably loose head with competition with Borland (which had always better market politic than Nantucket, at least in other continents than America.) From windows Clipper and windows market let make step out to take a CLIPBBS 1-06 Page 17 4 Mar 1991 look to actual state of Clipper. First, i have to say, that it's mine personal view and i'm living in EUROPE and not in U.S. where will probably everything looks different and probably much better. Nantucket, producing one from most complex xBASE compatible programming systems Clipper made a big step with not enough saved back. Clipper 5.0 was a crash of many possible customers and dealers because was unbelievable buggy. Game with dates of releasing 5.01 (which actually should be 5.1 because it's not only bugfix) and final decision NEVER telling any date, NEVER telling any bug reports means lost of trust from many sides. Lost of money in case of firms were trusting in 5.0 version can be sometime very big number. Promises of informing registered users about "something" from acting in Nantucket are also left. Only one what i got as registered user was Clipper 5.01, but not ANY informations. It's matter of local supporters and dealers of course, but then it's also in hands of Nantucket itself. Is possible also say, that Clipper itself is fulfilled with not filled promises. From biggest ones is easy to point out: NEVER distributed dBASE index drivers NEVER working incremental linking of RTLINK NEVER published Clipper Programmers Interface NEVER published Replaceable Database Drivers information, nor any existing drivers STILL problematic use of Clipper programs in small memory of PCs Still repeating stories with bugs (so called ANOMALIES by Nantucket) which are invented by many parties, but for sure not Nantucket is also not giving good image to Clipper. Especially if some problems are staying permanently from version to version which should be a bugfix. For European users is GOOD technical support of Nantucket something as midnight dream. FAXES or telephones to Nantucket offices are mostly unaswered or without ANY professional answer when they are little bit difficulty than reading of standard manuals. Possibility to go to COMPUSERVE is in Europe very hard and when, then is very expensive. Therefore putting attention into Compuserve from side of Nantucket is maybe good for U.S., but not for any other parts of world. Very good idea of setting up Nantucket Bulletin Board in England is slowly going to nowhere because staff of Nantucket is not able to decide WHO and HOW OFTEN has to work on this Bulletin Board and Clipper World (that's name of this BBS) is dying and staying on dead point. With all this oscillating background will be very difficult to come into professional WINDOWS market. Especially because will be needed probably to get second, drastic way. .DD. ----------------------------------------------------------------------------- CLIPBBS 1-06 Page 18 4 Mar 1991 ============================================================================= SOFTWARE ============================================================================= TIPS&TRICKS - how to found largest string in array In some of mine function for displaying MENU on screen (generic function) was coming a need to count SIZE of window for MENU, because text of every field can come in different length. For this sample, we have to say: INPU is aMenu - array of texts for MENU function. Something like: aMenu := {"First","Second","Third","Latest one"} And then, it's simple to use aeval() function for get what we need. local nMaxLen := 0 aeval(aMenu, {|n| nMaxLen:=max(len(xx),nMaxLen)} ) And that's all. Some explanations: aeval(aMenu, {|n| nMaxLen:=max(len(n) ,nMaxLen)} ) │ │ └──────┬───────┘ │ │ └── getting maximal from │ │ two values and assigning │ │ it as new maximal length │ └──────── It's needed for aeval() function, every │ time it's getting value of one portion of │ array └────── aeval() function is scanning WHOLE array, field by field and is using codeblock defined as second parameter ----------------------------------------------------------------------------- DANIEL.LIB - NEW version of DANIEL.RMK After getting NANFOR.LIB got inspired with their .RMK file and made totally new one for DANIEL.LIB. Old one is of course working still, but this one is far better. Feel free to get it as inspiration also, because it's nice piece of easy to change .RMK file. I'm absolutely sure, that using of RMAKE is someting that is staying at lowest level of interest to upgrade in knowledges and in praxis in most of programmers minds. Mine opinion is different and sometime i'm getting mad to make some upgrades of my using of .RMK files. This is one from them. <----------------------------------FILE BEGINS HERE------------------------> /*************************************************************************** * $Author$ * $Date$ CLIPBBS 1-06 Page 19 4 Mar 1991 * $Revision$ * $RCSfile$ * $Source$ * $State$ * $Locker$ * * $Log$ * ***************************************************************************/ // ------------------------------------------------------------------------ // PATH macros for compiling and/or linking. They MUST be set correctly // ------------------------------------------------------------------------ l = \LIB o = \SOURCE\DANIEL\OBJ s = \SOURCE\DANIEL c5 = C:\CL5\BIN tc = C:\TC\BIN mc = C:\MSC\BIN ma = C:\MASM ta = C:\TASM ex = \SOURCE\DANIEL // ------------------------------------------------------------------------ // Killing working giles used for manipulation or checkings // ------------------------------------------------------------------------ wrk = DANIEL.RSP !iffile $(s)\$(wrk) #! del $(s)\$(wrk) !endif // ------------------------------------------------------------------------ // Now some internal definition dependent from DEFs from outside // ------------------------------------------------------------------------ // DEBUG is defined in case of request for DEBUGGING capable version !ifdef DEBUG SW = /W/N/M/B SW1 = /DDEBUG !stdout "DANIEL.RMK - DEBUGGER version generated" !else SW = /W/N/M/L !stdout "DANIEL.RMK - NOTDEBUG version generated" !endif // ------------------------------------------------------------------------ // And some defintions which are changing internal parameters // ------------------------------------------------------------------------ PLL = /PLL:FULLBASE CLIPBBS 1-06 Page 20 4 Mar 1991 // ------------------------------------------------------------------------ // apply above defined paths into internal RMAKE defintions for DEFAULTS // default will be used when will not be specified directly. Can save some // time and work with filling all informations // ------------------------------------------------------------------------ makepath[.OBJ] := $(o) makepath[.PRG] := $(s) makepath[.EXE] := $(ex) makepath[.LIB] := $(l) makepath[.CH] := $(s) makepath[.INC] := $(s) makepath[.C] := $(s) // ------------------------------------------------------------------------ // inference rules for TC files // WARNING $(tc)... TWO lines are actually ONE line. For space in magazine // was divided in two. BEFORE USING PUT BACK!!!!!! // ------------------------------------------------------------------------ .c.obj: $(tc)\TCC -I\TC\INCLUDE -L\TC\LIB -C -d -ml -wmlt -K -G -O -Z -k- -c -o$(o)\$* $(s)\$*.C echo -+$(o)\$* & >> $(s)\$(wrk) // ------------------------------------------------------------------------ // inference rules for .PRG files // ------------------------------------------------------------------------ .prg.obj: $(c5)\clipper.exe $(s)\$* $(sw) $(sw1) /o$(o)\$* echo -+$(o)\$* & >> $(s)\$(wrk) // ------------------------------------------------------------------------ // inference rules for library from RSP // ------------------------------------------------------------------------ .RSP.LIB: echo ,$(l)\DANIEL.LST; >> $(s)\$(wrk) if not exist $(l)\DANIEL.LIB LIB $(l)\DANIEL.LIB LIB.EXE $(l)\DANIEL.LIB @$(s)\$(wrk) DEL $(l)\DANIEL.BAK // ------------------------------------------------------------------------ // check existency of CLIPPER.EXE and then dependece of CLIPPER files // ------------------------------------------------------------------------ !iffile $(c5)\CLIPPER.EXE dd_lib1.obj : dd_lib1.prg dprinter.ch daniel.ch dd_lib2.obj : dd_lib2.prg dprinter.ch daniel.ch dd_lib3.obj : dd_lib3.prg dprinter.ch daniel.ch dd_lib4.obj : dd_lib4.prg dprinter.ch daniel.ch dd_lib5.obj : dd_lib5.prg dprinter.ch daniel.ch dd_netw.obj : dd_netw.prg dprinter.ch daniel.ch configer.obj: configer.prg dprinter.ch daniel.ch CLIPBBS 1-06 Page 21 4 Mar 1991 !else // clipper.exe doesn't exit !stderr "DANIEL.RMK error: CLIPPER.EXE file not found...." !endif // ------------------------------------------------------------------------ // check existency of TCC.EXE and then dependece of C files // ------------------------------------------------------------------------ !iffile $(tc)\TCC.EXE cleanup.obj : cleanup.c faxlog.obj : faxlog.c isprinte.obj: isprinte.c screen.obj : screen.c wptext.obj : wptext.c wputil.obj : wputil.c !else // TCC.EXE doesn't exist !stderr "DANIEL.RMK error: TCC.EXE file not found....." !endif // ------------------------------------------------------------------------ // final dependence of library // ------------------------------------------------------------------------ DANIEL.LIB : $(wrk) <------------------------------END OF FILE--------------------------------> .DD. ----------------------------------------------------------------------------- CLIPBBS 1-06 Page 22 4 Mar 1991 ============================================================================= ANOMALIES ============================================================================= ANOMALIES and their comments This part of Clipper BBS Magazine is dedicated to all discovered anomalies and comments about them in Clipper products. Because Nantucket is still unable to give own bug and anomalies reports (as actually did in past with Summer 87 version) is very handy to have results of many investigations done on many user places. I'm also doing my own investigatings, because i'm always very good when someting has hidden problems. Everything what i buy will first show all problems and then all normal things. This amazing part of my live is sometime making me crazy, but for testing of programs it's great <grin>. Daniel ----------------------------------------------------------------------------- Clipper Debugger anomalies CLD is now great program and i'm really using it every day. Previous CLD was unusable because of memory problems of CLIPPER (no one my application was able to fit in memory together with CLD.EXE nor CLD.LIB). There is lot of improvements in this new debugger, but of course there are same nonlogical and non working things: Totally new HELP system for CLD is enough descriptive, but not CONTENTS SENSITIVE. When one will press F1 (Help key), then will get first page of help and MUST scan through all of them to page which need. In case of small memory left is this process VERY slow (because of lot of disc swaps). Don't try to use HELP system when you are editing (for example) colour codes or path string for your source files. If one will press F1 at this moment,then will get some graphics codes which are equal to code of F1, but will not get help. Somebody who made it forget that F1 is help and not a valid key for editing characters... One must still be carefully when specifying command line options for SPLIT and EGA/VGA screen. CLD /S/50 and CLD /50/S is really not the same for CLD.. Just use spaces between them and then it's working well (CLD /S /50 is the same as CLD /50 /S) Be aware of using ? row() or ? col(). Both of them will return not value from YOUR program, but value from inside of Debugger. Debugging a program and needs to know values of cursor position is at this moment totally impossible. Some from people using CLD will report - When i pressed F8, then CLIPBBS 1-06 Page 23 4 Mar 1991 CLD didn't make steps through my procedure, he just did action for F10 (execute procedure). Be aware at this moment if your .OBJ files ALL had specified /B switch for CLIPPER command line. If not, then CLD will not warning you of course, will just step through. You can then be surprised and thinking what is wrong! There is IMPOSSIBILITY to debugging a code which is using #include statements for include PART of CODE (not only #define and other definitions). Because Clipper is allowing you to use #include for normal part of program (with all commands...), then i'm expecting that will be able to correct do debugging of this includes. ABSOLUTELY not, when one will do this, will have a debugger step cursor somewhere in .PRG file, but on wrong place of course, because of mixup of line numbers. Small example of this: Program.prg 1 function Main() 2 #include "small.inc" 3 ? 'Hello this is NOT a first line of program' 4 ? 'and there will be a RETURN coming' 5 RETURN NIL File SMALL.INC 1 ? 'This is a first line of program of course' 2 ? '-----------------------------------------' Of course, after compiling here is another program: 1 function Main() 2 ? 'THis is a first line of program of course' 3 ? '-----------------------------------------' 4 ? 'Hello this is NOT a first line of program' 5 ? 'and there will be a RETURN coming' 6 RETURN NIL CLD will start a debugging and display PROGRAM.PRG file on screen with debugging step cursor on line "2" of PROGRAM.PRG file. Executed instruction but will be a instruction from line "1" of SMALL.INC. I'm so sure, that this stupid acting is possible to eliminate just by better recognizing of includes and modifying screen display routines of source code. There is of course EASY solving for this - NEVER use #include for part of your code. All your code you can easy put to separate .OBJ files and call them like function. Then will CLD working GREAT. Possibility to use Alt-D function for stop program in working is limited. If there will be for example: WHILE rlock() ENDDO (of course, this is a mistake. But i made it few time, just forgotten "!" negation) Then one will NOT be able to stop program and see what is working. When i had this problem i did lot of looking what is wrong and after LONG period i found that there is missing one negation for CLIPBBS 1-06 Page 24 4 Mar 1991 condition. .DD. ----------------------------------------------------------------------------- INCONSISTENCY One funny inconsistency of program language is in new Clipper 5.01 version. I found it when i was thinking: OK, there are @ SAY .. COLOUR, @ .. TO .. COLOUR command (something new in 5.01, extension of @ command). Then there OF COURSE is @ .. CLEAR .. COLOUR. GOTCHA, there is NOT @CLEAR with COLOUR parameter. Maybe somebody is thinking that CLEAR is without colour, but then is wrong. .DD. ----------------------------------------------------------------------------- New RTLINK version, INCREMENTAL LINKING Another expectation which i had when new Clipper 5.01 was coming - incremental linking in new version of Clipper will work. Truth is again different. DO NOT use incremental linking. It's working successfully only for few first links and then programmer will receive nice set of error from program. Probably will then start looking to errors in own program, but reality is - incremental linking is NOT working. First, second and sometime third link is WITHOUT problems, but after this you can receive errors of this kind: Syntax error: + Be sure, that on your program line where report is reported doesn't exist a "+" operation Syntax error Just funniest error, because it's SYNTAX error on compiled and linked code. Maybe there is processor of your PC reporting non existed function.... Cannot create overlay file: !^%#@(* Really, i cannot reproduce file name reported by overlay manager, because this magazine will not probably have all control and graphical symbols which one can get in IBM PC environment. All this errors you will get in part of code which was a subject of changes, everything other (not changed) is still working pretty perfect. FAX which i then sent to Nantucket, together with uploads of these error reports to Nantucket special message area on CLIPPER WORLD BBS are WITHOUT answers. CLIPBBS 1-06 Page 25 4 Mar 1991 FAX to Pocket Soft Inc., authors of RTLINK, is with answer: RTLINK version shipped with Clipper 5.01 is working with INCREMENTAL linking without problems. It passed our and Nantucket test suites. If you have some problems, please contact Nantucket..... .DD. ----------------------------------------------------------------------------- CLIPBBS 1-06 Page 26 4 Mar 1991 ============================================================================= ECHODUMP ============================================================================= CONFERENCE DUMP part (5) Subj: NTX Structure Prvt: N Read: N I think Rick Spence's book on Clipper 5.0 has description of NTX file structure and others used in the Clipper environment. Subj: OBJECT VISION Prvt: N Read: N Borland has released a new toy called "Object Vision". Has anyone out there used this product? Looks real interesting in the advertisements. Subj: PKLITE/Blinker? Prvt: N Read: N very interesting. Had assumed Lite e.a. would not work with overlayed programs at all. Do you have any experience using it with Blinker? Subj: PKLITE/BLINKER? Prvt: N Read: N I could be wrong, but strictly speaking from my shareware copy of the PKLITE program, it advises not to compress overlayed .EXE-s ... Having used it on programs not utilizing overlays say about 320K, I have reduced sizes down to say about 125-130K ... It's neat . One thing one must remember is that there is no reduction in memory used size form the original program . Say if it was 450K before, it will still be 450K even though it has been compressed. However, I can't state what the commercial version is capable of < least not yet > ... Subj: PKLITE/Blinker? Prvt: N Read: N I principally use Warplink, with Tlink for short things and in development cycles, though Warplink seems to be taking over that as well. I have played with Blinker, but I prefer the MSLink-style of linker invocation; hence Warplink. I can't say whether Pklite will work or not with Blinker, but I heard it will. Apparently, the issue has to do with external vs. internal overlays. I believe I saw a message thread to that effect either on this Clipper net or one of the others. BTW, note that the issue of linker speed is irrelevant when using incremental linking. When relinking incrementally the link time is one second or so. In a large program with complex overlaying done internally Pklite is the longest part of cycle, so save it for release versions. Subj: Randomness Prvt: N Read: N Is it my imagination or does Clipper not have a random number function? All I could find in the manual is a program to maipulate the time function to simulate a random number generator. If Load-us and the ("Out to")-C programs can do it, why can't dBase or Clipper without CLIPBBS 1-06 Page 27 4 Mar 1991 all the contortionism? Subj: Randomness Prvt: N Read: N Standard Clipper just don't have a Random() function. Some add-on libs does thou, eg. FUNCky. ----------------------------------------------------------------------------- CLIPBBS 1-06 Page 28 4 Mar 1991 ============================================================================= CLIPPER NET ============================================================================= Following is COMPLETE list of all published file descriptions in Clipper BBS magazine in previous numbers. Purpose of this index list is to allow anybody find needed file descriptions in growing number of described files. Short description after name will give first possible close image about file. Number enclosed in "[]" will mean number of Clipper BBS magazine. ┌─────────────┬────┬────────────────────────────────────────────────┬──────┐ │FileName │Src │Description │Where │ ├─────────────┼────┼────────────────────────────────────────────────┼──────┤ │ACH2TB.ARJ │Cln │Convert ACHOICE to TBROWSE │[1-05]│ │ADHOC302.ARJ │Cln │Summer 87 inteligent report program │[1-04]│ │CL5103.ARJ │Cln │Report of 5.01 anomaly number 3 │[1-04]│ │CL5REP6.ARJ │Cln │5.01 replacement of REPORT command │[1-04]│ │CLIP110.ARJ │Cln │Clipper Documentor program │[1-05]│ │CLIPLINK.ARJ │Cbs │Complete text of R.Donnay about linkers │[1-04]│ │CLIPSQL.ARJ │Cln │Demo of complete SQL library for CLipper │[1-05]│ │CLPFON.ARJ │Cln │Set of fonts for EXPAND.LIB from author │[1-03]│ │COND.ARJ │Cln │Builder of conditional indexes like SUBNTX │[1-03]│ │DBSCN2.ARJ │Cln │<sorry, forget, will add in next number, hope> │[1-05]│ │HGLASS.ZIP │Cln │Hour glass for indication of index progression │[1-04]│ │INDXSL.ARJ │Cln │User Fields selection builder for index generate│[1-03]│ │IOBASYS9.ARJ │Cln │Demo of S87 library and calling Clipper from C │[1-03]│ │MK30.ARJ │Cln │Mouse library demo version │[1-03]│ │MOVEGETS.ARJ │Cln │GETSYS change for moving between gets via VALID │[1-03]│ │NOTATION.ARJ │Cln │Complete text of article about hungarian notat. │[1-04]│ │PACKUP.ARJ │Cln │ASM source of PACK/UNPACK replacement SCRSAVE.. │[1-04]│ │POPUPCAL.ARJ │Cln │Popup calender │[1-05]│ │READPW.ARJ │Cln │GETSYS change for password invisible reader │[1-03]│ │SCRSAVE.ARJ │Cln │Screen AntiBurning utility (inactivity snake) │[1-05]│ │SYMBOL.ARJ │Cln │Dumper of symbol tables of Summer87 .EXE │[1-03]│ │VSIX711.ARJ │Cln │Vernon Six Clipper utilities and library │[1-05]│ └─────────────┴────┴────────────────────────────────────────────────┴──────┘ Src can be: Cln File is accesible on ClipperNet Cbs File is accesible in HQ BBS of CLipper BBS Magazine ----------------------------------------------------------------------------- Routing-Plan for ClipperNet Followin text is routing schema of ClipperNet from last friday. Someone interested in connecting to ClipperNet, find your nearest RC or NC node and ask for information. Someone for downloading files from ClipperNet advertised here, find closest BBS, take a look into FIDONET NODELIST (all number are FIDONET NODENUMBERS.. CLIPBBS 1-06 Page 29 4 Mar 1991 CLN-HQ 2:240/100 Clipper WorkShop BBS (Hamburg - FRG) RC 24 | | RC 23 |234/64 TugBoat BBS (Glumsoe - DK) | | | |231/62 TBSM's Opus (Kolding - DK) | |240/1 Ramos II (Hannover - FRG) NC 240 |240/102 MuM BBS (Hamburg - FRG) |240/600 IGEPOS Mailbox System (Lage - FRG) |241/2004 NordSee BBS (Leer - FRG) | | | |241/5302 Terrania City (Bonn - FRG) | |241/2103 H.C.C. (Hamburg - FRG) | |241/5609 BLM (Wermelskirchen - FRG) | | | |241/4008 Stone's Box (Haan - FRG) | | | | | |241/5603 Funboard (Velbert - FRG) | | | | | |241/4512 BS-1 (Bad Salzuflen - FRG) | | | | | |241/4511 MICOS BS (Bad Salzuflen-FRG) | | | | | |241/4501 Life-Box(Osnabrueck-FRG) | | | |241/5303 NEPTUN Box (Bonn - FRG) | | | | | |243/93 Piano-Bar (Urbar - FRG) | | | |241/5608 Pit's BBS (Wermelskirchen - FRG) | NC 241 |241/7605 Baden Area (Offenburg - FRG) | | | |241/7000 Popey-Box (Donaueschingen - FRG) | |242/47 Dungeon of Despair (Aachen - FRG) | NC 243 |243/29 Clipper Ship (Nieder Olm - FRG) | | | |247/31 Tritech BBS (Seelzen - FRG) | | | | | |247/9 Wildbox (Heidelberg - FRG) | | | | | |247/1 Schawott-Host (Brensbach - FRG) | | | |249/29 Guenter's Mailbox (Mainz - FRG) | NC 245 |245/60 Bonny's (Dortmund - FRG) | | | |241/4152 Quick!-Box (Meerbusch - FRG) | | | | | |242/92 Skyhigh Mailsystem (Krefeld - FRG) | | CLIPBBS 1-06 Page 30 4 Mar 1991 | |245/9 MANNY'S BBS (Essen - FRG) | | | | | |245/102 WiInf_BBS (Essen - FRG) | | | |245/36 SoftStream (Dortmund - FRG) | NC 246 |246/15 MagicBox Munich (Muenchen - FRG) | | | |241/8007 Merlin BBS (Nuernberg - FRG) | |246/5 PERI-BBS (Kassel - FRG) | |246/10 Beta-System (Muenchen - FRG) | |246/26 AGS-Box (Leutkirch - FRG) | |246/32 Aischgrund Mailboxsystem (Gerhardshofen-FRG) | | | |246/12 AlphaSoft Network (Fuerth/Stadeln - FRG) | RC 28 |285/610 Lobster BBS (Schiedam - NL) | | | |285/402 AllTime Board (Koudekerke - NL) | |285/608 NETCONSULT BBS (Rotterdam - NL) | |500/223 HEKOM BBS (Harmelen - NL) | | | | | |283/314 KISS Software Support (Almelo - NL) | | | |500/236 Touch-One (Den Haag - NL) | |512/43 Flash BBS (Utrecht - NL) | RC 29 |296/10 BBS D.C.V.V. (Maaseik - B) | | RC 20 | |201/254 Clipper(Sweden) BBS (Vallingby - S) | | | | | |200/306 Lanthandelen (Kalmar - S) | | | |294/5 CIS BBS (Antwerpen - B) | RC 30 |302/801 Gepard's Oracle (Zuerich - CH) | | | |302/807 SENECA (Zuerich - CH) | | | | | |302/804 Triton (Zuerich - CH) | | RC 32 | |321/100 Cyrielle (Montargis - F) | | RC 33 | |331/110 DB-Line (Biandronno - I) | RC 31 |313/9 EuroSoft Net 2 (Wien - A) | RC 34 |341/25 Clipper U.C.P. (Majadahonda - E) | |341/8 AlpaSoft (Madrid - E) |341/14 Rafa_STD-1 (Madrid - E) ----------------------------------------------------------------------------- CLN - SOUND.ARJ CLIPBBS 1-06 Page 31 4 Mar 1991 File Name: SOUND.ARJ Other Names: File Size: 2,336 bytes File Contents: SOUND.CH 1447 SOUND.TXT 1762 SOUND.PRG 411 Sound function using TONE() function for playing a melody. Point of this function is, that it's getting parameter as array of arrays contains in every field tone and length of piece of melody. In SOUND.CH file are defined some particular notes and two sample sounds for your programs. It's also nice way how to implement note names in clipper programs anyway. .DD. ----------------------------------------------------------------------------- CLN - TBWHL4.ARJ File Name: TBWHL4.ARJ Other Names: File Size: 5,164 bytes File Contents: NAMES.DBF 1458 READ.ME 1317 NAMES.NTX 2048 TBWHILE.PRG 7041 MAKEIT.BAT 49 COMPUSERVER coming well commented function for WHILE browsing a database with Tbrowse objects. It's good as example how to handle own cases, because it contains complete sample with included database (.DBF). .DD. ----------------------------------------------------------------------------- ClipNet - NFDESC2.ARJ File Name: NFDESC2.ARJ Other Names: File Size: 5,129 bytes File Contents: NFDESC.TXT 12451 NANFORUM library is one from greatest Clipper projects which i ever seen. It's great idea of sharing user supplied, test and used sources into one big library. It's managed by some CompuServe users to final state of CLIPBBS 1-06 Page 32 4 Mar 1991 complete library with .NG documentation. First day when i looked into NANFORUM library i decided immediately take it inside of mine used libraries set. I can only reccomend to do this for you also. If want to take a look to list of functions in this library before taking over of complete needed library files, get this file, because .TXT inside is complete list of functions with descriptions. See also articles about NANFORUM in "SOFTWARE" section of magazine. .DD. ----------------------------------------------------------------------------- ClipNet - NFLIB2.ARJ File Name: NFLIB2.ARJ Other Names: File Size: 167,730 bytes File Contents: NANFOR.LIB 197161 FTINT86.CH 1250 FTMENUTO.CH 2329 FT_ACH2T.CH 4572 NANFOR.NG 333548 THIS IS MAIN FILE NEEDED for NanForum library. Complete documenation (.NG file) and complete 5.01 library (NANFOR.LIB) with needed definition files (.CH files) are included. Don't worry about size, it's great and needed for everybody to save LOT of programming. See articles about NANFOR.LIB in this magazine! .DD. ----------------------------------------------------------------------------- ClipNet - NFSRC2.ARJ File Name: NFSRC2.ARJ Other Names: File Size: 308,973 File Contents: ADAPTER.ASM 3428 BYTENOT.PRG 2531 AINT86.ASM 4453 BYTEOR.PRG 2659 CHDIR.ASM 2021 BYTEXOR.PRG 2788 DEFAULT.ASM 4535 CALENDAR.PRG 7395 DIRFUNC.ASM 2971 CLRSEL.PRG 24956 DISPA.ASM 4524 CNTRYSET.PRG 2403 ENDCAP.ASM 6266 COM34.PRG 2687 FLOPTST.ASM 9955 D2E.PRG 2453 FLUSHCAP.ASM 4195 DATECNFG.PRG 11619 CLIPBBS 1-06 Page 33 4 Mar 1991 HANDCNT.ASM 6372 DAYOFYR.PRG 3939 INP.ASM 3427 DAYTOBOW.PRG 2403 ISPRINT.ASM 8548 DECTOBIN.PRG 1763 KSPEED.ASM 6627 DFILE.PRG 6574 LOGOUT.ASM 2719 DIR2DBF.PRG 4195 MKDIR.ASM 2149 DISKFUNC.PRG 3349 NOVDMP2.ASM 2844 DISPMSG.PRG 8113 NOVPURGE.ASM 2219 DOSVER.PRG 2275 NWTTS1.ASM 3555 E2D.PRG 1749 NWTTS2.ASM 2788 EASTER.PRG 3683 NWTTS3.ASM 3044 ELAPMIL.PRG 1635 NWTTS4.ASM 2660 ELAPSED.PRG 4579 NWTTS5.ASM 2532 ELTIME.PRG 2275 ORIGIN.ASM 3299 EXIST.PRG 1837 OUTP.ASM 4196 FIELD.PRG 11491 PUTKEY.ASM 10595 FINDITH.PRG 3427 REBOOT.ASM 2660 FIRSTDAY.PRG 1763 RESTATT.ASM 6600 GCD.PRG 3043 RMDIR.ASM 2149 HEX2DEC.PRG 1772 SAVEATT.ASM 7908 INVCLR.PRG 2543 SHADOW.ASM 9060 ISBIT.PRG 2787 TMENU.ASM 38602 ISBITON.PRG 2019 _WHEREIS.ASM 23526 ISSHARE.PRG 2403 ALT.C 1507 LASTDAY.PRG 1891 CAPLOCK.C 2019 LINKED.PRG 4963 CINT86.C 8552 MADD.PRG 4451 CTRL.C 1507 MENU1.PRG 20198 DISPC.C 27601 MENUTO.PRG 12522 FTTEXT.C 16982 METAPH.PRG 14691 GETENVRN.C 5219 MILTIME.PRG 7523 NUMLOCK.C 2275 MIN2DHM.PRG 1636 PEEK.C 2147 MONTH.PRG 4195 POKE.C 2275 MOUSE1.PRG 23032 PROPER.C 3299 NETPV.PRG 3043 PRTSCR.C 2019 NOOCCUR.PRG 2531 SHIFT.C 1507 NTOW.PRG 4164 STOD.C 1507 NWLSTAT.PRG 2219 VIDEO1.C 11049 NWSEM.PRG 16077 FTINT86.CH 1250 NWUID.PRG 3811 FTMENUTO.CH 2329 PAGE.PRG 2531 FT_ACH2T.CH 4572 PCHR.PRG 8435 CINT86.H 1174 PEGS.PRG 8675 DFKEY.H 5117 PENDING.PRG 4835 FTTEXT.H 1410 PICKDAY.PRG 1763 TMENU.INC 3746 POPADDER.PRG 45352 MINTER.MAC 9931 PRTESC.PRG 2915 AADING.PRG 4772 QTR.PRG 3939 AAVG.PRG 2510 RAND1.PRG 2275 ACCTADJ.PRG 3427 RESTSETS.PRG 1891 ACCTMNTH.PRG 4579 ROUND.PRG 7267 ACCTQTR.PRG 4579 SAVEARR.PRG 6774 ACCTWEEK.PRG 3811 SAVESETS.PRG 2275 ACCTYEAR.PRG 3299 SCANCODE.PRG 3171 ACH2TB.PRG 27161 SCREGION.PRG 8705 ADESSORT.PRG 2636 SETDATE.PRG 2787 AEMAXLEN.PRG 3811 SETTIME.PRG 2915 CLIPBBS 1-06 Page 34 4 Mar 1991 AEMINLEN.PRG 3683 SINKEY.PRG 2787 AMEDIAN.PRG 3913 SQZN.PRG 4030 ANOMATCH.PRG 3127 SYSMEM.PRG 1930 ANY2ANY.PRG 6698 TBWHILE.PRG 14947 AREDIT.PRG 10932 TEMPFILE.PRG 3427 ASUM.PRG 2769 VERTMENU.PRG 6207 AT2.PRG 3555 VIDCUR.PRG 3965 ATPROMPT.PRG 17385 VIDMODE.PRG 3043 BITCLR.PRG 2915 WDA.PRG 2659 BITSET.PRG 2787 WEEK.PRG 3939 BITTEST.PRG 7779 WHEREIS.PRG 3939 BLINK.PRG 2403 WORKDAYS.PRG 2787 BYT2BIT.PRG 2609 WOY.PRG 5987 BYT2HEX.PRG 2459 XBOX.PRG 8599 BYTEAND.PRG 2787 YEAR.PRG 3171 BYTENEG.PRG 2147 NANFOR.RMK 18351 COMPLETE (real) complete sources of NANFORUM library available for everyone who wants to make ANY changes in sources. They are really free for sources. This files is absolutely no needed for to have NANFORUM library working (you need only file with .LIB !!), therefore take it only when you need it, or just want it. Finaly i was wondering if include complete list of files inside archive, but i found it useful for some cases. .DD. ----------------------------------------------------------------------------- ClipperNet - ACCESS.ARJ File Name: ACCESS.ARJ Other Names: File Size: 411 File Contents: ACCESS.PRG 1198 Source of TEST program for speed of Clipper. Very easy and showing only time needed for accessing variables in FOR cycles. .DD. ----------------------------------------------------------------------------- ClipperNet - ACHOO2.ARJ File Name: ACHOO2.ARJ Other Names: File Size: 6,851 bytes CLIPBBS 1-06 Page 35 4 Mar 1991 File Contents: ACHOOS.PRG 21976 ACHOICE "replacement" allowing combine achoice() features with direct GET, searching for whole menu line (not only one letter)... .DD. ----------------------------------------------------------------------------- CLIPBBS 1-06 Page 36 4 Mar 1991 ============================================================================= CLIPBBS ============================================================================= CLIPBBS Distribution CLIPBBS is special magazine about CLIPPER and CLIPPERing (or about another related problems and xBASE languages). This magazine is for free and articles aren't honored. Nobody can make a profit from the distribution of this magazine. CLIPBBS can be freely downloaded and uploaded to any BBS or any other public system without changes of original contents or number of files in original archive (kind of archive can be changed, but we are sup- porting ARJ archive because is best and smallest). If you are interested in CLIPBBS and would like to become a DISTRIBUTION site, contact publisher on 2:285/608@fidonet or 27:1331/4412@signet or just call to 31-10-4157141 (BBS, working 18:00->08:00, 2400bps) or voice to 31-10-4843870 in both cases asking for DANIEL (Docekal). Distribution sites: United Kingdom Welsh Wizard, SYSOP Dave Wall, phone 44-656-79477 United States of America The Southern Clipper, SYSOP Jerry Pults, phone 1-405-789-2078 ----------------------------------------------------------------------------- CLIPBBS 1-06 Page 37 4 Mar 1991 How to write articles in CLIPBBS? Submission of articles to CLIPBBS is really easy: Maximum of 78 characters per line, as long or as short as you like ASCII text. Choose from the list of extension which most describes your text, or just name it .ART as ARTicle and send it to publisher or to any distribution site via modem to BBS or with mailer as file attach. Article will come automatically appear in the next free issue. Extensions are: Articles (anything) .ART Software .SOF News .NEW Question and Answers .Q&A ANOMALIES and their comments .ANO Letters to editors .LET Advertisement .ADV Wanted .WAN Comments .CMS DUMP from conferences .DMP Clipper Net .CLN That's all at the moment, there will probably be changes later, as the magazine evolves. If you have any ideas for a new section of CLIPBBS, please tell us, or just write an article about it. Daniel, publisher -----------------------------------------------------------------------------