PC Online 1996 February

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

/ PC Online 1996 February / PC02_96.ISO / software / ever / tips / cfosd / run / whatsnew < prev

Wrap

Text File | 1995-12-29 | 27.1 KB | 754 lines

What's new? 'cFos' ISDN CAPI MultiPort FOSSIL driver ---------------------------------------------------------------- Bitte die folgende Liste der Aenderungen/Erweitungen aufmerksam durchlesen, da sich zu frueheren Versionen einige Punkte geaendert haben. Um es etwas einfacher zu machen, eine kleine Legende: ! Software-Aenderung erfordert Konfigurationsaenderung * Neues Feature x Bugfix - Information Revision history: ---------------------------------------------- cFos/2 0.42 Beta 671, 29-Dec-1995 x I introduced a new bug in beta 654, which caused a system crash at least with TELES boards. This bug has been fixed. It seems that after all I was able to fix a bunch of problems with key-files. Thanks a very big bunch to Heinz-Otto Scheid, Wim Abels and Patrick Bosch for their neverending patience with me. If there are still bugs out which cause cFos/2 to crash the system or to complain about the key-file, please write eMail to me. cFos/DOS 1.94, 28-Dec-1995 cFos/2 0.42 Beta 654, 28-Dec-1995 cFos/Win 0.80 Beta, 28-Dec-1995 x Bug fixed which caused CRC errors with V42b x ATB6 bug fixed cFos/Win * cFos/Win now supports the registration key file. In the [cFos] section in SYSTEM.INI you can insert the line key=<path>\cfos.key, so that cFos can find the key-file. The key file removes the time limitation an enables all features known of cFos/DOS Professional, such as cFos Channel Bundling and V.42bis data compression. (-jd data dump and time sync are not yet supported) * "CAPI unaccessable" error message splitted into several error messages. - WHATSNEW.WIN no longer supported cFos/DOS x Added detection of V.110 protocol and modem support for dual mode CAPI's (1.1 and 2.0) x Hopefully fixed the "executable corrupted" bug. x Again: fixed a bug in the deallocation of memory. x Bug fixed, which caused memory corruption in rare cases cFos/2 x Bug fixed, which caused the system to crash in V.42b mode with an quite unspecific error message. x -ks switch always set, thus 'cFos/2' will NOT deregister from the CAPI at shutdown. Some CAPI's are quite unstable when the system goes down. cFos/DOS, 1.93 Beta, 14-Nov-1995 cFos/2 0.42 Beta 508, 14-Nov-1995 cFos/Win 0.78 Beta, 14-Nov-1995 (see WHATSNEW.WIN) * cFos now comes with an installation and configuration program, featuring easy update and electronic ordering. * cFos and cFosN may be both used to deinstall the other without those nasty "could not free segment" errors. * AT&K modem command can be used for inband negotiation of X.75 frame-length and window-size, even without V.42bis. * V.42bis data compression now runs with cFos Channel Bundling. We achieved transfer rates of up to 35kcps using two channels (on the FIDO nodelist file)! * New statusline "LED": ...C-B3 ACOD* 0R:0000... The asterisk indicates the V.42bis mode/negotiation: left arrow V.42bis on reception right arrow V.42bis on transmission arrow in 2 direct. V.42bis on rx and tx dot cFos inband negotiation x Fixed problem with re-registering with some CAPIs. x Fixed a bug in memory allocation in CFOSN.EXE. x Fixed a severe bug (introduced after 1.05) with tx-buffer purging. x Fixed a bug in the rejection of calls with no caller-id x Tons of minor bugfixes x Major bug in inband negotiation fixed. x FOSSDUMP bug fixed. - cFos, cFosN and cFosCFG are now distributed in LZEXE format and are thus a bit smaller. cFos/2 0.42 Beta 504, 12-Nov-1995 x Fixed a bug that caused cFos to crash the system if not loaded with -ja. cFos/DOS, 1.91 Beta, 20-Oct-1995 x Fixed some bugs in the V.42bis code, fixed typos and enhanced some passages in the documentation. cFos/DOS, 1.90 Beta, 19-Oct-1995 ! WE CHANGED OUR DISTRIBUTION AND PRICES, SEE ORDER.DOC, ORDER.ENG {++} * 'cFos' has now V.42bis data compression ! When transfering uncompressed data you can achieve 16000 cps and more ! To enable V.42bis data compression you can load 'cFos' with -4 [1|2|3] [r|t|b] switch. | | | +--- direction (r=receiving, t=transmitting, | b=both. If you use V.42bis only in one | direction, you can save memory, default is | both. | +--- compression (low, medium, high). The higher the compression, the more memory & computing power is needed, default is low. Example: cFos i -4 loads 'cFos' with bi-directional V.42bis with low compression. To enable V.42b negotiation also set AT&K1. We recommend at least a 486 class computer for good performance. Note: 'cFos' V.42bis data compression is currently in beta state and can currently not be used with CCB. Under some circumstances it can produce CRC errors. In the next version we will fix these limitations. To circumvent the usual memory problems under DOS, we plan to release a DPMS version in the future. This version will need about 15 kb of DOS memory, the remaining memory will be allocated above 1mb. {++} * 'cFos' can emit an audible RING if you set Register S10.9=1, even if the RING count is less than S48. * Since ISDN charges seem to cause problems NO CARRIER/CHARGE=xxx messages are also issued in the shareware version. * ATH1 set 'cFos' in off-hook state. This will reject all calls until you issue ATH0. * The help screens displayed by AT? can be scolled one line at a time by pressing RETURN. All other keys display the next page. * After ATD dial commands exactly one CR or LF char are ignored for compatibiliy with UNIX dial scripts. * Changes of Registers S13 and S14 did not issue LISTEN_REQ message. Fixed. * Register S10.5 is obsolete. To enable padding '0's in the Caller-ID use the -j0 switch. - Changed default RX-Buffer size to 2048 bytes. cFos 1.05f, 06-Oct-1995 {++} * 'cFos' has a full featured charge counter! The ISDN charging info can be used to calculate the costs of a call, if Register S70 contains the costs of one charging unit (e.g. Pfennig, Groschen, Cent,...). Register S71 contains the amount of money for the last call in Pfennig, Groschen, ... Register S72 contains the amount of money for the last call in DM, Schilling, Dollars. The costs are also diplayed in the status line. The accumulated costs are written into the PROFILE. AT&G0 or AT&G can be used to display the accumulated costs. AT&G1 reset this charging counter (the accumulated costs). Note: Some CAPIs don't deliver ISDN charging info messages, so that 'cFos' can not evaluate them. Use ATI5 to check this if you get no charging info. * In ATI2 display, the disconnect reason is given in a more specific way. This helps debugging carrier-loss problems. * S10.8=1 rejects calls with no caller-id x Small bug fix for BTX mode x CallerID is displayed even when containing # or * characters cFos/2 0.42 Beta 459, 31-Aug-1995 [OS/2 only] ! We accidentially dropped the possibility of loading 'cFos/2' with more than one COMport in Beta 455. Now this is possible again. * The SPV support should work now. cFos/2 0.42 Beta 455, 30-Aug-1995 [All platforms] {++} ! Just to remind you: support for the analogue modem features of ISDN boards and the VT-100 BTX support are 'cFos/PRO' features. {++} * Tested the BinTec BIANCA board. Added support for the analogue modem emulation of it. [OS/2 only] ! If you have problems with lost connections when shelling out to other applications load 'cFos/2' with -kx and if this does not help with -kc as well. CHECK CAREFULLY THAT YOUR CONNECTIONS ARE PROPERLY TERMINATED if you set this switch! ! Changed the defaults for receive and transmitter buffer sizes. They are now (windowsize+2)*blocksize, i.e. we increased them by <blocksize> bytes. This should give good performance with almost all applications. * Changed the behaviour of the -kc switch a littlebit: now it disables changing the status of DTR on open or close function calls. {++} * Analogue modem support should now run ok under OS/2. * Added a new switch: -kx. If set, 'cFos/2' will mimic the behaviour of SIO, which is (as far as i know now) not always exactly the same as documented in the IBM refs. E.g. PCBoard/2 needs this switch. This MAY fix hangup problems with OS/2 programs shelling to DOS applications. * BTP EE-32 did not send with full speed with 'cFos/2'. After some debugging we implemented a little workaround and now you get the FULL speed: 7880 cps (with testdata). x InterMail mailer uploads are speedy now; the slow uploads from the terminal and the CRC errors still persist. x IOCTL Cat 1, Function 0x65 now correctly sets Bit 0 if write requests are blocked. cFos/2 0.42 Beta 360, 03-Aug-1995 [All platforms] x Fixed some CCB bugs, which caused transmission to stop when channels were removed. * We added an ATI5 display to show the number of INFO_IND's with date/time info and chargeinfo that 'cFos' received. This display is only available in the debug versions of 'cFos', i.e. not in CFOSN.EXE. * 'cFos' now accepts AT commands not only after a carriage return, but after any preceeding characters. [DOS only] x Fixed a bug that caused 'cFos' to issue an "Out of ports" if you tried to load it with 4 ports. x Maybe fixed some more problems with UMB allocation. x Fixed a bug in the appendage function 0x05 (get ISDN time). * The optimization of 'cFos'-buffers into the UMB blocks is now aborted after a specific number of tries, even if not all possibilities were checked. With 4 ports loaded on a iDX4-100, 'cFos' would need at least 20 seconds, maybe more (i reset my system then). Now it should be ready after half a second maximum. * There is a new ISDN appendage function (function 0x06), see APPEND.DOC for for info. * If 'cFos' is loaded and the -d switch is used, the tracefile name is displayed to show that dumping is active. * The ISDN appendage is now available in every version of 'cFos', i.e. even in the 'cFos/Plus' and shareware version. Only the 'add/remove channel' function (function 0x04) is restricted to the 'cFos/Pro' version. [OS/2 only] x Fixed a bug that caused 'cFos/2' to behave strangely if e.g. a "ECHO ATD....>COM3" command was issued. x Fixed a bug that inhibited OS/2 to change to the root directory if the -a switch was set. * Added the -kc switch to prevent 'cFos/2' from closing the connection if you do a first level close while the carrier is still up. * 'cFos/2' is working up to October 31st, 1995. * Enhanced the maximum number of ports to 8. Fixed a bug that crashed the system when the maximum number of ports were used. cFos/2 0.42 Beta 322, 26-Jun-1995 [DOS only] x Fixed a bug that caused 'cFos' not to reject incoming calls properly if no more ports were available. [OS/2 only] x Fixed a bug that cause 'cFos/2' to crash if a call came in and the port was already in use by an active connection. x Fixed a bug that the statuslines of different ports were displayed all in row 0 of the alternate display. cFos/2 0.42 Beta 318, 02-Jun-1995 [OS/2 only] * Now the following appication should run ok with 'cFos/2': Warp Internet Access Kit, Cantaloup/2 Beta 0.73, UUGate's uucico 0.33 (see readme.os2), CKermit 5A(190). x Implemented -jy for e.g. uugate's uucico. x Fixed a bug in ioctl cat1, function 0x64 (query com status). This prevented CKermit from working with 'cFos/2'. cFos/2 0.42 Beta 313, 26-May-1995 [OS/2 only] ! This version runs until July 31st, 1995 * Changed some more stuff in the set_dcb() routine. MainDoor/2 tested and worked (after some changes) ok. x One more fix in the set_dcb() function. This should (hopefully!) finally fix the problems with Xenia/2. [All platforms] x Fixed a bug in AT&L. x Fixed a bug that cause every call to be auto-answered. cFos/2 0.42 Beta 308, 22-May-1995 [DOS only] {++} * CFOS STATUSLINE:<port>:-1" switches the statusline off for port <port>. * If you select -e2 on 'cFos' startup, 'cFos' will not only disable the FOSSIL functions, but additionally it will overwrite the FOSSIL signature in memory. * Added switch -jf to ignore the flush_output() FOSSIL function. This may be useful e.g. for QuickBBS, which sometimes sends a flush_output() after each character sent. {++} * CFOS STATUSLINE:<port>:<line>" sets the statusline number for port <port> to line <line>. * Added a -jh switch to ignore the high byte of DX in FOSSIL function calls that use DX for the portnumber. Some programs (e.g. D'Bridge) trash the upper part of DX. Use this switch if your program works without a -c switch, but not if you specify one. x Hopefully finally fixed the "Out of memory" bug. [OS/2 only] x Fixed a bug that prevented Xenia from dropping DTR. ! The WHATSNEW.OS2 file is discontiniued and incorporated into the WHATSNEW file of cFos. [All platforms] * 'cFos' now ignores any modem input if DTR is low. This is for enhaced modem compatibility. If you have problems with that, load 'cFos' with the -jy switch. Note that this may cause problems when dropping the carrier. {++} * You may now switch to VT-100 BTX mode if you set ATB6 before calling BTX. This sends the VT-100 switch sequence to the BTX network just at the beginning of the session. * To make things easier for you, you may now select the services you want to listen to with &I: 'A' selects analogue calls (si=2 and si=4) and 'D' selects data calls (si=0x80). Use &IAD to listen to both of them. * You may now suppress the first <n> RING messages, where n is the value you have to write into register S48. Furthermore you may now specify in S0 the number of RINGs after that 'cFos' will automatically answer the phone (before that 'cFos' autoanswered the phone at the first RING, regardless of the S0 setting, if S0 was > 0). E.g., if you set S48 to 3 and S0 to 2, 'cFos' will not sent any RING messages, but will connect to the other side after 2 incoming RINGs (i.e. after 8 seconds). * 'cFos' now rejects incoming calls when doing outgoing calls. This results in less NO DIALTONE modem messages. {++} * New register S44 to select the way 'cFos' handles incoming calls with service "telephony" and "a/b services". If this register is set to 0, 'cFos' selects bittransparent and issues a "CONNECT VOICE" message as usual. If this register is set to 1, 'cFos' selects the modem protocols supported by certain manufacturers (AVM, ITK, ELSA) and issues a "CONNECT 14400" message (independant of the actual baud rate). {++} * ATB7 selects the modem protocols supported by certain manufacturers (AVM, ITK, ELSA) for outgoing calls. Note that we are not able to check if the protocols are really implemented. We just select the right parameters according to the manufacturer. {+} * In Multiport Mode, you can now specify EAZs and services, 'cFos' should listen to, for each port separately. Overlapping settings result in automatic ringdown as before. {++} * If bit 7 of register S10 is set (S10.7=1) then a CONNECT string with the bausrate specified in S47 will be displayed. * If a dialattempt was aborted before the CONNECT message was sent and charge for this call was reported by the ISDN network a CONNECT message will be displayed and followed immediately by a NO CARRIER message. This is e.g. for mailer-software to notice that a unit for this call was charged. * When listening to "a/b services" 'cFos' now treats these calls like "telephony" calls, i.e. selects protocol bittransparent and issues CONNECT VOICE message. * Added some more registers. To display the registers, use "AT&V<x>", where x=0 shows registers 0 to 79, x=1 shows registers 80 to 159 and so on. * New design of the debugscreen. Maybe this clears up the mess a littlebit. x Some minor fixes for the handling of "broken" CAPIs. {++} x Some cFos Channel Bundling (CCB) fixes, when using dynamic channel bundling. x Bugfix in ATI4 message dump. x some minor BTX bugfixes 1.05b, 16-Jan-1995 {++} * Within the first charging unit, 'cFos' now displays the number of seconds the connection is already running at the right hand side of the status bar. With the begin of the second unit, it will count the seconds down. {++} * 'cFos' now supports ISDN leased lines as they are supported at least by the TELES CAPI. Use ATDF to start such a connection. Attention! If you don't have a leased line your machine is likely to hang! - Changed the BIOS modem and line status bits a littlebit to make them more modem-like. {+} - The TIES hangup-string can now be 11 chars long ! Register s6 now specifies the column of the callerid display (if at&d1, at&d2 and s10.1 == 1), default 61. {++} ! Bit 5 of Register 10 is no longer used to activate the Caller-ID repair. Use the -j0 switch instead. 1.05, 15-Dec-1994 {++} * Register S69 can be used to specify a delay between two dynamic CCB tries. A value can be specified in seconds. * AT&Cn will be ignored for compatiblity reasons {+} * The +++-sequence hangup string can now be modified with a "AT&H<String>" command. This sets the hangup string to "+++<String><cr>". E.g. this enables setting the hangup sequence to "+++ATHE1<cr>" for the pcANYWHERE "Hayes compatible modem". {++} x minor bug fix in dynamic CCB. ' cFos' sometimes did no dynamic CCB if register S0 was not 0. {++} x user appendage function 4 (add/remove channel) did not work correctly with dynamic CCB. fixed. it also ignores the CCB tries (S68) and AT&Bn settings. x Numerical result codes in MODEM.DOC added. x bug in caller ID evaluation fixed. x "cfos {eaz!callerid!bps}" now work even with COMMAND.COM (COMMAND is not compatible with the MS specs on environment blocks). x The phantom "segment at ... destroyed" error message fixed. 1.04d, 12-Dec-1994 x Fixed one more bug in the date routines and enhanced the NetWare support with some more security. 1.04c, 22-Nov-1994 * 'cFos' now handles garbled caller id's as "unknown" {++} * FOSSIL dump function (-jd) does now dump data in ASCII as well. x FOSSIL function 0x14 (enable/disable watchdog) could not be used when FOSSIL was not init'ed. Fixed. {++} x The appendage function 0x03 with BX == 0xffff didn't work as expected. Fixed. x Fixed 1 source typo. 1.04, 27-Oct-1994 * We now offer a special Christmas Deal for a limited time! For more information see XMAS.DOC and XMAS.ENG {++} * There is a smaller 'cFos' executable available, named CFOSN.EXE, which uses about 5kb less resident memory. This version does not contain any debug code, no trace and/or dumping functions (no -d[d], no -jd, no -jm). It is only functional if you have a 'cFos' registration key. {++} * Slight extensions in the ISDN appendage, see APPEND.DOC. * MODEM.DOC updated. {++} * Appendage function 2 can now be called without carrier. * Detailed chapter in CFOS.DOC/ENG about CompuServe's DosCIM/ WinCIM. x Fixed an odd behaviour in the timer interrupt code, that may have not called the rest of the applications that hooked INT 1C. {++} x Fixed a bug in the "cfos t" code, that caused cFos to ignore all tranx requests made. x Fixed a bug in ISDN appendage if 'cFos' was loaded with -c. 1.03c, 12-Oct-1994 * From this 'cFos' version on, 'cFos' is compiled with a new C compiler, resulting in a saving of about 6kb resident memory. * New switch: -jk to enable set_baudrate hangup; the hangup on INT 14 function 0 in INT 14 mode is now off by default. * New command: "cfos init:<port>". * leading plusses '+' in front of an AT command are now ignored. * Documentation updated and expanded; listed a lot of new products in software/hardware compatibility lists. * Support for Gerdes PowerISDN. {++} x Fixed a bug in charge limited calls. x Fixed a bug in the allocation of DOS UMB's (.e.g under EMM386). {++} x Fixed some problems when writing to the master environment. Added some code to ensure that the system won't hang is cases where it wasn't possible to write to the master env. x Fixed the HS/Link hangup problem. {++} x Tried another fix to catch the time-sync problem. x ATBx commands now set s22, s26 and s28 to defined values. If You like these registers to be modified, be sure to modify them *after* each ATBx command. x FOSSIL func 1c maybe didn't work correctly. Fixed. - Small changes in the ISDN appendage. Function 3 may be called even without carrier. {++} - -m90 is now default, i.e. the maximum difference of the ISDN time and your systems time may be 90 minutes {++} - No-debug version of 'cFos' will require a keyfile - ATH now does the same as ATH0 - RING and RINGING messages now are sent with a delay of 4 seconds 1.03, 26-Aug-1994 {++} * 'cFos' hat jetzt eine ISDN FOSSIL appendage. Autoren von Applikationsprogrammen koennen jetzt einige zusaetzliche Infos abfragen. Siehe APPEND.DOC. {++} * Es gibt eine neue Disconnect Reason, Nr. 8: Appendage Reset, wenn eine Verbindung durch eine Funktion 9001 ausgeloest wurde. * Sollten mehrere erweiterte RING Meldungen Probleme machen, so kann man mittels S10.6=1 einstellen, dass nur beim ersten RING die erweiterten Infos geschickt werden. {++} * CFOS EAZ Kommando: 'cFos' gibt waehrend einer Verbindung in die Environmentvariable EAZ die angerufene EAZ aus. Fuer Batchfiles. {++} * Register S12 kann jetzt benutzt werden, um einzustellen, wie lange (in 1/18 sec. Schritten) die CONNECT Meldung nach erfolgreichem Verbindungsaufbau verzoegert wird. Default ist 50. * Noch mehr "safety code" beim Verbindungsaufbau... Versionen davor: - die komplette Revision History kann bei Zaphods BBS in der Filearea ISDN downgeloaded werden. ------------------------ End of file ---------------------------