FDChk - The FAST DMS Checker Adder 'n Remover! Version 2.3 Programmed by SiNTAX/TsL 1. Introduction --------------- An again... a new update... bringing you more speed and hopefully less bugs! My major concern for this release was, to get the bugs fixed. Although there weren't that many serious once, only found 1 actually. The version 2.2 didn't detect Short Reads anymore, this one does... but... be careful, cuz' it is VERY strict about the filesize... What I mean is, if there is 1 extra byte after the DMS file, that shouldn't be there, the dms checker will say it is a corrupt dms (which is ofcoz correct! We don't want some lame trader to get extra credz just by adding some 0's to the dms file! Do we!) But, be careful cuz' DMS itself won't see it as an error. So it might happen that FDChk detects an error, and when ya test it with DMS, it will say it is OKAY. 2. But first ------------ CHECK The files in this archive to avoid nasty backdoored patches of this production, although this is no 100% guarantee, it sure is safer than just COPY 'n RUN... Now let's put it more formal: The utility 'CRC' (included in this archive) calculates a so-called `WRAP' checksum. For all the files in this distribution, this checksum MUST be zero. If this is not correct then delete the file immediately because then you know that it has been modified! FDChk23_68000 (CRC = $75E8D627, Size = 4496 Bytes) FDChk23_68010 (CRC = $9E1C7FF2, Size = 4496 Bytes) FDChk23_68020 (CRC = $DC30CDE3, Size = 4528 Bytes) CRC (CRC = $4F5177B9, Size = 4104 Bytes) FDChk.DOC (Size = 28881 Bytes) InfoScreen.FDC (Size = 394 Bytes) AddText.FDC (Size = 2620 Bytes) DMS (Size = 27 Bytes) UL_MSG.REXX (Size = 1805 Bytes) SetEnv (Size = 448 Bytes) GetEnv (Size = 336 Bytes) Although I included an internal check to see if somebody has tempered with my programs, you better check out these values to be 100% sure that the programs are in their original form! 3. The USe ---------- o Novice-Level First of all make sure you have the ARP.LIBRARY in your LIBS: directory, otherwise you won't even be able to start the program! Secondly, ENV: must be assigned, else you'll get a nice requester asking you to insert volume ENV... That's about all you HAVE to do to start. o Profi-Level Now I'll describe why this isn't just another DMS Checker, Adder 'n Remover. - Adding DMS texts FDChk will look for a text file called 'S:AddText.FDC' and will append it to the end of a DMS file. This file may include some special %-sequences to instruct the program to insert some information about the dms file itself. These %-sequences will be described later on. NOTE: The text file will be truncated if it's larger than 20.000 bytes NOTE2: to define multiple texts, read the section about "Configurating FDChk" - Adding a customized look You are able to add an InfoScreen to the program. Just create a file called 'S:InfoScreen.FDC' and include all the stuff you want, ofcourse you are also able to use the same %-sequences in this text. This info screen will be shown to the standard output channel, so you can use it as some kind of info-feedback to the user. NOTE: to define multiple texts, read the section about "Configurating FDChk" - Configurating FDChk The program is guided by means of Environment strings. So it's fairly easy to change the fucntions of the program. All you need are 2 cli programs called 'SetEnv' and 'GetEnv' although you don't really need the last one. To change an option just issue the following in your cli 1> SetEnv FD_ The available environment strings are: FD_BUF : Defines the Max amount of memory FDChk should use as buffer to read in the DMS File Default: '901120' (880Kb) FD_CFG : Configuration for FDChk (described below) FD_UMSG : A message from the user that can be included in the InfoScreen and AddText! Max size 70 chars, if the text is longer then the default will be used. Default: 'UNKNOWN' FD_DATF : Define date format (see further down for more info) Default: 'DOS' FD_CNTINF: Number of InfoScreens you have made, and wish to use FD_CNTADD: Number of AddTexts you have made, and wish to use NOTE: ¹) Max value is 100 ²) When you use the multitexts then the texts must be named: 'S:InfoScreen.FDC.xx' 'S:AddText.FDC.xx' Where 'xx' represents a decimal number starting from '00' upto FD_CNT???-1 FD_UMSG ends in '.X' where X is the node number it is used for so FD_UMSG.0 is for node 0.. etc.. FD_CFG can contain the following switches: 'A' : Add DMS Text Ad 'D' : Disable DMS Texts that are present in the source file 'R' : Remove DMS Texts (slower than disabling) The difference between disabling and removing is that remove will REALLY remove the text ads from the DMS file where disable will only disable the displaying of the texts 'O' : Enable displaying of the strings 'TEXTS DISABLED' and 'TEXTS REMOVED' So if you want to add your text and Remove the other texts, you should write SETENV FD_CFG "AR" FD_DATF can be set to the following strings: 'DOS' : AmigaDOS format (dd-mmm-yy) 'INT' : International format (yy-mmm-dd) 'USA' : American format (mm-dd-yy) 'CDN' : Canadian format (dd-mm-yy) 4. Control sequences -------------------- The format of the control sequence is: %[flag][width.limit][option] Where: [flag] can be a '-' to use left alignment [width] field width. If the first char is a '0', the field will be padded with leading 0's . must follow the width if you want to use the limit option [limit] maximum number of chars to output from a string [option] the actual option to use, ranges from 'A' to 'O' using a char out of this range, will just result in the text after the % being displayed See below. If you want to use a '%' in your texts you need to type '%%' so that the program will know you don't want to use a control sequence. FDChk knows the following options: A - Name of the day when file was UL'ed (e.g. Monday) B - Date when the file was UL'ed C - Time when the file was UL'ed D - Low track used in archive (as in information given by DMS when using the VIEW option) E - High track F - Packed Bytes G - Unpacked Bytes H - Serial No. of creator I - CPU of the creator (68000, 68010, 68020, 68030, 68040, 68050, 8086, 8088, 80188, 80186, 80286, 80386, 80386SX, 80486, 80586) J - CO Processor of the creator (68881, 68882, 8087, 80287, 80387SX, 80387) K - Computer Type of the creator (AMIGA, PC/CLONE, ATARI ST, MACINTOSH) L - Compression mode used to make the archive (NOCOMP, SIMPLE, QUICK, MEDIUM, DEEP, HEAVY1 - HEAVY5) M - Filesystem (AMIGA OFS, AMIGA FFS, MSDOS, AMAX, MAC) N - Name of user who uploaded the file (should be used in conjunction with the ARexx Door) O - Message from the uploader (dito) 5. A word about Multi-Nodes --------------------------- Finding a suitable way to get the node number on a multi node system needed some thinking, because it had to be a fast and memory friendly way. So I decided to do the following: FDChk will scan it's parameter for the word 'NodeX' in the pathname. If the scan is succesful, it will use the node number found in the path. This way you will be able to use the external file checker protocol from AmiX, without having to do some patch work! Just add the supplied (in this archive) DMS FCheck entry to your BBS:FCheck dir and off you go. Now this methode works fine as long as you use the external file check protocol of AmiX, but when you have some 3rd party util, it won't always be so easy to use this methode... so I included another way to pass the node number to FDChk. Just enter a -X to the parameter, where X is the actual node number for which FDChk is doing the checking. This will allow you to use the util in conjuntion with utils like OffLine Checker etc. All you need to do, is make seperate logoff scripts for the nodes. For the Pseudo-ops OffLine Checker you would for instance create a library file for each node, in which you put the program name and the parameter, which is then - Here's an example of my LibFile for the OffLine Checker: 00000000|444D5321|FDChk|-0 %s| ^ ^ | Node number Program name 6. Available version -------------------- FDChk comes in 3 versions, each optimised for their processor, and I mean REALLY optimised NOT some Lame-C like optimising! So be smart and use the version for your CPU! Ohwell you'll have to do that anyway, coz' they will only work 100% on the CPU they were developped for! This has something to do with priviledged instructions and shit, but then WHO CARES?!? '030 users should use the '020 version, which is actually also made for '030 processors. BUT as I don't have an '030 or even an '020, I was unable to test them thoroughly. (HINT: accelerator card is very welcome! hoho) 7. Speed Comparison ------------------- File 'EDK-POWE.DMS' size 906493 HSCHK V4.19B | DMS V1.11 | FDChk V1.2 (68000 version) --------------------------------------------------------------------- Checking| 13.160 seconds | 34.680 seconds | 12.120 seconds (BUF=300000) | | | | 12.380 seconds (BUF=100000) | | | | 11.820 seconds (BUF=901120) | Disable | not possible | not possible | 12.000 seconds (BUF=901120) | Remove | not possible | not possible | 14.880 seconds (BUF=901120) | --------------------------------------------------------------------- | NOTE: FDChk V1.3 does the same checking in a mere 8 seconds!!!!! | --------------------------------------------------------------------- Well as you can see FDChk beats the shit out of the others! Even with a small buffer, FDChk is able to archive high speeds! Ofcourse the lower the buffer, the more diskaccess you get. So with a small buffer, the speed much depends on how fast your HD is. 8. The ARexx doors ------------------ If you want to use the %N (UploaderName) and/or %O (UserMsg) option, then you have to install the supplied ARexx Doors. The first door 'LUser.REXX' sets the FD_UNAME.x environment with the name of the user who just called. You need at least version 2 of AmiX to use the door, because the door should be called just after the user logged in. This can be done with BULL.AXM (AmiX V2.05) or with the MCI Commands (AmiX V2.08+). For AmiX V2.05 you should add the following to your BULL.AXM file: *LU RM030DOORS:FDChk/LUser.rexx The name of the door itself may be changed to something else. It doesn't matter what it's called. In AmiX V2.08+ the BULL.AXM has been replaced by the MCI Commands, so you should make a file called BBS:BULL.TXT and put the following in it: ~CC_LU| The door definition has to be present in your CustomCommands file, and it should be defined as above (*LU RM030DOORS:FDChk/LUser.rexx). Now when a user log's in, the door will be executed and the user name will be saved to the right environment string. The second door is used to set the User Message. To install it, add the following line to your CustomCommands file: *U RM030DOORS:FDChk/UL_MSG.rexx Now when a user selects 'U' to upload, he will be asked if he wants to add a one-liner to the DMS Text Ad. After that, the FD_UMSG.x environment will be set and the normal 'U' command will be executed. (NOTE: the UL_MSG door accepts a commandline parameter 'N' or 'Y' 'Y' for adding a one-liner and 'N' for no one-liner) You can also change the minimum access level a user needs to be able to set a one-liner. To change this, load the door and change the MINSEC variable to the level you want. (NOTE: the FD_UNAME environment has been replaced by STATS@, so that ya don't need to run the LUser.rexx door anymore!) 9. History ---------- Version 1.0 o First release of this wonderful DMS checker and text adder (NOTE: This version was released accidently due to a human error, mine) Version 1.1 o Improved commandline handling o FULL PC-Relative code o Removed the EGO-TRIP message (you know, the one that sticked to all your dms files) o Bug fixed (Crashed when not enough memory was available!) o The 68020 version now really works! There was a bug in the processor identification routine that caused the program to think you only had an MC68000! (sorry about this one.. but I don't have a turbo card so...) o Internal checksum added, to stop some lame name changers! (note to those lamers: I have put time and effort in this program, so I think I deserve some credits for it.. and if you don't agree then go and code your own tool!!!!!!) NOTE to the honest users: check the CRC value of your version, because the program won't tell you if a lamer has tempered with it.. so DO it, OR suffer the consequences! o Now really worx with Ami Express V1.1V !!!!! All faulty DMS's will go to the HOLD dir !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! o Just received Ami-Exp V1.1w... I guess the previous improvement doesn't matter that much anymore coz' you can now configure AE yourself. There's also an example AE-Checker file included, to make it easier to install the tool. NOTE: you don't have to change AE with NEWZAP anymore from 1.1W on, just use the external checker ability o Also included an 68010 version (read 4. for more info) NOTE: I didn't have a chance to test this, but it SHOULD work OK. It's practically the same as the 68000 version, only 2 instructions are different SO what can go wrong?!?!?! Version 1.2 o Added MaxMem to limit the size of the buffer for the checking o Added ability to check a dms file even if there isn't enough memory available to read it totally in o Added a VERY Fast methode to kill the existing DMS Texts o Bug fixed: Length of text now set correctly in DMS File (Copied the upper word, instead of the lower word!) o Some peephole optimising done o Added new % options: %xxN - The user who uploaded the file %xxO - Message from a user (Max. 70 chars long!) (NOTE: Requires the arexx door to be installed!) o Added Environment support. If the environment isn't found then the defaults will be used. Current environment vars are: FD_BUF : Defines the Max amount of memory FDChk should use as buffer to read in the DMS File Default: '901120' (880Kb) FD_CFG : Configuration for FDChk (described below) FD_UNAME : Name of the user who uploaded the file Use this to be able to include the name of the uploader in your text. I selected this methode of getting the uploader name, rather than scanning the logfile because it is a hell of a lot faster! Default: 'UNKNOWN' FD_UMSG : A message from the user that can be included in the InfoScreen and AddText! Max size 70 chars, if the text is longer then the default will be used. Default: 'UNKNOWN' FD_DATF : Define date format (see further down for more info) Default: 'DOS' o Fixed a bug (wrote 4 bytes to a var that was only 3 bytes long!) o Rewrote the sloppy % format routine. The new routine is more user- friendly! The routine offers: - an OPTIONAL length indicator, so you don't have to add the 00 anymore. If you need to align something, then just put the size indicator between the % and the option No need to make it 2 digits. NOTE: don't use extreme values!! (Like 80+) - right aligning!! Mucho nicer! - an error detection routine, so a %Z won't crash the computer anymore - a possibility to put a % in the text, just enter %% - with digits you can also fill up with zeros, just enter a 0 before the size (e.g. if %6F would give ' 7834' then %06F would give '007834' nice hey!) - optional left aligning by adding a - after the % - text limiting by entering a .limit after the size where limit is the max amount of chars to copy NOTE: This is only valid for text!! When Adding it to a number option, the routine will just skip it. This to be able to use it with the Serial No. option which can be a number or a string. So in short: %[flag][width.limit][option] Where: [flag] can be a '-' to use left alignment [width] field width. If the first char is a '0', the field will be padded with leading 0's . must follow the width if you want to use the limit option [limit] maximum number of chars to output from a string [option] the actual option to use, ranges from 'A' to 'O' using a char out of this range, will just result in the text after the % being displayed o Improved speed of slow scan (removed the _LVOSeek) o Available strings for FD_DATF environment: (case-insensitive) 'DOS' : AmigaDOS format (dd-mmm-yy) 'INT' : International format (yy-mmm-dd) 'USA' : American format (mm-dd-yy) 'CDN' : Canadian format (dd-mm-yy) The default is 'DOS'. The default will also be used when there is an invalid string in FD_DATF. o Configuration setup for FD_CFG: 'A' : Add text to DMS file 'D' : Disable other DMS Texts e.g. if you want to Add & Remove text ads then set FD_CFG to 'AD' o Speeded up slow scan a bit more (faster move routine) o Made the util multi-node compatible. The environment strings changed a bit to make it possible: FD_UNAME -> FD_UNAME. FD_UMSG -> FD_UMSG. Where is ofcourse the node number ranging from 0 to 9. I used the following methode to determin from which node the util was started: FDChk will scan the parameter for '/NodeX/' or ':NodeX/' if this string is present in the program parameter then FDChk will use the X char as node number. This methode is the fastest and simplest way, and it will work with the way AmiX checks files (using the BBS:FCheck dir) o Added another methode to determin the node number, just add '-X' before the DMS filename. X must be in range 0-9. Using this way, you are also able to use FDChk with the pseudo-ops OffLine Checker, just create a library file for each node. (This methode overrides the previous one!) o Bug fixed in slow scan (odd address error in fastcopy routine) o New config option 'R' When set, the existing text ads will be removed from the source file. NOTE: This requires creation of a temporary file and thus results in slower speed! So it's up to you to determin whether you wanna trade in speed for a few bytes diskspace This option overrides the 'D' option. o Speeded up the checking routine a bit more! Version 1.2b o Fixed bug in the 68020 version, which caused it to crash... here's why: I accidently optimized a JSR Tab(PC,d1.w) instruction to JSR Tab(PC,d1.w*2) where it was not needed to Multiply D1! Ohwell nobody's perfect! o Some minor adjustments to the source (you won't notice anything) Version 1.3 o Made the program re-entrant,so now you can set the pure bit and make it resident!!! This is especially handy for Multi Node Systems because all the nodes can use the same code at the same time! No need to have XX Copies of the code in memory. o As a result of the previous change, the filesize reduced by almost 1Kb o Speed improved by 25%! Speed gain: approx. 3 seconds on a 900.000 bytes DMS file!!!! This increase in speed was achieved by replacing the CRC calculation routine, by a faster one. o New Config Option: (as requested by someone) 'O' - Enable output of 'TEXTS REMOVED' and 'TEXTS DISABLED' Default is disabled o Fixed a bug: The path for the temp file wasn't always correctly duplicated from the original file. The routine missed to copy a char when entering a path like 'DH0:Filename' o Fixed a little optical bug: when the DMS file contained an error, the 'TEXTS REMOVED' string still got printed o Now supports multiple InfoScreens & AddTexts upto 100 texts!!! To set these environments to the number of texts that are available: FD_CNTADD : Number of AddText.FDC files FD_CNTINF : Number of InfoScreen.FDC files If the program can't find the appropriate environment, then it will operate as it did before. But if you use multiple texts then they must be named as: 'AddText.FDC.xx' and/or 'InfoScreen.FDC.xx' where 'xx' is a number starting from '00' upto '99' NOTE: If the program can't open the file, then it not try to find another text that does exist... so be sure that all texts are present. NOTE2: Remember that the 'xx' starts at '00' so setting the counter to 2 means that you must have 2 textfiles, one ending in '.00' and one in '.01' o Changed the random routine a bit... now it uses the Vertical Raster Position and the DMS Info Header CRC to calculate a random number o Corrected a fuck up in the DOC file! The LIBRARY file for the Pseudo-Ops OffLineChecker should ofcourse be: 00000000|444D5321|FDChk|-x %s| Instead of: 00000000|444D5321|FDChk -x|%s| >>> Where x = Node Number Because otherwise OLC would look for a program called "FDChk -x" instead of loading "FDChk" and passing the commandline "-x %s" to it... Thanks to MR.SOLO for spotting this! o Fixed a bug: I changed the offset tables to word-size, to decrease the program size, but forgot to change the move-size to word. This had as effect that the next element was taken out of the table. Version 2.2 o Skipped to version 2.2 to avoid misunderstandings, because I accidently named the .LHA archive of the V1.2 FDChk21.LHA o Fixed a bug with the options (the DISABLING was alwayz ON!) o Changed the resultcode from -20 to 20 if the program can't allocate the memory for the dynamic variables. (dunno why in hell I wrote a minus sign in front of it... I must have been sleeping) NOTE: for all you guyz who like to write their own ADos Scripts, you can use the following keywords: ERROR : When there was an error during the check (e.g. outta mem, wrong dms, unable to lock file, ...) FAIL : When FDChk couldn't allocate it's variable memory So, you would write something like: FDChk -1 ;Check file upped on node 1 IF ERROR ECHO "File screwed up!" .... ELSE IF FAIL ECHO "Go buy some memory!" .... ENDIF ENDIF Got da picture? o Changed the name of the environment that held the user name, now it will read the environment called 'STATS@'x where x is the node number This way, the LUser.REXX door is obsolete when you use Ami Express! (The program will remove the status bytes from the environment and will removed the spaces between the '-' char and the last char of the name. e.g. 'John Doe -00' will be translate into 'John Doe') For SysOp uploads, I recommend creating a 'fake' STATS environment, instead of using the one of the private node, because when the node is not active, the environment will be empty... so create a STATS@9 file with your name, and use FDChk -9 o Rewrote the code that create the temp filename, because it wasn't always doing it's job 100% okay. The problem wasn't that severe tho, it only caused the checker to create a wrong temp filename... so nothing to write home about o Optimised the start code a little bit (you know.. use a TST.B instead of doing a TST.L.... that kinda stuff) o Tracked down a bug that would cause FDChk to use the number/character after the NODE string in the pathname.. so, the program would also accept BBS:NodeS/... and then change the node to check to 'S' (the check routine to find out if it was a number or not, didn't clear the char when it wasn't a number)... boy, this SOUNDS complicated o Whoops! Zapped a serious bug in the 68020+ version that would let Short Read errors in a DMS pass the check! Thx to WAREMONGER for telling me about this!! (I accidently wrote the 'BNE.S xxx' in the processor-conditional so it was only assembled in the 68000/68010 code... STUPID!) Version 2.3 o Bug fix: Using '-9' to indicate and upload to node 9 didn't work it could only handle upto '8' o Bug fix: File wasn't closed when there was an error in the dms file o Bug fix: SHORT READ Error's weren't always detected! o WONDERFUL! I just discovered MungWall!! Great debugging tool C= made! and even better news... FDChk passed the test! It didn't zap into forbidden memory area's (used PRE/POSTSIZE 64) Pitty I can't use enforcer to test my util (U know... 68000) o Optimised the check loop a BIT more: I Had the following codelines 2 times in the check loop: IFNE MC68020_CPU move.w $6(a4),d0 ELSE movep.w $6(a4),d0 ;16 cycli move.b $7(a4),d0 ;12 cycli ENDC ;==== Total: 28 cycli ;==== So: 28 cycli * 2 times = 56 cycli Replaced by a simple MOVE.W D4,D0 !! Saving some clock cycles.. I know it isn't much, but every bit helps. This one will have most effect on an 68000, because it doesn't have to execute the VERY time consuming MOVEP.W instruction (MOVE.W D4,D0 = 4 cycli!! Saving 52 cycli!) And since this is executed for every track, and there are usually 80 tracks in a DMS file... result: 80*52= 4160 cycli saved!!! o Managed to squeezed off a few bytes in the Add routine (+ Faster, but not worth mentioning) 10. Words of thank ------------------ To all people who motivated me to make this new release, to all people who brought me some useful ideas to incorporated in this update, and to all my friends! Now some personal thanks: ByteMaster - For keeping Ami Express up2date Relayer - For reporting the 'Short Read' bug Modem Man - dito Sigma Seven - Keep up da good work Ali Baba - dito OLI - thx & the entire /X Development team For more details contact me at my board Sin City - The Silents Belgian HeadQuarters (AMIGA ONLY!) Home of the SiLENTS Trainers! Node 0: +323-825-2274 (DUAL STANDARD) Node 1: NEVER! >X<>X<>X<>X<>X<>X<>X<>X<>X<>X<>X<>X<>X<>X<>X< >X< End of this LOOOONG BOOORING DOC File >X< >X<>X<>X<>X<>X<>X<>X<>X<>X<>X<>X<>X<>X<>X<>X<