home *** CD-ROM | disk | FTP | other *** search
- +---------------------------------+
- | BUGFIXES, CHANGES, and UPDATES. |
- +---------------------------------+
-
-
- CFX 5.102 - 03/Nov/91
-
- - Minor changes to some texts.
-
- - On registered versions (when I used the compiler's optimiser), a certain
- bug slipped past my eye. This 'bug' is really a feature of my
- global optimiser! A certain initialised auto-variable was being
- 'un-initialised' by the optimiser. This caused CFX 5.100 to 'stick' on a
- current-directory scan containing one file. I have now optimised CFX
- 5.101+ manually, and no longer use the global-optimiser for finished
- releases.
- NOTE: This was apparent *only* in registered versions.
-
- BOB
-
- /*------------------------------------------------------------------------*/
-
- CFX 5.103 - 19/Nov/91
-
- - On registered versions of CFX Pro, some MUTANT Imploder filetypes were so
- mutated that CFX had NO real chance of uncrunching them for the file-virus
- check. This has still not changed, and any attempt to use the "-U" option
- on mutated files may result in a severe system crash. CFX has, however,
- been changed so that it now attempts to determine the extent of mutation,
- and if found to be severe, CFX will not uncrunch the file. This does not
- preclude the possibility that the crunched file is knackered, and CFX may
- still crash the machine when attempting such examinations, even though
- CFX now has tighter control on the situation.
- NOTE: This was apparent *only* in registered versions.
-
- BOB
-
- /*------------------------------------------------------------------------*/
-
- CFX 5.104 - 20/Nov/91
-
- - A slight slip-up during programming allowed certain name-changes to CFX
- (which included node extensions like "CFX.blah") but as viruses are
- becoming more dangerous and devious by the day, I have now ensured that
- NO name-changes are possible on CFX.
-
- BOB
-
- /*------------------------------------------------------------------------*/
-
- CFX 5.105 - 21/Nov/91
-
- - Practically an inter-author test. First phase of wildcard support (NOT
- AmigaDOS wildcard support, UNIX wilds!) including splats (*), hooks (?)
- and character classing ([]). This implementation of wildcarding DOESN'T
- work with the -d option. CFX will happily accept this mixture, but
- recursive directories will not be parsed. The carding will however work
- as desired. There are slight problems in this version, however. Some
- filenames are skipped for some reason. Classing is weird, and won't
- work exactly properly yet. An example of classing:
- 1> CFX c:[ABC]* will examine all files from the c: directory begining
- with A, B or C. Some files may be skipped however (why!?) Fixes are
- forthcoming.
-
- BOB
-
- /*------------------------------------------------------------------------*/
-
- CFX 5.106 - 25/Nov/91
-
- - Further testing has revealed that when attempting to use the "-u" option
- a machine with completely stuffed memory (ie. my machine while compiling
- CFX) or severely fragmented memory, the machine may crash when/after
- uncrunching the file. We have attempted to make the uncrunching process
- as clean as possible, but there's not much more that we can do about this
- problem. We will be further looking into it however!
-
- - This version sees the addition of further graphics types for the Amiga
- and PC. Additional supported filetypes are: JPEG image compression for
- Amiga and PC; binary Postscript images; BMP 24-bit images; HAME images
- (which appear to be giving CFX the wrong messages about image sizes and
- bitplane depth, but we don't really know as we can't find the specs for
- HAME); and finally, Impulse's RGB8 8-bit images.
-
- - The AmigaDOS protect bits for the CLI can make CFX's life a misery. If
- a file has only it's 'H' bit set, or 'A' bit set, then CFX will report
- a "READ ERROR" and refuse to examine the file. This is not CFX's fault
- per se, it's the AmigaDOS function "Read()" which won't allow us the
- privilege. Never mind, just check the protection bits for the file if CFX
- reports a "READ ERROR" without a DOS "READ/WRITE ERROR" requester, and
- change the bits via "protect" if you'd like to examine this file.
-
- BOB
-
- /*------------------------------------------------------------------------*/
-
- CFX 5.107 - 26/Nov/91
-
- - Well, it's a post-release Beta test version!! This version can now lay
- claim to being able to uncrunch mutant Imploders as well as Imploder pure
- crunched files. I have optimised this version with the SAS global
- optimiser, and it should be interesting to see if anyone is able to lock
- up their system with this version. Have a good test you guys!!
-
- BOB
-
- /*------------------------------------------------------------------------*/
-
- CFX 5.108 - 02/Dec/91
-
- - Added 'APE' archiver. Don't ask me why, either. On an accelerated machine
- this stupid piece of work takes 4-10 times longer than LZ to compress,
- and it doesn't even WIN...
-
- - The wildcarding capabilities of CFX have now been minimised, leaving
- only the '*' wildcard. No other wilds work in CFX (unless you are using a
- Dillon-esque shell (ie. CSH)) AmigaDOS wilds don't even work because of the
- particular C front-end that we've linked CFX with. The splat works as a
- standard splat wild (ie. CFX c:A* will examine all files in the C:
- directory starting with the letter 'A'.)
-
- BOB
-
- /*------------------------------------------------------------------------*/
-
- CFX 5.109 - 22/Dec/91
-
- - Fixed redirection so that no unprintable characters are printed during
- a redirected run.
-
- BOB
-
- /*------------------------------------------------------------------------*/
-
- CFX 5.110 - 03/Jan/92
-
- - Well, after adding ALL of the new options for PowerPacker 4, I found a
- couple of nasties crawling around...
-
- Fixed:
- wildcard (splat) directory path bug
- fixed PowerPacker 3 overlay bug
-
- Fixed & Added:
- significant changes made to the mutation code. Now, CFX may take
- slightly longer to figure out if a file's mutated or not. The method
- employed is a double-width method, which allows fallback if CFX
- thinks it may have found a weird mutant. What this means to you, is
- that CFX should now find nearly ALL mutants, known or unknown
-
- Added:
- '-x' option. This will enable the registered user to access further
- information about certain files. At this time, only DMS archives
- have been added. CFX 5.110, when run over a DMS archive with the
- '-x' option, will display relevant info about the DMS archive.
- (This is currently limited to ONE line of info, which I think is
- nice and discreet, clean, and wholesome.) This line of info tells
- the user the starting track number, ending track number,
- compression mode used, and the uncompressed filesize (for good
- measure!) Further filetypes will be added fairly quickly, as I'm
- interested in perfecting this function. BTW, the coding for the new
- functions are all in assembly, so things shouldn't be slowed too
- much...
-
- If anyone has any further ideas (ie. about '-x' support for other
- filetypes) then let me know. I'm looking at next doing PowerPacker data
- files, and most of the major Amiga archivers...
-
- BOB
-
- /*------------------------------------------------------------------------*/
-
- CFX 5.111 - 06/Jan/92
-
- - Shit! I stole some code from several PC Lharc archivers, and f*ck me
- dead if the very same code wasn't completely rooted! What a joke the PC
- is, I think that they're still pissing themselves over that one, but
- unfortunately, some people haven't realised that it is a joke, yet.
- Anyhow, after, ooh, let's see, 14 hours straight, I have finally finished
- the new option, which is just the turbo 'VIEWER' for Lharc/LHA archives.
- This is again under the '-x' option, and works practically like the 'V' or
- 'L' command in Lharc, except that CFX cuts out all the crap, and gives you
- one line of info. If that isn't enough for you, stiff shit. I had to
- write my own PC integer to Amiga integer converter (in assembly) as the PC
- stores its number like this:
-
- our magic number (on the Amiga) is 12345678
-
- on the PC it's stored like this: 78563412. God, how brilliant! What a
- pack of complete legends they really are. What a logical way to store
- 32-bit numbers! Anyhow, just as I was getting REALLY pissed with the
- coding for the Lharc viewer, I put on my Pat Benatar "Wide awake in
- Dreamland" CD, and voila!, she cured the problem. (I put a calculation in
- the wrong place, one line off *:-( Anyhow, give it a test for me, and see
- if you can break it. I haven't adhered to any official LZH structures, as
- the only source that I've got for LHA/LZH is on my PC, and they like to
- store long-words on 3-byte boundaries. Ami don't like that. So I wrote my
- own silly routine, which shouldn't work, but does (up until now, that is!)
- Have fun...
-
- BOB
-
- /*------------------------------------------------------------------------*/
-
- CFX 5.112 - 07/Jan/92
-
- - Well, Mr. O'Callaghan found three bugs:
- + the need for an extra newline after a '-x' dsm;
- + the need for losing the precision of '-x' Lharc dsm percentiles;
- + the fact that EVERY PowerPacker3 file was known as PP3 MASTER.
-
- All of these have been fixed (I hope) so please get back to me if these
- appear dicky, or if anything gets to you...
-
- BOB
-
- /*------------------------------------------------------------------------*/
-
- CFX 5.113 - 07/Jan/92
-
- - Under Marcus's supervision, CFX's '-x' Lharc dsm failed to count the
- number of files, and the correct explosion size of a certain LHA archive.
- This is because we weren't checking for EOF properly. Now, that has
- changed, and the Lharc dsm option should now prove ok (!?)
- Please finalise your wants/dislikes for CFX as I'm just about ready to drop
- this stinkweed of a project (as with all other projects...)
-
- BOB
-
- /*------------------------------------------------------------------------*/
-
- CFX 5.114 - 12/Jan/92
-
- - Another Brett O'Callaghan-ism! He'd like CFX to be able to report just
- WHAT files passed or failed an integrity check, while skipping other
- non-relevant information. Well, this function is now half-completed, as
- you can now specify either:
-
- 1) a straight integrity check of all files, with reporting of all file's
- statuses, or;
-
- 2) an integrity check of all files, but only receiving reports about which
- files FAILED the check, or;
-
- 3) an integrity check of all files, but only receiving reports about which
- files PASSED the check.
-
- To gain a complete integrity check of all files (and receive full
- reporting), use the '-i' option. For a full check, but only receiving
- reports for FAILED or PASSED files, use '-if' or '-ip' respectively. This
- option is subject to all other dynamic switches/options, as in using the
- '-s' option with the '-i' option, you won't receive any information (unless
- you use the BOC '-sl' option) about individual files, but only a summary
- report. Then check the number of entries in the 'Compromised files' area
- for how many bad files there are.
-
- Why I say 'partly completed' is because AmigaDos still puts up the "Volume
- xxx-xxx has a read/write error..." requester, and therefore interrupts your
- CFXing. I do know how to stop AmigaDos from doing this, but it doesn't
- appear to be working, but I'll get it right soon. What this means, is that
- CFX will force AmigaDos NOT to display the requesters, while remaining in
- control of error situations. After CFX has finished it's business, AmigaDos
- will again be allowed to put its requesters back up. Have fun...
-
- BOB
-
- /*------------------------------------------------------------------------*/
-
- CFX 5.115 - 14/Jan/92
-
- - Thanks to the inimitable Mr. Fish (and my co-author) there have been some
- new additions to the CFX brainfile. These are:
-
- Sonic Arranger songs
- Sonic Arranger synthesiser instruments
- Sonic Arranger sampled instruments
- AMOS part-cruncher (reloc) (I couldn't think of anything else to call it!)
- GadToolsBox data files
-
- Marcus and I found a mutated PowerPacker 4 file last night (on a Fish disk)
- by PowerPacker's author, Nico Francois. So, already people are mutating new
- cruncher-headers (even the man who doesn't like doing this!)
-
- I must ask for all of the contributing CFX users to send me any new
- filetypes, or crunchers that you find. The brainfile listing in CFX is now
- fairly up-to-date, so it'll be quite easy to check if I have not added a
- certain filetype to CFX. The more cruncher/filetypes we add, the better
- for the users! See ya...
-
- BOB
-
- PS. Here's the latest registered user list (known as the legend-list):
-
- Bob Rye
- Marcus Mroczkowski
- Stephen Walsh
- Darius Ignasiak
- Mike West
- Brett O'Callaghan
- Richard McElvenny
- Justin Downey
- Brendan Pratt
- Amiga Quarantine
- Andrew Garth
- David Margerison
- Wolfgang Schoenfelder
- Commodore Business Machines Pty. Ltd. (Aust.)
- Ron Hocking
- Andrew Rogl
- Ralph the wonder llama
-
- Thanks, guys...
-
- /*------------------------------------------------------------------------*/
-