home *** CD-ROM | disk | FTP | other *** search
- Bugs:
- ~~~~~
-
- Things to do:
- ~~~~~~~~~~~~~
- - For multiple files, tell the net size gain/loss.
- - XPK support, instead of PowerPacker support.
- - Change the (ax,xxxx) addressing syntax to xxxx(ax), so SNMA can be used.
- - When running out of space, source file is deleted.
-
- Things to do for the documentation:
- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-
- Unimportant things to do:
- ~~~~~~~~~~~~~~~~~~~~~~~~~
- - Checks for ANSI & InterNet modes waste CPU time.
- - Autoalloc of bigger buffers when needed, copying the old data.
- - When converting SF7, if there is no letter around the supposed SF7 char,
- don't translate it.
- - Option: Turn temp dir off or on from the commandline.
- - Encrypt tables and texts so no-one fiddles with them. Flush cache
- afterwards.
- - In NoBuffer mode, the stats flash annoyingly between reading & writing,
- since the buffers are almost the same size.
- - Double space. (Undouble)
- - At least writing the write buffer could be made into a subroutine,
- perhaps read too.
- - Progress information on reading & writing.
- - Clean OpenFiles.
- - In the end of the conversion, tell bps stats.
- - Use the regs more effectively.
- - Don't allocate CrunchInfo multiple times.
- - TabConversion could probably be optimized by using tables.
- - AmigaGuide stripper.
- - Don't scan back in the file when checking for PP.
- - Intuition interface.
- - If the destination can't be written, and the mode is buffered, retry
- option.
-
- Port Operation Information:
- ~~~~~~~~~~~~~~~~~~~~~~~~~~~
- OpenFiles():
- Determines which buffering mode to use, and jumps to OpenFiles1() if we
- have a full source buffer.
-
- OpenFiles3():
- If we have no conversion buffer, this is the last routine before
- ReadBlock() which is already in the conversion loop.
-
- ReadFinished():
- is jumped to when the source has been opened and read. Destination will
- be opened there, after which MoreData() is jumped into. This starts the
- real conversion loop.
-
- Coded Port on:
- ~~~~~~~~~~~~~~
- 18-May-94: 1h46m
- - Researched the MatchFirst() routine.
-
- 22-May-94: 4h6m
- - Changed to PhxAss.
- - Kick V36 required now, started matching code.
-
- 23-May-94: 0h15m
-
- 31-May-94: 3h47m
- - Got a first version of MatchFirst() code to work, only handles one file.
- - File opening made more modular.
- - CodeError() routine fixed.
-
- 02-Jun-94: 3h23m
- - No longer loads the table in a3 at every character cycle.
- - First version of multi-file matching works.
- - Tests for a lame CPU and exits if found.
- - Doesn't put the old processor checking code in the 68000 version.
- - No longer saves TranslateAddr.
- - Checks for break on Match*() routines.
-
- 03-Jun-94: 5h9m Total: 1105m (18.4h)
- - No longer takes a file destination for wildcards.
- - Wildcards routine finished.
- - Use of bold removed.
- - PowerPacker decrunch added.
- - Also supports encrypted files.
- - Shows the filename being converted.
- - Always shows the copyright info.
- - Replaced DestIsDir() with the ROM AddPart().
- - Started changing the documentation to AmigaGuide format.
-
- 04-Jun-94: 1h35m Total: 1200m (20.0h)
- - Finished converting to AmigaGuide documentation.
- - LFs taken out of error messages, now ShowErr() automatically appends a LF
- and a clear EOL to all error messages.
-
- 05-Jun-94: 7h32m Total: 1652m (27.5h)
- - Worked on the strange bug which causes MatchNext() to take same files
- over and over. It appears to be a bug in MatchNext() itself.
- - Bugfix: The routine which uses Fault() to print an error didn't have a
- linefeed.
- - Added a warning about the wildcard bug in the docs.
- - Told how to make an "ALL" option for the wildcards in the doc.
- - 68000 version.
- - Sync'ed text output, now updated once a sec on all machines.
- - Read speed sync'ed, reads around once a sec.
- - This version is 37% faster than V1.46.
- - Normal text output with NoBuffer mode.
-
- 06-Jun-94: 2h56m Total: 109680s (30.5h)
- - Started optimizing the translation code and killing useless logic.
- - Fixed a bug in the conversion progress routine.
- - Trashed SpacePointer, now only uses a2 instead. Code 6% faster.
- - All modes use tables now.
- - All modes use tables, all modes kill all characters under $20 except
- LFs & tabs.
- - SF7 now also converts å.
- - Fixed a bug, FromPC upper 128 characters caused weird results.
- - Started coding a new WhiteProcess routine.
-
- 07-Jun-94: 3h46m Total: 123240s (34.2h)
- - Recoded the whole Space-Tab-EOL routine, and it seems to work
- beautifully.
- - Moved the White routines in SUBs section.
- - Inserts one LF in the end of the file.
-
- 08-Jun-94: 1h52m Total: 130018s (36.1h)
- - Designed test files for White processing routines.
- - With EndSpace removal on, an incorrect EOL was inserted.
- - SpaceBuf pointer (A2) was not initialized in the beginning.
- - MakeEOLs was triggered only when there were more spaces than specified
- with the -e option.
- - Now Port has an auto-updated header.
- - Quietly skips directories.
- - Added Combined (PC+SF7) mode.
- - All modes use tables, all modes kill all characters under $20 except
- LFs & tabs & ESCs.
- - New address added to PortOpts2
- - NoBuffer mode no longer added suffix with same file, fixed.
-
- 09-Jun-94: 11h31m Total: 171537s (47.6h)
- - Coded the matching routine which goes through the directory before
- converting anything.
- - Tells when a partial file is deleted.
- - PowerPacking of output file added.
- - Allocates source and dest buffers based on free memory.
-
- 10-Jun-94: 7h31m Total: 198632s (55.2h)
- - Full Amiga->PC character support.
- - Recognizes as PC file from the use of $80-$90 characters.
- - AllocBuffers now allocates the write buffer based on the source file
- size, taking into account if it was crunched.
- - When doing reads and writes, tells so. Immediately updates the stats
- afterwards.
- - Coded the preferences decoder.
-
- 11-Jun-94: 1h23m Total: 203627s (56.6h)
- - Coded option supressing.
-
- 12-Jun-94: 5h2m Total: 221753s (61.6h)
- - Fixed a bug in PrintStats.
- - Made Quiet mode more effective, now no progress information is printed.
- - PrintStats showed the written bytes instead of converted, fixed.
- - Coded the Stat update rate as an option.
- - In SpacesToTabs, tabsize can be 1-16. In TabsToSpaces, tabsize can be 1-16.
- - Planned ANSI & Internet strippers.
-
- 13-Jun-94: 5h48m Total: 242651s (67.4h)
- - Planned ANSI & Internet strippers.
- - Made WhiteProcess modular.
- - Coded Internet header stripper.
-
- 14-Jun-94: 3h59m Total: 257048s (71.4h)
- - Coded ANSI stripper, in 35 minutes.
- - Coded two options for the InterNet stripper, one inserts a "HEADER
- REMOVED" message and the other inserts nothing.
- - Fixed a bug in StrToNum, now works for values 0-99999.
- - Coded user selectable temp dir support for crunching.
- - Planned LhA packing of destination.
-
- 15-Jun-94: 2h14m Total: 265105s (73.6h)
- - Coded LhA packing of destination.
-
- 16-Jun-94: 5h36m Total: 285267s (79.2h)
- - Fixed a bug in MakeLhAName().
- - Alphabetized preferences options.
- - In the prefs file, added a short description of each option.
- - Cleaned EndProgram.
- - Cleaned OpenFiles a bit.
- - Made tests on conversion stats rate.
- - Changed a lot of data references from absolute to register relative.
- - Fixed some bugs in OpenFiles.
- - Added some support for error return codes.
- - Coded the 68000 code.
- - Edited the documentation.
-
- 17-Jun-94: 1h26m Total: 290427s (80.7h)
- - Edited the documentation.
- - Fixed a bug in the InterNet header stripper. Now takes all lines that
- start with "From " and have numbers and colons & lines that start with
- "Path: " as the start of the header. This is a bit dangerous.
-
- 18-Jun-94: 2h56m Total: 301039s (83.6h)
- - Worked on the documentation.
- - Commented the code.
- - Optimized code, and made memory accesses DT relative. As a result, the
- code shrank 2448 bytes. While doing that, omitted some useless
- instructions and variables.
- - Now continues reading the preferences file even after an unrecognized
- option.
-
- 19-Jun-94: 1h51m Total: 307701s (85.5h)
- - Fixed a bug in the PowerPack dest ShowStats.
- - 6147 3431
- - Changed "LHA a" to "LHA r".
- - Fixed the bug resulting in returning bad return codes.
-
- 20-Jun-94: 2h11m Total: 315564s (87.7h)
- - MakeLhAName() handles suffix lengths from 1-3. (Previously only 3).
- - Added the InterNet explanations in the quick help.
- - Dropped Automatic mode, unreliable.
- - Dropped Quiet mode, as it was found useless.
-
- 26-Jun-94: 1h20m Total: 320417 (89.0h)
- - Worked on the documentation.
- - Released V2.00 R417.
-
- Work after initial V2.00 release:
-
- 17-Jul-94: 1h6m Total: 324436 (90.1h)
- - 'å' & 'Å' didn't get translated in ToSF7, fixed.
- - Added an automatic test version report to Debug flag compiles.
- - Made a lot of automatic release logic.
- - Released V2.00 R422.
-
- 19-Jul-94: 0h57m Total: 327914 (91.1h)
- - Looked into some bugs.
- - Bugfix: When converting on the original file, stack was mangled while
- writing.
- - Bugfix: Null files caused Port to abort with error "not enough memory",
- now they are quietly skipped.
-
- 01-Aug-94: 0h18m Total: 329050 (91.4h)
- - Released V2.00 R426.
-