home *** CD-ROM | disk | FTP | other *** search
- v0.35 changes the way I was rewinding files... it didn't work
- properly, which caused NCR compression ("packing") to fail.
-
- v0.34 fixed a bug that always caused deletion of the $$$.SUB file
- even if there were no errors.
-
- v0.33 fixes a bug in the date code support that was inadvertently
- added in 0.32... (ever have one of those days?). Only affects
- those with Date support in their operating system. This time I
- was able to test it, so now it _does_ work. (It was miscalc-
- ulating the minutes.)
-
- v0.32 fixes a bug in the date code support that made dates
- entered by patching act strangely.
-
- v0.31 was an unreleased version that fixed the bug regarding the
- length of the command line.
-
-
- Changes made in version 0.3 of ARK.COM:
-
- 1) Further speed increases in the actual compression code (read:
- more hand-optimization... what fun!).
-
- 2) Complete rewrite of all disk I/O code. This is the biggest
- change. This makes the program run fine and dandy on my 52k
- TPA Z3 system. This should mean no more incredibly long
- runs for no particular reason.
-
- 3) Support for time/date!!!! Yep, I finally got around to it.
- This is only partially supported at present, but I do plan
- on adding full support for Z80DOS and DateStamper in the
- next few weeks (after I get my version of Z80DOS working
- with my Dallas Semiconductor SmartWatch as well as some
- insight into how DateStamper works). For now, ARK makes a
- BDOS call (105) to get the present date. If the OS returns
- a date, that is used for dating _all_ changes in the
- archive. This works fine with most systems that support a
- real-time clock. (It also works on standard CP/M systems
- with no RTC -- see #4 for further info...)
-
- 4) To make life easier for those using another program to insert
- a date into ARK.COM, I have moved the four bytes that hold
- the date and time to 103h, and there they shall remain. In
- fact, the first few lines of the program look like this:
-
- org 100h
- jp start
- db year ; 0103h
- db month ; 0104h
- db day ; 0105h
- db hour ; 0106h
- db min ; 0107h
- ...
-
- Note that these are stored as simple binary values, and
- _not_ in the strange format that ARK uses internally. This
- should make coding any external program much easier.
-
- 5) I have also added a new option -K[runch]. This will skip
- the "analyzing" stage of compression, which will further
- enhance the speed (since only very small files and maybe a
- couple of larger ones would be best stored uncompressed).
-
- I have also given consideration to doing PKARC-type "squashing"
- of files. Alas, it does not seem feasible under CP/M. In order
- to do this, I would need to create a table of over 9000 4-byte
- entries... in addition to another table of 2-byte entries. That
- adds up to 54k of data tables, a net increase of 24k. Until I
- can get a Z280 system together, that will just have to wait.
-
- And... just before I was ready to release this, a bug showed up.
- The problem is that you can not do this:
- A>ark test b:*.*
- and have it always work. It seems to get angry when it has to
- pack the file and the file is not on the default drive. I don't
- know why it does this (and I wrote the drivers!). I will look
- into it some more, but this release will help silence those
- people (like myself) that can't live without their memory-
- resident programs and Z3.
-
- As always, I can be reached in the following places:
-
- B.E.MOORE [GEnie]
- Brian Moore [The Machine BBS -- (503) 747-8758]
- (editorial note: The Machine is _not_ MY BBS, it is run by a
- friend of mine, Tony Newman, and is one of the nicest boards
- around.)
-
- or at:
- 1550 High St., #2
- Eugene, OR 97401
- (503) 687-8531 (note: new address and phone #!!!)
-
-
- 8/8/88, 8:50pm
-