jdg@elmgate.UUCP (Jeff Gortatowsky) (06/09/86)
I got around to playing with 1.2 beta this weekend. All the playing I did
was at a "user" level. Meaning I haven't even begun to program with the 1.2
fixes and enhancements. The report from my experiance isn't looking good.
If C-A is planning on releasing 1.2 *NEARLY* bug free, don't expect to see
the release anytime soon. My list of bugs is 2 pages long and growing.
As I said, I haven't even begun to program in it yet.
The workbench in 1.1 WAS fairly reliable. At least in my experiance.
Ah-hem, it ain't so in 1.2b2. Now I don't know if CBM-Amiga wants
developers airing dirty laundry out in public so I'll keep this
non-specific for the moment.
The workbench's error recovery and window refreshing is damaged.
The notepad has numerous problems making it a risky item to use. I
certainly wouldn't trust my bugs notes to it!
What happen? My guess is, like in my own programs, too many fixes were
installed all at once. In my own code, after 3 or patches to the same
routine the code becomes so different from what I originally had intended
that I can't follow it or fix it any further. Time to scrap that routine
and rewrite it. C-A better do the same (I suspect).
It's not that I don't expect bugs in a beta release. However the bugs I'm
finding are *MAJOR* items that worked before (ie. workbench windows). The
idea behind a new minor release (1.x to 1.x+1) is to fix documented stuff
that didn't work in the previous release and hone everything else.
Unfortunatly 1.2 is more like the 1.0 major release (at least so far as
I've explored it) in that it is frustrating to work with due to all the
simple (simple to the 'user', not the programmer) things that seemingly
don't work.
OK guys.... I KNOW kickstart is a very complex hunk (hunks) of code. I
know it's easy to sit here on my SUN and fire away at you. The question is
what to do about it? With as many bugs cropping up in things that just the
user is going to use you can't possibly (can you) be considering releasing
1.2 anytime soon. You kill the machine (ie frustrate the user). I suggest
you might try releasing 1.11. In it keep the paths, expansion stuff, and
disk caching. Beyond that start doing what I do when routines start
getting outta hand. Break 'em apart and rewrite incorporating the patches
from the start. Beyond that you *might* (just a suggestion) want to hire a
couple of those guys back at least as part time consultants. It's gonna
take some effort (I think) to clean 1.2 up before a real release.
Sorry to be so long winded, but I'm concerned. Things are supposed to be
getting better not worse. I haven't even started compiling old/new code.
Maybe that'll go better.
Oh yes, the bug reports will go out Wednesday. I've gotta run off copies
(many) and type 'em up. Later.....
--
Jeff Gortatowsky {allegra,seismo}!rochester!kodak!elmgate!jdg
Eastman Kodak Company
<Kodak won't be responsible for the above comments, only those below>grr@cbmvax.cbm.UUCP (George Robbins) (06/10/86)
In article <468@elmgate.UUCP> jdg@elmgate.UUCP (Jeff Gortatowsky) writes: > >The workbench in 1.1 WAS fairly reliable. At least in my experiance. >Ah-hem, it ain't so in 1.2b2. Now I don't know if CBM-Amiga wants >developers airing dirty laundry out in public so I'll keep this >non-specific for the moment. Please go ahead and post any bugs you find - It's better to find them once and report them, than for n people to be complaining about bugs and nobody knows if they are new ones or old favorites. >What happen? My guess is, like in my own programs, too many fixes were >installed all at once. In my own code, after 3 or patches to the same >routine the code becomes so different from what I originally had intended >that I can't follow it or fix it any further. Time to scrap that routine >and rewrite it. C-A better do the same (I suspect). Pretty close - too many goodies and fixes were thrown into the beta release at the last moment. The previous alpha release was probably better as far as day to day reliability. >Oh yes, the bug reports will go out Wednesday. I've gotta run off copies >(many) and type 'em up. Later..... >-- >Jeff Gortatowsky {allegra,seismo}!rochester!kodak!elmgate!jdg Could I suggest that bug reports posted to the net try to approach the formalism used for posting bugs to net.bugs.4bsd? This is not an attempt at being officious, but rather a very pragmatic suggestion. The more clearly you can describe a bug and how to invoke it, the more likely that someone at Amiga or Commodore can repeat the problem and fix it. Here is an outline based on the 4.2BSD bugformat... Subject: Short summary of the problem Index: <program|file|system call> 1.2b2 Description: Detailed description of the problem, suggestion, or complaint. Repeat-By: Procedure to repeat the problem. Fix: Description of how to fix and/or work around the problem. -- P.S. Flames to /dev/null on this - if you've ever worked in technical support or software maintenance you can understand the problem... -- George Robbins - now working with, uucp: {ihnp4|seismo|caip}!cbmvax!grr but no way officially representing arpa: cbmvax!grr@seismo.css.GOV Commodore, Engineering Department fone: 215-431-9255 (only by moonlite)