danbabcock@eklektik.UUCP (/dev/ph1) (08/15/90)
The following are two bug/feature reports that I emailed to Commodore. I'm posting them because I'm sure some other people are interested, and I'm not sure if Commodore received my email (no response as of this writing). I noticed a problem with task exceptions while writing a program that uses them (a serial.device compatible driver for a super-cheap multiserial, multiparallel board that works with any Amiga - I use a task exception to handle non-queued breaks). I hear you groaning already. Yes, I *have* read the warnings about using task exceptions, but this code only uses Exec - and Exec supports exceptions internally, right? Well, I found a case where it (Exec) doesn't. Exceptions fail to work when the task is (at the time of the exception) in Wait() and the exception routine itself uses Wait(). The problem is that Wait() writes over tc_SigWait. The solution is very simple: the exception code should save tc_SigWait before calling Wait(), and restore it afterwards. ---------------------------------------------------------------------------- I found a bug in the skeleton device driver. The bug appears in both the printed 1.3 RKM and the latest (to the outside world, at least) electronicly distributed version (as well as the earlier versions). The problem is in the flush routine. In the following code (the tail end of the flush routine)... tst.b d0 beq.s 1$ bsr InternalStart 1$: bra TermIO ...the BEQ should be a BNE. I could explain the "why", but you probably won't understand (or trust me :-)) until you look at the code for yourself anyway. I'm surprised that no one reported this one before, although I assume many people noticed it and fixed it in their own drivers. -- Dan Babcock People/Link: DANBABCOCK Internet: danbabcock@eklektik.pgh.pa.us <-- Note: Invalid on August 17+ Voice: (412)-373-1753 <-- up to and including August 17 New Voice: (814)-862-2931 <-- August 19+ (I'm going off to college...)
valentin@cbmvax.commodore.com (Valentin Pepelea) (08/18/90)
In article <5177@eklektik.UUCP> danbabcock@eklektik.UUCP (/dev/ph1) writes: > > The following are two bug/feature reports that I emailed to Commodore. I'm > posting them because I'm sure some other people are interested, and I'm not > sure if Commodore received my email (no response as of this writing). If you mailed the bug reports to bugs@cbmvax, hances are the mail was received. But you did not get a reply, because no reply was ever sent to you. It would be ideal for bugs@cbmvax to reply to you with the bug number that was assigned to your report, but alas that is not the case. I'm resubmiting your reports to bugs, just in case. Valentin -- The Goddess of democracy? "The tyrants Name: Valentin Pepelea may distroy a statue, but they cannot Phone: (215) 431-9327 kill a god." UseNet: cbmvax!valentin@uunet.uu.net - Ancient Chinese Proverb Claimer: I not Commodore spokesman be
andy@cbmvax.commodore.com (Andy Finkel) (08/21/90)
In article <13867@cbmvax.commodore.com> valentin@cbmvax (Valentin Pepelea) writes: >In article <5177@eklektik.UUCP> danbabcock@eklektik.UUCP (/dev/ph1) writes: >> >> The following are two bug/feature reports that I emailed to Commodore. I'm >> posting them because I'm sure some other people are interested, and I'm not >> sure if Commodore received my email (no response as of this writing). > >If you mailed the bug reports to bugs@cbmvax, hances are the mail was received. >But you did not get a reply, because no reply was ever sent to you. It would be >ideal for bugs@cbmvax to reply to you with the bug number that was assigned to >your report, but alas that is not the case. Sorry, he has no idea what he's talking about. We do a reply on emailed bug messages as an ACK before its entered in the database, but rely on the return address being accurate. They seem to bounce a lot :-( andy -- andy finkel {uunet|rutgers|amiga}!cbmvax!andy Commodore-Amiga, Inc. "Of course it's the murder weapon. Who would frame someone with a fake?" Any expressed opinions are mine; but feel free to share. I disclaim all responsibilities, all shapes, all sizes, all colors.
xanthian@zorch.SF-Bay.ORG (Kent Paul Dolan) (08/21/90)
andy@cbmvax (Andy Finkel) writes: >valentin@cbmvax (Valentin Pepelea) writes: >>danbabcock@eklektik.UUCP (/dev/ph1) writes: >>> >>> The following are two bug/feature reports that I emailed >>> to Commodore. I'm posting them because I'm sure some other >>> people are interested, and I'm not sure if Commodore >>> received my email (no response as of this writing). >> >> If you mailed the bug reports to bugs@cbmvax, chances are >> the mail was received. But you did not get a reply, >> because no reply was ever sent to you. It would be >> ideal for bugs@cbmvax to reply to you with the bug number >> that was assigned to your report, but alas that is not the >> case. > > Sorry, he has no idea what he's talking about. Now children, don't quarrel in-house; go out on the net to play. Oops! Wrong newsgroup! Thought I was in misc.kids. ;-) > We do a reply on emailed bug messages as an ACK before > it's entered in the database, but rely on the return > address being accurate. They seem to bounce a lot :-( Now there at least is an easy problem to solve! Take a page from the vote takers' manual; if your ACKs bounce, bulk post them! I think _everybody_ would like to see a monthly posting of new bug-number, bug-reported-by, bug-assigned-to, bug-priority, bug-one-line-descriptions from CBM show up in comp.sys.amiga.tech, and the load as a percent of the newsgroup volume would be infinitessimal. Add to that a list of known outstanding bugs, and known fixed bugs with fixes not yet released, and it would be a lot easier to decide whether to report the latest quirky behavior seen in AmigaDOS 1.1. ;-) If someone at CATS has the time, that would be a good PR and public service kind of a thing to do on USENet, and would save developers and users mountains of time figuring out that the problem is in the OS, and not in our stars. Kent, the man from xanth. <xanthian@Zorch.SF-Bay.ORG> <xanthian@well.sf.ca.us>