On 2005-01-29, Zooko O'Whielacronx
<zooko-F7hWHBaSm8MAvxtiuMwx3w@xxxxxxxxxxxxxxxx> wrote:
>
> #101: Fail: bug in darcs! Problem resolving conflicts in resolve!
> http://bugs.darcs.net//Ticket/Display.html?id=101
> Unfortunately I've lost the buggy repo tarball that I generated for
> this when I first hit the bug last November. Unless David has a copy
> of the tarball that I submitted we'll have to treat this one as
> irreproducible.
>
> #119: Error applying patch to working dir.
> http://bugs.darcs.net//Ticket/Display.html?id=119
> I've lost the tarball for this one, from last December. We had a disk
> space crunch on my server. I guess from now on I'll submit the
> tarballs as attachments to the bug tracker instead of storing them on
> my web site. Although they are perhaps too big to be submitted that
> way.
>
> #132: bug report: Fail: r_u commute bug, contact droundy@xxxx!
> http://bugs.darcs.net//Ticket/Display.html?id=132
> I submitted this last December, and although I do have a tarball I made
> later that same day, it is probably not a complete tarball of the
> problem. What a shame!
This are all problematic, because without a way to reproduce the bug,
it's difficult to know if they have been fixed or not unless they appear
again. That doesn't make them less important, just difficult to work on.
> #161: Error applying patch to working dir.
> http://bugs.darcs.net//Ticket/Display.html?id=161
> I've submitted a minimal buggy repository showing this bug.
>
> #176: darcs bug report: indelible conflict
> http://bugs.darcs.net//Ticket/Display.html?id=176
> There are many repos available that show this bug. This is my least
> favorite bug right now (excluding the "matter/antimatter O(2^N)" design
> issue), because it bites me approximately once a day.
According to the ticket comments, these should be solved by the new
'Conflictors' work happening now in unstable. Because that work is a
major upgrade, it may well be a best of use time to not try it and put
it into a near-term stable release.
> #166: unrevert results in "WARNING: You may now have a corrupt (but
> fixable) repository."
> http://bugs.darcs.net//Ticket/Display.html?id=166
> The tarball containing this repo is still available.
Unfortunately, only David has access to the URL, so other developer's
can't easily work on it, or even confirm if it's already been fixed.
> #190: some kind of failure and possible corruption
> http://bugs.darcs.net//Ticket/Display.html?id=190
> There is a tarball available containing the state of my repo *after*
> this bug bit. Hopefully that is sufficient to find the bug.
I'll poke around a bit with this now.
#####
Zooko,
I think your bug reports are very helpful. These generally seem like
important and hard to reproduce bugs.
I imagine the hardest part with some of these will not be the coding,
but simplifying the case to figure exactly what's triggering the bug in
the first place.
I think it's worthwhile to go ahead and release 1.0.2, because it's
better than 1.0.1. :)
Mark
--
http://mark.stosberg.com/
|
