[haiku-development] bug tracker
- From: Michael Mounteney <gate03@xxxxxxxxxxxxxxx>
- To: haiku-development@xxxxxxxxxxxxx
- Date: Mon, 21 Apr 2014 21:16:08 +1000
For those who aren't on the general haiku mailing list, here's a
summary of a thread started there:
>>> From me:
Hello, here is an offer of help.
Although I'm a programmer, I already program for eight hours a day as
my job so frankly the last thing I want to do in my scant leisure time
is more programming. However, it occurs to me, looking at the bug
tracker, that of the almost 3000 open bugs, half have not been updated
for 21 months and I could help to pull that number down by sifting
through them. In most cases, it needs a programmer's knowledge to do
that, even though it's not actually programming.
Some, like https://dev.haiku-os.org/ticket/8756 would be very quick to
fix if only the information were conveyed to the right person, while
others like https://dev.haiku-os.org/ticket/8777 are probably invalid
now (no disrespect to the reporter at the time).
Another task is proofreading documentation but I think there are
already people doing that.
So would my intervention in going through the bug list be welcome ? If
say 500 to 1000 bugs could be closed, that would make the list less
daunting.
>>> To which Humdinger replied:
I'd say the number of 500 to 1000 easily closable non-coding related
tickets is a bit optimistic. But your work would be very much
appreciated in any case!
>>> My reply to Humdinger:
I'd like to know I had a mandate of some sort, even just to the extent
that the main characters, Axel, Ingo, Humdinger etc. approved what I
was doing, rather than the usual open source prima donna nonsense of me
turning up and deciding what I'll contribute in the absence of any real
need. Otherwise it's like arriving at someone elses party and saying
here's how things are going to be run from now on. A bit rude to say
the least, as I'll be contacting people and asking for more
information, deciding that bugs will be closed etc. That's a
significant responsibility.
Some of the tickets may very well be code-related. I'd say any
non-specific it-just-crashes older than say 12 months ought to be
closed unless it can be reproduced again now. In fact, *any* that
doesn't provide a means to reproduce; sorry guys but what can we do
with that ?
>>> And Humdinger again:
Who knows, maybe you'll find some you can't resist the itch to come up
with a patch for. More eyes are always good, more hands are better. :)
>>> My reply to Humdinger:
Maybe so but I won't be able actually to use Haiku myself until
https://dev.haiku-os.org/ticket/9932 is solved except maybe on a VM
just for testing.
>>> Humdinger:
The closing of tickets is reserved to people with commit access (I
think). Insofar, you're free (and encouraged) to go through the
tickets, add comments with requests for more information or, when you
decided that a ticket doesn't provide clues to its solution, request
its closure.
All committers are subscribed to Trac and should see your comments and
have time to react to them.
I suggest to add a keyword to a closable ticket, so we're able to
quickly list them all. Maybe simply "purge"?
I'd volunteer to go through your list of tickets every other week and
close where there was no veto from anybody.
> Some of the tickets may very well be code-related. I'd say any
> non-specific it-just-crashes older than say 12 months ought to be
> closed unless it can be reproduced again now. In fact, *any* that
> doesn't provide a means to reproduce; sorry guys but what can we do
> with that ?
Our experienced coders will have to answer that. While I agree that
tickets that only say "it-just-crashed" should be save to be purged.
OTOH, when backtraces are attached, I'm not sure those can't help
track down the bug...
I just realize that this thread should've been started on the
haiku-dev list. If you don't get replies from more competent people
here, you may want to post a summary there.
So to summarise: is my offer of help, in cutting down the list of
outstanding bugs, welcome ?
Michael.
Other related posts:
