Sign in Join

Live chat for Free Pascal, Lazarus, Delphi and every Pascal programmer.

Mailing lists / lazarus / 2011-08

Moderating the bugtracker

10 messages · 6 people · –

Reply by email Subscribe to this list Your subscription

Read-only archive. Use j and k to move between messages.

Bart

Hi,

I have a question on moderating the bugtracker, more specifically on when I should assign (and unassign) issues to myself.

In the Lazarus bugtracker, I have the "status" (don't know how it is called) of developer.
I do however not have commit rights (nor do I aspire to get them) on svn.

In the past I was told I could assign issues to myself, if I thought I could fix them, and unassign them if it turned out I wasn't able to fix it after all.

I have done so for quite a while now, but recently when I assigned an issue to myself and fixed it, I was asked why I assigned it to myself, since I didn't have commit rights.
It was suggested to me that I should unassign the issue as soon as I had fixed it and set a relevant target.

I must admit that a little bit of vanity comes into play here as well: if an issue assigned to me (and fixed by me) gets resoved (without changing the assignment) it'll be reflected in the bugtracker statistics.

But putting that aside, what should be the general guidline for this scenario, where one has assigned an issue to oneself and then fixed it, but one doesn't have commit rights?

I can see several options:

1. Notify someone with commit rights a patch is available (the bugtracker doesn't offer such functionality)

2. Move the issue to "Patches", do not change the assignment (for me , and probably others, this has the drawback that in "Patches" I am "reporter", so I cannot move the issue back, or unassign if needed)

3. Set the target and just wait until someone commits the patch (this is what I did in the past)

4. Set the target and unassign the issue.

5. Never assign issues to you if you do not have commit rights.

The wiki article about moderating the bugtracker
(http://wiki.freepascal.org/Moderating_the_bug_tracker) does not describe this situation, so it might be a good idea to update it?

B.t.w. Is it OK to set the current trunk (currently 0.9.31) as target?

Bart

Felipe Monteiro de Carvalho
to Bart

This is the bug report which brought this issue up:

http://bugs.freepascal.org/view.php?id=19638

And there I already spoke about the subject.

The main issue here for me is that having issues with waiting patches assigned to people without commit rights will cause confusion: One cannot easily see which patches are unassigned, unless you have a list of who are the developers without commit rights. So I propose to use option 4: Set the target and unassign the issue.

B.t.w. Is it OK to set the current trunk (currently 0.9.31) as target?

No, only future releases should be targets, so the nearest possible target is 0.9.99 and this should be the target for any issue with a valid patch

Felipe Monteiro de Carvalho
Juha Manninen
to Bart: I have a question on moderating the bugtracker, more specifically on when I should assign…

2011/8/24 Bart

Show quoted text 5 lines
In the Lazarus bugtracker, I have the "status" (don't know how it is called) of developer. I do however not have commit rights (nor do I aspire to get them) on svn.

How does one end up having developer status but no commit rights?

I personally would give commit rights to some people who constantly provide patches, if only they asked for the rights.
I also realize it is important the person asks for it himself and it is not pushed anyhow.

Now some previously active developers have lost their motivation or energy, or they are busy with other things, maybe temporarily or maybe for longer time.
On the other hand, some reporters are extremely active and have both motivation and energy.
IMO the changing reality should reflect into the commit rights, too. Otherwise the result is a stagnant and bureaucratic development process.

I have committed many patches from reporters during past few months. Without those commits the authors would still be waiting and soon loose their energy for this project. Ignoring a patch is a strong negative feedback, at least it feels like that for the author.

Yet, I don't feel the situation is very bad now but it could be better.

Juha
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20110824/c4738d86/attachment.html>

Bart
to Felipe Monteiro de Carvalho: This is the bug report which brought this issue up:…
Felipe Monteiro de Carvalho wrote
This is the bug report which brought this issue up: http://bugs.freepascal.org/view.php?id=19638 And there I already spoke about the subject.

I know, but since I was instructed otherwise in the past I would like to know the opinion of other devels.

Show quoted text 5 lines
The main issue here for me is that having issues with waiting patches assigned to people without commit rights will cause confusion: One cannot easily see which patches are unassigned, unless you have a list of who are the developers without commit rights. So I propose to use option 4: Set the target and unassign the issue.

I would think moving to "Patches" (with or without unassigning) would be even more indicative?

Show quoted text 5 lines
> B.t.w. Is it OK to set the current trunk (currently 0.9.31) as target? No, only future releases should be targets, so the nearest possible target is 0.9.99 and this should be the target for any issue with a valid patch

0.99.0 or 0.99.1? I can see no 0.9.99 ;-)

Besides, if "we" can agree about the issue, the wiki page can be updated accordingly.

Bart

Bart
to Juha Manninen: 2011/8/24 Bart How does one end up having developer status but no commit rights? I…
Juha Manninen wrote
How does one end up having developer status but no commit rights?

I have no idea. One day I just was "developer".
It was done so that all MaskEdit issues could be assigned to me.

I personally would give commit rights to some people who constantly provide patches, if only they asked for the rights. I also realize it is important the person asks for it himself and it is not pushed anyhow.

I have never considered asking for it.
It would be yet another thing to learn.
Also I am afraid I might seriously f** up the repository by issuing a stupid command
(I am the kind of person that can accidentally do a rm -R *)

Yet, I don't feel the situation is very bad now but it could be better.

An option in the bugtracker to notify a patch is provided for review would be nice.

Bart

Felipe Monteiro de Carvalho
to Bart: On 8/24/11, Felipe Monteiro de Carvalho I know, but since I was instructed otherwise in…
Bart wrote
I would think moving to "Patches" (with or without unassigning) would be even more indicative?

I don't think this solves the confusion. A patch might be in patches for other reasons then you moving it there. The original reporter of a patch might have put it there or a developer might have moved it there for whatever other reason.

0.99.0 or 0.99.1? I can see no 0.9.99 ;-)

0.99.0

Felipe Monteiro de Carvalho
Flávio Etrusco
to Juha Manninen: 2011/8/24 Bart How does one end up having developer status but no commit rights? I…

2011/8/24 Juha Manninen :

Show quoted text 8 lines
2011/8/24 Bart > > In the Lazarus bugtracker, I have the "status" (don't know how it is > called) of developer. > I do however not have commit rights (nor do I aspire to get them) on svn. How does one end up having developer status but no commit rights?

AFAIK 'developer' is just a standard 'role' in Mantis with permissions to edit bugs.
I agree this can generate some confusion, even more when you're commenting on a bug and the reporter think I'm speaking as a Lazarus developer...

-Fl?vio

Alexander Klenin
to Felipe Monteiro de Carvalho: I don't think this solves the confusion. A patch might be in patches for other reasons…
Felipe Monteiro de Carvalho wrote 8 lines
On Thu, Aug 25, 2011 at 3:08 PM, Bart wrote: > I would think moving to "Patches" (with or without unassigning) would > be even more indicative? I don't think this solves the confusion. A patch might be in patches for other reasons then you moving it there. The original reporter of a patch might have put it there or a developer might have moved it there for whatever other reason.

In general, I think "Patches" subproject is confusing and should be eliminated.

Alexander S. Klenin
Vincent Snijders
to Alexander Klenin

2011/8/28 Alexander Klenin :

Show quoted text 13 lines
On Fri, Aug 26, 2011 at 00:31, Felipe Monteiro de Carvalho wrote: > On Thu, Aug 25, 2011 at 3:08 PM, Bart wrote: >> I would think moving to "Patches" (with or without unassigning) would >> be even more indicative? > > I don't think this solves the confusion. A patch might be in patches > for other reasons then you moving it there. The original reporter of a > patch might have put it there or a developer might have moved it there > for whatever other reason. In general, I think "Patches" subproject is confusing and should be eliminated.

The idea is / was to have a special place for patches, so that they can get priority treatment, This was before Mantis supported tags, so maybe this can be handled wth tags better these days. OTOH, I would miss the possibility to have statistics for patches only.

Vincent