Stream: t-compiler

Topic: maintenance/triage


nikomatsakis (Jun 07 2019 at 14:41, on Zulip):

PS, @pnkfelix, @centril had some thoughts about maintenance/triage they was relating to me yesterday. TL;DR was whether we should consider doing things like beta-backports another way

centril (Jun 07 2019 at 14:42, on Zulip):

Yea; talking through beta-nominations seems to be taking a lot of time atm

centril (Jun 07 2019 at 14:43, on Zulip):

maybe it would be better to do things more async with a yay/nay process on the issues?

centril (Jun 07 2019 at 14:43, on Zulip):

Or at least you could start with asking folks async before the meetings on the PR and then they can leave "votes" and then you can consider that on the meeting

pnkfelix (Jun 07 2019 at 14:44, on Zulip):

in principle I'm in favor of doing more things async

centril (Jun 07 2019 at 14:45, on Zulip):

Another thing you could try to gain in meeting time efficiency re. ICEs is that you could empower the release team triage process to label things with P-medium for more obvious things

pnkfelix (Jun 07 2019 at 14:45, on Zulip):

but sometimes I do get the idea that there is a kind of group-mind that can emerge

pnkfelix (Jun 07 2019 at 14:45, on Zulip):

especially with things like evaluating backports

pnkfelix (Jun 07 2019 at 14:45, on Zulip):

where a bunch of people are in favor, but then someone brings up a point, and it causes a shift in mindset

pnkfelix (Jun 07 2019 at 14:46, on Zulip):

of course this can happen with any decision making process, including a number that the project is already doing asynchronously

centril (Jun 07 2019 at 14:46, on Zulip):

right

pnkfelix (Jun 07 2019 at 14:46, on Zulip):

(I also want to try to ensure there is a fair amount of social pressure against approving backports...)

pnkfelix (Jun 07 2019 at 14:47, on Zulip):

regarding ICE's, I do think we need to put together some guidelines that would allow others, like the release team, to do more automatic labelling

pnkfelix (Jun 07 2019 at 14:47, on Zulip):

things like "if you're confident that this only arises under a feature-gated thing, then its P-medium"

pnkfelix (Jun 07 2019 at 14:48, on Zulip):

but we should probably collectively discuss those rules

centril (Jun 07 2019 at 14:48, on Zulip):

Yea that sgtm

centril (Jun 07 2019 at 14:48, on Zulip):

how do you want to approach that discussion?

pnkfelix (Jun 07 2019 at 14:51, on Zulip):

i don't know yet

pnkfelix (Jun 07 2019 at 14:51, on Zulip):

it probably would make sense to just fold it into the design meeting planned for June 28th

centril (Jun 07 2019 at 14:52, on Zulip):

In that case we should involve the release team members also I think

centril (Jun 07 2019 at 14:53, on Zulip):

cc @simulacrum

simulacrum (Jun 07 2019 at 14:53, on Zulip):

I can't make the 28th personally but in principle, yes, I agree

simulacrum (Jun 07 2019 at 14:53, on Zulip):

we can bring it up at next meeting

nikomatsakis (Jun 07 2019 at 14:53, on Zulip):

in principle I'm in favor of doing more things async

my concern is that we need to have some set apart time

simulacrum (Jun 07 2019 at 14:54, on Zulip):

(if compiler team would be interested)

nikomatsakis (Jun 07 2019 at 14:54, on Zulip):

basically I think async works great until people get overloaded :)

nikomatsakis (Jun 07 2019 at 14:54, on Zulip):

and..we are overloaded...

nikomatsakis (Jun 07 2019 at 14:54, on Zulip):

(not everyone, though)

nikomatsakis (Jun 07 2019 at 14:54, on Zulip):

regarding ICE's, I do think we need to put together some guidelines that would allow others, like the release team, to do more automatic labelling

this seems very good tho

nikomatsakis (Jun 07 2019 at 14:54, on Zulip):

I mean I think just in general forming guidelines even just for ourselves is super useful

nikomatsakis (Jun 07 2019 at 14:55, on Zulip):

we have some kind of "heuristics" we apply but I don't think we've ever written them down (similar to the one you mentioned, i.e., does it affect stable code, etc). Similarly, we seem to value preventing new regressions higher than fixing old ones (righly or wrongly, hard to say, but there's a logic to it:)

centril (Jun 07 2019 at 15:10, on Zulip):

Yea; I think the release team could also use some written down heuristics :+1:

nikomatsakis (Jun 21 2019 at 12:21, on Zulip):

@pnkfelix so I'll hijack this topic for comments on your edits to the maintenance/triage hackmd

nikomatsakis (Jun 21 2019 at 12:21, on Zulip):

first off, I really like the P-daily etc, although I think maybe we should call it F- ("frequency")

pnkfelix (Jun 21 2019 at 12:21, on Zulip):

that makes sense

nikomatsakis (Jun 21 2019 at 12:21, on Zulip):

the issue of I-nominated (and, for that matter, the priority) tags not being sufficiently targeted is annoying, I'm not sure how much of a problem it is in practice, but at least maybe a bit

pnkfelix (Jun 21 2019 at 12:22, on Zulip):

(Also, F-<frequency> would allow us gradual migration from P-priority to F-frequency; rather than being in a state where some bugs use one semantics and others use another)

nikomatsakis (Jun 21 2019 at 12:23, on Zulip):

Ah, yes, interesting

pnkfelix (Jun 21 2019 at 12:23, on Zulip):

I think the problem of non-targetted nomination is only an issue for a small subgroup, namely the triagers for each team.

pnkfelix (Jun 21 2019 at 12:23, on Zulip):

so while its super annoying for me

pnkfelix (Jun 21 2019 at 12:23, on Zulip):

better that I alone be annoyed

pnkfelix (Jun 21 2019 at 12:24, on Zulip):

than annoy our whole contributor community with an even larger slew of labels

nikomatsakis (Jun 21 2019 at 12:24, on Zulip):

the harder questions of course are kind of... what are we going to do about things that have to get fixed, and how are we going to do those visits

pnkfelix (Jun 21 2019 at 12:24, on Zulip):

I am curious about what other teams do, actually

nikomatsakis (Jun 21 2019 at 12:24, on Zulip):

I'm pondering your comments regarding "best use of time"

pnkfelix (Jun 21 2019 at 12:25, on Zulip):

I know of course that Centril does a fair amount of triage and paper doc construction to organize the T-lang nominations

nikomatsakis (Jun 21 2019 at 12:25, on Zulip):

Some part of me thinks "we need the whole hour, and trying to squeeze in wg checkins hasn't worked out that great" -- though I think the wg checkins have come up with good stuff

pnkfelix (Jun 21 2019 at 12:25, on Zulip):

I think we're still learning how to even do the wg-checkins

pnkfelix (Jun 21 2019 at 12:25, on Zulip):

the principle is really good

nikomatsakis (Jun 21 2019 at 12:25, on Zulip):

Yes, I agree

nikomatsakis (Jun 21 2019 at 12:25, on Zulip):

On both counts

nikomatsakis (Jun 21 2019 at 12:25, on Zulip):

One thing I've been wondering about that feels related

pnkfelix (Jun 21 2019 at 12:25, on Zulip):

but I know I've failed to follow-through on ensuring that people are ready to do them

nikomatsakis (Jun 21 2019 at 12:26, on Zulip):

So there is the "meta" working group

pnkfelix (Jun 21 2019 at 12:26, on Zulip):

in part because we haven't set up the schedule

nikomatsakis (Jun 21 2019 at 12:26, on Zulip):

We've been kind of getting .. not to the end of the work we had in mind, but sort of to the end :)

nikomatsakis (Jun 21 2019 at 12:26, on Zulip):

And I've been wondering whether it makes sense to have a "standing" working group that is dedicated to meta issues -- or triage issues, perhaps. Basically to "making things run"

pnkfelix (Jun 21 2019 at 12:26, on Zulip):

didn't @davidtwco have a PR to automatically construct a schedule for WG-checkin's ?

nikomatsakis (Jun 21 2019 at 12:27, on Zulip):

Yes, they do, it's blocked on another PR that's been making good progress to overhaul the compiler-team repository

pnkfelix (Jun 21 2019 at 12:27, on Zulip):

i see, okay.

pnkfelix (Jun 21 2019 at 12:27, on Zulip):

I probably should join WG-meta

nikomatsakis (Jun 21 2019 at 12:27, on Zulip):

What I'm getting at here is

pnkfelix (Jun 21 2019 at 12:27, on Zulip):

or at least read what its doing more actively

nikomatsakis (Jun 21 2019 at 12:27, on Zulip):

Well I'm wondering whether it would make sense for triage to be done by a more focused group

nikomatsakis (Jun 21 2019 at 12:28, on Zulip):

But I think having that group be you is ungreat

nikomatsakis (Jun 21 2019 at 12:28, on Zulip):

I guess the question is

nikomatsakis (Jun 21 2019 at 12:28, on Zulip):

I think the aim would be to look at the bugs at the given frequency and ensure that "progress is being made" -- and if not, to open up (I think) a dedicated Zulip topic and ping the assignees --

nikomatsakis (Jun 21 2019 at 12:28, on Zulip):

the hard part is going to be the bugs that aren't assigned, though

nikomatsakis (Jun 21 2019 at 12:29, on Zulip):

(I mention the dedicated topic because I think it's harder to process pings that come from a massive "pre-triage" topic, at least I find that to be the case)

Pietro Albini (Jun 21 2019 at 12:29, on Zulip):

btw can y'all avoid F-*? the release team was thinking of using it for feature gates

nikomatsakis (Jun 21 2019 at 12:29, on Zulip):

How about Frequency-Foo :)

pnkfelix (Jun 21 2019 at 12:29, on Zulip):

Freq

nikomatsakis (Jun 21 2019 at 12:29, on Zulip):

Or that

Pietro Albini (Jun 21 2019 at 12:30, on Zulip):

Hz-

pnkfelix (Jun 21 2019 at 12:30, on Zulip):

Let your Freq flag fly

nikomatsakis (Jun 21 2019 at 12:30, on Zulip):

I kind of love Hz

pnkfelix (Jun 21 2019 at 12:30, on Zulip):

I won't veto Hz

nikomatsakis (Jun 21 2019 at 12:30, on Zulip):

despite it's utter obscurity :)

pnkfelix (Jun 21 2019 at 12:31, on Zulip):

struggling to work in a Car Rental joke...

nikomatsakis (Jun 21 2019 at 12:31, on Zulip):

Re: finding assignees

nikomatsakis (Jun 21 2019 at 12:31, on Zulip):

One thing we've talked about but never done

nikomatsakis (Jun 21 2019 at 12:31, on Zulip):

is trying to make a group of "icebreakers"

nikomatsakis (Jun 21 2019 at 12:31, on Zulip):

I feel like there's a power in making a named group of folks with a job :)

pnkfelix (Jun 21 2019 at 12:31, on Zulip):

ah yeah

nikomatsakis (Jun 21 2019 at 12:31, on Zulip):

I've also talked about, but never quite gotten organized enough to do,

nikomatsakis (Jun 21 2019 at 12:32, on Zulip):

creating some kind of "help wanted" thread on internals

nikomatsakis (Jun 21 2019 at 12:32, on Zulip):

though I had two distinct things in mind for it

pnkfelix (Jun 21 2019 at 12:32, on Zulip):

i did at least start a dedicated topic at some point to go through the issues labelled ICE

nikomatsakis (Jun 21 2019 at 12:32, on Zulip):

the idea being to have a dedicated place that we post stuff

nikomatsakis (Jun 21 2019 at 12:32, on Zulip):

and then trying to draw attention to it, perhaps via a twitter account or other things

nikomatsakis (Jun 21 2019 at 12:32, on Zulip):

yeah so I think ICE-breakers is a good idea, but maybe there are two visions for it

nikomatsakis (Jun 21 2019 at 12:32, on Zulip):

what I was thinking is that it might be useful to have:

pnkfelix (Jun 21 2019 at 12:33, on Zulip):

oh I just thought of something to add to the hackmd doc

pnkfelix (Jun 21 2019 at 12:34, on Zulip):

The question of how to prioritize bugs in feature-gated code where the feature is on the "short list" of things to be stabilized soon

nikomatsakis (Jun 21 2019 at 12:34, on Zulip):

Definitely one of the challenges is always the feature-vs-maintenance struggle

nikomatsakis (Jun 21 2019 at 12:34, on Zulip):

The question of how to prioritize bugs in feature-gated code where the feature is on the "short list" of things to be stabilized soon

Ah, yes, so -- more generally I think trying to write down our prioritization rules

nikomatsakis (Jun 21 2019 at 12:34, on Zulip):

we have some kind of "heuristics" we apply but I don't think we've ever written them down (similar to the one you mentioned, i.e., does it affect stable code, etc). Similarly, we seem to value preventing new regressions higher than fixing old ones (righly or wrongly, hard to say, but there's a logic to it:)

i.e., this comment from before

nikomatsakis (Jun 21 2019 at 12:36, on Zulip):

Maybe I'll try to edit the hackmd, @pnkfelix

nikomatsakis (Jun 21 2019 at 12:36, on Zulip):

Definitely one of the challenges is always the feature-vs-maintenance struggle

More on this: e.g., I've been feeling a lot of pressure to try and get async-await going

nikomatsakis (Jun 21 2019 at 12:37, on Zulip):

Though I think maybe that's a bit artificial-ish.

nikomatsakis (Jun 21 2019 at 12:37, on Zulip):

(On this topic, it feels like the compiler team should be more invested in that conversation)

nikomatsakis (Jun 21 2019 at 12:37, on Zulip):

Or at least some part of it

pnkfelix (Jun 21 2019 at 12:39, on Zulip):

hmm

pnkfelix (Jun 21 2019 at 12:39, on Zulip):

weird, I thought I had put something in the hackmd mentioned that we need to even just document the systems we are trying to do

nikomatsakis (Jun 21 2019 at 12:39, on Zulip):

hmm yes I think you did

pnkfelix (Jun 21 2019 at 12:39, on Zulip):

but maybe I ended up not writing that down, or thought it was somehow part of the "automation" discussion

pnkfelix (Jun 21 2019 at 12:40, on Zulip):

oh oh yes I did put it there

nikomatsakis (Jun 21 2019 at 12:40, on Zulip):

yeah there's some notes, I guess I expected a separate heading somehow

nikomatsakis (Jun 21 2019 at 12:40, on Zulip):

I'm thinking back to this idea of more dedicated groups

nikomatsakis (Jun 21 2019 at 12:41, on Zulip):

This is not the original intent of working groups, in that it's not "short term", but that's ok

nikomatsakis (Jun 21 2019 at 12:41, on Zulip):

I think the danger is that you can't create people from thin air

nikomatsakis (Jun 21 2019 at 12:41, on Zulip):

Or .. I expect people might show up, but to get bugfixing happening there is some need of expertise

nikomatsakis (Jun 21 2019 at 12:41, on Zulip):

(Sometimes)

pnkfelix (Jun 21 2019 at 12:41, on Zulip):

right, and acting like the existence of new structure will somehow magically make things smoother is ... questionble

nikomatsakis (Jun 21 2019 at 12:41, on Zulip):

Still I feel like there is a power to naming activities directly and giving them weight

nikomatsakis (Jun 21 2019 at 12:47, on Zulip):

right, and acting like the existence of new structure will somehow magically make things smoother is ... questionble

so @pnkfelix the obvious place to start though is with a triage group

nikomatsakis (Jun 21 2019 at 12:47, on Zulip):

advantages:

nikomatsakis (Jun 21 2019 at 12:47, on Zulip):
nikomatsakis (Jun 21 2019 at 12:47, on Zulip):

well, experience is relevant, but does't require deep knowledge of the compiler, maybe? at least less :)

pnkfelix (Jun 21 2019 at 12:49, on Zulip):

yes

pnkfelix (Jun 21 2019 at 12:49, on Zulip):

you are absolutely right; given that I'm going on leave for a month in mid-July

pnkfelix (Jun 21 2019 at 12:49, on Zulip):

we probably should try to spin up a triage group, mmm, yesterday.

nikomatsakis (Jun 21 2019 at 12:50, on Zulip):

confirm

nikomatsakis (Jun 21 2019 at 12:51, on Zulip):

the truth is that with both you + I heading for vacation this summer, we don't have much of a backup plan, do we ?

nikomatsakis (Jun 21 2019 at 12:51, on Zulip):

/me sighs

pnkfelix (Jun 21 2019 at 12:52, on Zulip):

yeah, and I'm not sure what @nagisa 's availability is going to be.

pnkfelix (Jun 21 2019 at 12:52, on Zulip):

I know @oli will be unable to contribute as much time as in the past...

oli (Jun 21 2019 at 12:53, on Zulip):

/me needs a cloning machine

nikomatsakis (Jun 21 2019 at 12:55, on Zulip):

/me sympathizes

nagisa (Jun 21 2019 at 12:57, on Zulip):

Discuss my availability now!

nagisa (Jun 21 2019 at 12:57, on Zulip):

advertisement jingle

nagisa (Jun 21 2019 at 12:57, on Zulip):

I can run triage meetings over the summer, sure.

nikomatsakis (Jun 21 2019 at 12:58, on Zulip):

oh, really?

nikomatsakis (Jun 21 2019 at 12:58, on Zulip):

I'm going to be in Greece July 15 -- August 9 I think? Not sure about @pnkfelix, so we're realizing we should figure out a plan

nagisa (Jun 21 2019 at 12:59, on Zulip):

yeah, I’ve no plans to go on any sort of leave this summer so far.

pnkfelix (Jun 21 2019 at 12:59, on Zulip):

I'm going to be on parental leave from July 13th through August 23rd

pnkfelix (Jun 21 2019 at 13:00, on Zulip):

(or at least that's what I've told my wife)

pnkfelix (Jun 21 2019 at 13:00, on Zulip):

or wait, no, through the end of august

nagisa (Jun 21 2019 at 13:01, on Zulip):

I definitely am able to drive the few meetings between July 15th and August 9th, we can decide how to split the load once Niko’s back.

pnkfelix (Jun 21 2019 at 13:02, on Zulip):

okay

pnkfelix (Jun 21 2019 at 13:02, on Zulip):

@nagisa would you be interested in being a member of a WG-triage?

pnkfelix (Jun 21 2019 at 13:02, on Zulip):

my intention is mostly that you could help onboard other people

nagisa (Jun 21 2019 at 13:03, on Zulip):

Hmm… sure, I guess?

Pietro Albini (Jun 21 2019 at 13:05, on Zulip):

btw, just to avoid naming conflicts, t-release is also creating a triage-wg :sweat_smile:

pnkfelix (Jun 21 2019 at 13:05, on Zulip):

hmm

davidtwco (Jun 21 2019 at 13:06, on Zulip):

Definitely one of the challenges is always the feature-vs-maintenance struggle

More on this: e.g., I've been feeling a lot of pressure to try and get async-await going

Sorry to backtrack a little to this message:

@nikomatsakis, I was wondering by this if you meant because of the time-pressure that's associated with async-await or something else?

pnkfelix (Jun 21 2019 at 13:06, on Zulip):

would there ... be any value ... in having that (t-release's triage WG, and t-compiler's triage WG) be a single WG?

nagisa (Jun 21 2019 at 13:06, on Zulip):

Consider consolidating them? I can’t imagine triage for T-release being too different from triage for T-compiler stuff.

Pietro Albini (Jun 21 2019 at 13:07, on Zulip):

release team triage is PR triage (aka pinging people until they review) and issue triage (aka applying the initial labels)

Pietro Albini (Jun 21 2019 at 13:08, on Zulip):

not sure how much that'll overlap with t-compiler triage

Pietro Albini (Jun 21 2019 at 13:08, on Zulip):

@simulacrum ^^

nikomatsakis (Jun 21 2019 at 13:09, on Zulip):

I was wondering by this if you meant because of the time-pressure that's associated with async-await or something else?

Yes, partly -- I would really like to see us get async-await out in the world. But I think in general there's always a bias towards 'building new things'.

davidtwco (Jun 21 2019 at 13:10, on Zulip):

I see, thanks for clarifying.

davidtwco (Jun 21 2019 at 13:13, on Zulip):

I feel like since I've been around (which admittedly isn't too long and I'm not that involved), there's been a trend towards deadlines, pushing for things to get in releases, just more time pressure in general - there was obviously the edition, but also async/await's (to the best of my knowledge) arbitrary deadline. It contrasts with working on NLL which felt more when-it's-ready (though that could just be me looking back on it).

davidtwco (Jun 21 2019 at 13:14, on Zulip):

Though, I think that might be slightly off-topic, apologies.

nikomatsakis (Jun 21 2019 at 13:15, on Zulip):

I think it's very much on topic

oli (Jun 21 2019 at 13:15, on Zulip):

But I think in general there's always a bias towards 'building new things'.

Don't take this as a complaint: not getting const eval RFCs merged or tracking issues approved for stabilization actually helped me to clean up stuff instead.

I agree though that focussing on features comes much easier, though actually working on it isn't necessarily more fun than refactoring

nikomatsakis (Jun 21 2019 at 13:15, on Zulip):

I'm not sure that trend is altogether good -- I've been revisiting it slightly. I think that it's a good idea for us to focus on what do we need for an MVP

nikomatsakis (Jun 21 2019 at 13:16, on Zulip):

@oli are you disputing my statement about the bias towards new things? i.e., because it's been hard to get const eval RFCs merged etc?

nikomatsakis (Jun 21 2019 at 13:16, on Zulip):

(I do think that's an issue too, it's very hard to keep up with review load, and I personally am never satisfied with how well I am able to do so)

nikomatsakis (Jun 21 2019 at 13:17, on Zulip):

I'm not sure that trend is altogether good -- I've been revisiting it slightly. I think that it's a good idea for us to focus on what do we need for an MVP

to elaborate slightly, but maybe not to try too hard to set dates for that MVP. But some amount of pressure will always be there I guess, and you need it to actually ship.

oli (Jun 21 2019 at 13:17, on Zulip):

I'm not disputing that there's bias towards building new things, I just want to state that (for me) implementing new things isn't necessarily more fun than cleanups and fixes, but easier to focus on

nikomatsakis (Jun 21 2019 at 13:17, on Zulip):

I see

nikomatsakis (Jun 21 2019 at 13:18, on Zulip):

Yeah. I'm partly just commenting from myself. In theory, I want to spend time on docs + regressions. In practice, I often feel some pressure to do other things.

oli (Jun 21 2019 at 13:22, on Zulip):

for me it's "regression fixes are merged fast" and everything else isn't, so I have like 7 trees of everything else and keep juggling that one tree for fixes (adding more trees doesn't help). I need to figure out how to look for fixes whenever a tree opens instead of reacting to pokes. Most pokes are "can I haz feature X".

davidtwco (Jun 21 2019 at 13:24, on Zulip):

Most pokes are "can I haz feature X".

I've felt (purely from observing conversations here and in Discord) that there have been more comments like this - people checking in on what the status of feature X or Y is - in recent months. Which, while I'm sure are well-intentioned, I could see would put pressure on people to focus on those things.

pnkfelix (Jun 21 2019 at 14:00, on Zulip):

@nikomatsakis BTW, we should ensure we talk about/advertise/establish the hypothesized WG-triage, at least for T-compiler, prior to next week's design meeting.

nikomatsakis (Jun 21 2019 at 14:01, on Zulip):

Yes. Not sure what the best "venue" is -- perhaps just doing a bit of a public call for a "first meeting"? We would want to have some idea what we are doing in tha tmeeting -- but a special meeting to review status of regressions might be good

pnkfelix (Jun 21 2019 at 14:03, on Zulip):

I guess a relevant Q is whether the hypothetical members of such a group would want to make sure they get a chance to participate in next week's design meeting

pnkfelix (Jun 21 2019 at 14:03, on Zulip):

since I imagine that design meeting will influence the scope/mission-statement for WG-triage

pachi (Jun 21 2019 at 14:09, on Zulip):

wrt triage, GNOME has a Bugsquad group and its webpage on the wiki https://wiki.gnome.org/Bugsquad may have interesting ideas, such as triaging guides, Bug days, and some other interesting ideas.

nikomatsakis (Jun 21 2019 at 14:20, on Zulip):

since I imagine that design meeting will influence the scope/mission-statement for WG-triage

right, there's kind of a cyclic relationship -- but I don't see that it hurts to "Take a stab" at defining the group first

nikomatsakis (Jun 21 2019 at 14:20, on Zulip):

wrt triage, GNOME has a Bugsquad group and its webpage on the wiki https://wiki.gnome.org/Bugsquad may have interesting ideas, such as triaging guides, Bug days, and some other interesting ideas.

interesting!

pachi (Jun 21 2019 at 14:40, on Zulip):

@nikomatsakis in fact, there are some interesting talks by Luis Villa (which IIRC started the Bugsquad) about bugmasters and triaging here: https://lu.is/talks/
- Why Everyone Needs A Bugmaster (talk, paper) (Linuxconf.au 2005)
- How GNOME Learned To Stop Worrying and Love The Bug (talk, paper) (Ottawa Linux Symposium 2003)
- Bugzilla Sanity (GUADEC 2002)
Glad it helps!

Last update: Nov 22 2019 at 05:25UTC