inamerrata

Joey replies to my post from yesterday with his own example of leaving:

The technical details remain as irrelivant today as they were at the time, the relevant problem being that developers who are not involved in the installer development rarely consider how their actions can affect it.

When a policy gets int the way of solving a problem, it’s time to leave, that’s all.

His posted logs suffer from unescaped angle brackets so aren’t very easy to follow; and also from the fact that for rather obvious reasons they don’t include any of the conversation that took place after he left. Extended logs and the other side of that story below the fold:

02:40 <joeyh> aj: sigh, did you realy have to break d-i yet again right as we’re trying ot release it?
02:40 <aj> ?
02:41 <aj> joeyh: (also, please try to avoid making things overly personal on this channel, cf the charter)
02:42 <joeyh> fine, did one have to ignore the d-i release and make a package upload with lots of changes that broke d-i internalls. Yes one did
02:42 <joeyh> for at least two values of one
02:42 <aj> joeyh: please explain what you’re talking about, in technical terms
02:43 <aj> joeyh: if you do so, it might get fixed
02:48 <aj> joeyh: hello?
02:50 <h01ger> joeyh, i _guess_ you’re referring to the debootstrap-upload of aj ?! /me wonders why an upload to unstable breaks the d-i beta, isnt that build+based on etch ?
02:50 <joeyh> sorry, busy fixing the beakage
02:51 <Sledge> looks like a debootstrap change is causing problems in CD builds
02:51 <joeyh> well, there has not been a usable combination of apt, debootstrap, and d-i in testing for well over a week
02:53 <aj> so you’re saying i didn’t break it just as you’re trying to release?
02:55 <joeyh> no you broke it just as the fixed NMU was ready to go into testing
02:56 <h01ger> .oO( ah )
02:56 <aj> sigh
02:56 <neuro> …and we’re still waiting for the technical details
02:56 <aj> joeyh: second warning. this channel’s for technical discussions, eg “here’s what broke: …”, not accusations, eg “you broke it” “why’d you have to break it” “yes one did”.
02:57 <aj> joeyh: if you want, there’s still time to put 0.3.1.9 into testing
02:58 -!- joeyh [joey@kitenet.net] has quit [Quit: Terminated with extreme prejudice – dircproxy 1.0.5]
03:01 <liw> when an operator here is involved in a dispute, might it be best to have someone other than them to do the warning off of people?
03:02 <aj> i’d love that
03:03 <Sledge> sorry, I was distracted looking into CD builds or I’d have done so… :-(
03:24 <fjp> I understand joeyh’s frustration. We have been this >< close to a beta for oh, the last three weeks. Every time we think the last major issue is fixed, whoops, there’s yet another upload or migration that breaks d-i in a major way and leaves us to trace the cause of the problem and most of the time provide the fix as well.
03:25 <aj> well, that’s great, but putting debootstrap 0.3.1.9 into testing is something that could happen right now if anyone could be bothered explaining what’s going on
03:25 <fjp> Not saying it’s anybody’s fault, it’s just frustrating as it takes a lot of energy tracking all components needed for a d-i releasy; especially now.
03:25 <fjp> aj: We will. Watch you mailbox for but reports…
03:27 <aj> well, no, this is a time limited thing; dinstall’s in just over an hour
03:27 <aj> and i should’ve already gone to bed
03:28 <aj> it’s been almost 50 minutes since joeyh’s initial complaint already, even
03:28 <fjp> aj: I doubt you’ll be able to help much today then. Maybe joeyh can do it on his own. I guess he’s either working on that or taking a hike :-)
03:29 <aj> huh?
03:29 <aj> i can put 0.3.1.9 in testing with one command
03:29 <fjp> OK. Please do. That gives us a fallback at least.
03:29 <aj> *sigh*
03:30 <aj> can someone else please try getting an explanation of what’s going on
03:30 * fjp looks up bug #
03:30 <fjp> aj: #335653
03:31 <fjp> aj: Sorry, just got home and catching up on everything.
03:35 <aj> fjp: debootstrap-udeb is already 0.3.1.9 in testing, that’s a 0.3.1.9 log…?
03:35 <aj> (or earlier, perhaps)
03:37 <aj> fjp: it’s missing ++ DEF_MIRROR=http://ftp.debian.org/debian
03:37 <aj> fjp: which would show up if it were 0.3.2
03:38 <fjp> In that case I’m probably unaware yet of the new problem…
03:38 <fjp> Hmm. Wonder how that can that have gone missing all of a sudden.
03:39 <fjp> Anyway. Personally I’m not going to be rushed by todays dinstall. Dinner first and then I’ll try to look into these problems as well.
03:40 <fjp> (Including the one debian-cd is apparently having.)
03:40 <aj> *shrug* if it’s not done by today’s dinstall, it’ll probably be done with the 2day SoE and it won’t be possible to revert
03:41 <fjp> aj: I see joeyh made this commit for base-installer:
03:41 <fjp> [18:50:02] <CIA-2> debian-installer: * Remove code that munged/demunged Packages and Release filename in the
03:41 <fjp> [18:50:04] <CIA-2> debian-installer: cache for debootstrap, since debootstrap no longer does that munging.
03:41 <fjp> [18:52:33] <CIA-2> debian-installer: joeyh packages * r31703 /base-installer/etch-beta1/debian/changelog: releasing version 1.35.2
03:41 <fjp> Hopefully that fixes the new problems.
03:44 <aj> for values of “no longer does that munging” approaching “undoes the munging itself” now
03:46 <fjp> Anyway, change in behavior with nasty consequences at an inconvenient time.
03:48 <aj> well, night
03:48 <Sledge> night…

Other relevant details — the conversation above took place on the morning of the 27th of October localtime here. The upload in question was debootstrap 0.3.2, made on the 23rd of October. The previous upload was debootstrap 0.3.1.9, made by Joey on the 19th of October, and which also entered testing on the 20th. The debootstrap udebs in testing (which are synced independently), remained at 0.3.1.7 until the 25th of October, at which point they were replaced with the 0.3.1.9 versions. They were then updated to 0.3.2 in the dinstall run after the above conversation, and then to 0.3.3 on the 30th of December.

Those changes were all mentioned on the #debian-release channel, since the d-i beta was a pretty high priority. On the 25th, there was the following:

20:34 <aj> hrm, joeyh pushed debootstrap in too
20:34 <aj> i guess that means it didn't break too horibbly
20:35 <vorlon> :)
20:35 <aj> # more changes than I would like, but unbreaks etch install
20:35 <aj>  ^- debootstrap
20:35 <aj> # necessary for d-i beta
20:35 <aj>  ^- apt

The comments are from the commands joeyh passed to the testing script to get apt and debootstrap updated in testing. A few hours later, I also said:

03:11 <aj> joeyh: i bumped debootstrap-udeb/testing to 0.3.1.9 (0.3.2 for m68k) since i figured the choice of 0.3.1.9 v 0.3.2 would probably be more useful than 0.3.1.7 v 0.3.2

Downgrading the udebs to the previous version is possible while they remain in the archive — which they do for about two days after they were last used. So the 0.3.1.9 versions left the archive in the run of the 29th.

The changes in that upload included three changes that were important for good installations: defaulting to –resolve-deps, catching failures when dpkg is invoked using –status-fd to provide progress information, and improved handling of log file capturing. They weren’t crucial, but they were also only uploaded after Joey’s NMU had already gone into testing.

Joey understates his level of self-recusement somewhat too; his departure wasn’t just from that channel, but IRC entirely, and he didn’t return even to other channels for quite some time.

This doesn’t really leave me anywhere useful: I did actually take a fair bit of care to avoid causing problems for d-i in advance as per the above, and months later I’m still not sure what actually went wrong, and Joey appears to still not be interested in actually telling anyone; so the only way I can prevent anything similar happening again is just not to work on debootstrap, or anything else that might be vaguely d-i related.

What’s going on here — what’s the point of just going up to someone and yelling at them, without even explaining why? Why should anyone put up with that shit?

Add that, the fact that just trying to talk about this stuff gets you called a “snarky asshole”, and Jacobo’s remarks from today (“the existence of a charter makes the environment almost as hostile as I can bear”) and I’m left not seeing any chance for any sort of “meeting of the minds” or quid-pro-quo here, just good odds that one way or another a whole bunch of people are going to end up following Joey’s and Jacobo’s advice: leave.