@Chris let me pick one of the thousand: the internet is getting DDoS'ed

@Gilles I don't know. I don't really follow Debian politics either, and I don't know JH personally, of course. I imagine it was not any particular thing, but a combination of things. He has been a DD for 18 years. Most people burn out long before this.

And of course, the whole systemd thing has caused a lot of stress to the project and been really unpleasant.

@derobert Funny.

@Gilles In any case, it is unfortunate. JH is/was an important person within the project. And losing an experienced developer is always a bad thing, anyway.

@FaheemMitha you also work for debian?

@Ramesh Not really, no. I did packaging for one piece of software, but the ftpmasters rejected it. And have been ignoring me since.

Not that I have tried that hard to get their attention.

I've submitted the occasional patch over the years. And many bug reports.

Anyway, JH was one of the few people left from the early days of Debian. Now, he is gone too.

JH interview Setup interview

JH- on leaving

Reddit comment thread on the JH resignation

@FaheemMitha We just need a JH simulator, then we could use git bisect to find the commit which broke Debian. Could be useful exercise.

@derobert I'm not ready to laugh yet. It is still a time of grief.

Laughing is the only way to stay sane reading a lot of -devel recently :-(

@derobert Laughing is good when dealing with lots of things in life.

Did you catch the post where someone attacked Russ Allbery for being the reason JH left Debian? Or something like that.

Haven't seen that one yet. Not surprised, though. (That kind of BS has been unfortunately normal on -devel, etc. for a while...)

@derobert It has? Glad I'm not following it, then.

This one is definitely short right?

@Ramesh yeah.

@Ramesh - definitely

Thanks. It showed up in the low quality answer. So, I thought I could instruct the answerer to consider expanding in case if he is not aware that one liners are generally not encouraged.

I was going to edit in its home page, but it appears dead.

@derobert Yes, I thought maybe I had hallucinated that email (it's not been a fun day) but there it is.

sigh. Yeah. Sounds like -devel. Or was it -project?

> if the question bores you because the asker put no effort into it and hasn't provided any clear guidelines for choosing a correct answer, then that's a good reason to close it. — Shog9

@derobert -devel. But most everyone in that thread (which I'm reading through now) pretty much ignored it. Since it was obviously a bit unhinged.

There are a lot of people who are (understandably) upset by this. Debian seems ready to have a wake, almost.

By this I mean the JH resignation, to be clear.

Yeah. JH has been around forever, and is very well known. And he's written a lot of the software that everyone in (and around) Debian uses.

@derobert He has. A big part of the reason why perl is in essential is because JH uses it. Or used to use it. He prefers Haskell now, apparently.

I don't get intimidated easily, but I once filed a bug report against etckeeper. JH told me it was not a bug. I still think I was right and he was wrong, but I didn't feel up to arguing with JH.

Is all this stuff ones reading about the big Systemd land-grab alarmist? And sorry about bringing up the S-word.

@Braiam starring this!

@FaheemMitha I've only been loosely following this JH business. Why did he leave?

@slm We don't know. We have what he has written, but it is less than clear. I posted the relevant links earlier, but they don't really tell you much.

As you know there has been much acrimony around systemd and the tech ctte decisions involving that. (I never knew init systems were so important. I basically ignore mine.)

yeah I read through them and thought I was missing something

@slm No, it just isn't clear. I think that in large part JH just decided this was the end, as sometimes people do.

He has been with Debian 18 years. That's a hugely long time. Many of the people who joined at the time he did are long gone.

the calling out of ppl as complicit too, I didn't know who the players were so it was even more confusing

Maybe most. It would be hard to get statistics.

ah ppl come and go. Sometimes it's better for everyone if they leave.

@slm That was people calling the tech ctte names. Russ is on the tech ctte. They have the final word in technical decisions involving Debian, and got dragged into the systemd thing, though I'm sure they didn't want to be.

@slm JH wasn't just anyone.

I know who JH is

@slm Ok

but sometimes it's better for a person that's doing everything and keeping things floating to leave. It forces the remaining folks to pick up the slack.

@slm Maybe, but there have to be enough people to pick up the slack. The main people in Debian are overstreched already.

get more ppl to help

As I mentioned a little while ago, I thought of writing an essay (or blog post) tentatively entitled "Debian and the Inverted Pyramid". About how few people really make Debian run.

@slm How? People who are capable of doing difficult technical work don't grow on trees.

Many important parts of Debian are basically run by one or two people.

instead of doing all the work themselves, they should be putting systems in place to expose the various tasks in a way so that other's can participate

Not news to anyone who follows the Project at all, and probably not unusual in free software, but still a big concern. See what has happened to aptitude for example.

@slm They should be, but for one reason or another it doesn't happen.

Well, sometimes it does. But for example debhelper doesn't have anyone to pick it up, afaik.

FOSS is very much a survival of the fittest.

Ideally the big important software like dpkg, apt, debhelper, debconf, the software that manages the archives should be managed by teams.

If it has value and is useful/important someone will step up.

right

@slm Not necessarily. apt is used by everybody. It is spent much of its lifetime unmaintained.

Sounds like they haven't done a good job of developing that aspect

Well, much is maybe an exaggeration.

@slm It's a social rather than a technical issue. Debian excels at technical stuff. Not so much at the other thing.

well s/w can get to a state where it's done.

Well, a good part of that is, I think, that apt basically worked. No one really had a huge feature that it was missing that they wanted to work on.

yeah so why continue maintaining something that is essentially done and works

@derobert Maybe, but it was still quite buggy and stuff. David has been doing a good job keeping up with the bugs, but what happens if/when he leaves? And nobody stays forever.

the source is there if an issue arises. Yum is similar to that

Part of the problem is that though these things are relatively small, they are quite complex. Hard to work with if you don't know what you are doing.

I mean, sure, apt has a billion bugs open against it. But how many have you run into? Any that'd be worth spending your time fixing?

the original author had passed it on to someone else who was partially maintaining it, but there's little new dev action on it, so it sits there

@slm Software is never done.

@derobert Sure, I've run into them. Still run into them. I don't report them as religiously as I used to, though.

@FaheemMitha sure it is, it's done enough that we don't care to put more effort into it

@slm Did you mean apt, or are you speaking more generally?

generally

@slm You mean aptitude?

I'm only speaking generally

I disagree. Software need maintenance, particularly important software. Otherwise it bitrots.

there's a subtle difference in having someone actively maintain it vs. passively maintaining it

System level stuff is important because often there is no substitute. I mean, you can't not use dpkg or apt.

software can go into a passive maintenance

@slm True. My point is, you need active maintenance.

is there someone actively maintaining ls?

Yeah. When it fails to compile because the compiler suddenly got stricter, you fix that bug. And otherwise leave it alone.

@slm ls in coreutils? dunno.

I doubt it

But ls is a comparatively small and simple program.

yeah, ls is part of coreutils

Anyway, I'm not talking about ls, or anything like it.

I picked that b/c it's a tool that's essentially done

@slm The stuff I'm talking about it rather bigger and much more complex. As I'm sure you know, software grows exponentially in complexity as its size increases.

Assuming a coupled system, of course.

@FaheemMitha which is a good argument to stop its growth. And say, not add a bunch of new features to dpkg or apt-get.

it's been my experience that even though the complexity of s/w can grow, you still do not have to have it be actively maintained, since it was typically built in a modular fashion, so long as you're familiar with the building of the s/w you can typically fix issues as they arise, w/o having to be a full on maintainer

@derobert Well, apt and dpkg are already big. And they need to be a certain size to fulfil their function.

I've supported a number of systems over the years where there were no original authors left and the software was essentially frozen, and from time to time when the underlying systems were updated that it ran on, this s/w would be rebuilt on the new systems and issues would be addressed then.

@slm Well, in the long term you need people who understand the software well, for when significant work need to be done. And at some point it does.

@slm Didn't anyone want it unfrozen?

I can promise you that I knew jack about these s/w systems when they were handed to me.

@slm Ok

often times in business they do not want to spend any more money on s/w once it covers a specific feature set

manufacturing esp.

need some help running a linux command

@slm Ok. Well, needs vary.

think of all the s/w that goes into running a manfacturing process

keep getting rm cannot remove ... permission denied

the s/w that ran on these systems was written 10-15 years ago. The hardware it originally ran on was no longer being manufactuered. Typically HP VAX type systems, so this s/w was moved from there to Solaris, and then from there to Linux on VMware later on.

These are large complex code bases

@DemCodeLines copy/paste exactly what's going on pls

@DemCodeLines - also you should be asking Q's on the mainsite, not in chat

apologies

Oooh! Nexus factory images are out.

@derobert What is that?

developers.google.com/android/nexus/images ... software images for Nexus phones. The Android L one is out now.

@derobert Why is this interesting?

@FaheemMitha Can upgrade my phone to Android L now.

Which I just did. Hopefully. Either that or I broke it :-P

Still booting.

DexFile_isDexOptNeeded failed to open oat file '/system/app/Hangouts/arm/Hangouts.odex' for file location '/system/app/Hangouts/Hangouts.apk': Failed to open oat filename for reading: No such file or directory

... wonder if that error is expected. Probably.

W/dex2oat ( 1069): Compilation of int com.google.android.location.activity.a.a.a(float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float) took 124.773ms ... wow that is a scary method signature!

@derobert And that is a good thing, presumably.