Anaconda storage Test Day tomorrow

It's Test Day time again. Tomorrow - Thursday 2009-10-01 - is Anaconda storage Test Day. Due to some changes in approach, this Test Day has had a bit of an identity crisis, but the end result is we're basically looking for testing of all possible storage scenarios during installation except the use of software RAID, which will be covered in its own special Test Day the following week.

Of course, as this is an installation-related Test Day, unlike most Test Days you will require a spare system, partition or virtual machine to which to install. Ideally we want testing on as wide as possible a range of storage configurations - different disk controllers, hard disks, partition layouts, LVM configurations and so forth. Many possible test cases are available on the Wiki page. Most people remember that Anaconda's storage code was extensively rewritten for Fedora 11, so we'd like to have lots of testing in Fedora 12 so we can be sure to be back up to the usual standards of support. Please come out and help test if you can!

As always, the Test Day runs all day in #fedora-test-day on the Freenode IRC network, and both QA team members and Anaconda developers will be present to help work on any problems that arise.

Miscellaneous annoyance of the day

I bought Dollhouse season 1 on Blu-Ray recently (not a recommendation, yet - I actually didn't watch it at all on TV, partly because of the poor initial reviews, but I tend to watch more TV in winter due to the weather anyway and thought I really ought to check it out). Sat down to watch the first episode last night, and marvelled as I had to sit through six different trailers - which could be individually skipped, but from which it's not allowed to skip to the top menu - before I could actually get in and watch an episode.

Hello? Fox? I just paid you my hard-earned money to buy the show. It's bad enough that you did such a cheap job of the packaging (there's not even a fricking episode guide insert), you absolutely do not get to spam me with irritating, experience-breaking trailers every time I load up the fricking disc. Please get a clue or I shan't be buying any more of your products in future (they are available in much more convenient form, you know. I'm sure you know what I mean). Thank you.

Side note: you may notice I've switched back to the old theme for the blog, and there's no ads any more. I've been meaning to stop running the ads ever since I got my new job, just didn't get around to it yet. Switching the theme back seemed kind of appropriate for switching back to just being a little personal blog site...

You know...

You know you've been watching too much sport when you see one of those commercials for the channel where they string together half a dozen highlights from all sorts of different events...and you realize you saw every single one live when it first happened.

(Also, when said commercial is in the middle of Poker Sports Star Challenge IV).

GMA 500 / Poulsbo driver for Fedora 11: soon to be in RPM Fusion!

Thanks to the folks at RPM Fusion for accepting my entirely private individual side (look at the timestamp on this post!) as a packager, my packaging of the Ubuntu-sourced GMA 500 / Poulsbo graphics driver - previously available in a repository I provided myself, as documented in this post - will be moving into RPM Fusion's non-free repository. I sent the builds through a few hours ago; they should show up in the next set of packages pushed to the RPM Fusion non-free updates-testing branch for F11, if I'm following the procedure correctly.

While doing this, I updated the kernel module package somewhat with some patches from Mandriva - many thanks to Olivier Blin for those. In return I gave him a couple of pointers from my builds. Ain't collaboration great? I also found a few minutes to test kernel 2.6.31 separately from X server 1.7; on my Vaio P, with a mock rebuild of kernel 2.6.31-40 but the regular F11 X server packages, everything works OK. So we're a-OK for kernel 2.6.31 on Fedora 11 whenever that happens, which is great. I still don't have any magic answers for X server 1.7 (Fedora 12), though. Sigh.

Sexism debate

Just wanted to reprint and expand upon something I wrote as a comment on the Geek Feminism blog post about Mark Shuttleworth's sexist comment. To preface - this isn't particularly about Mark at all. It was a slip of the tongue, it happens to everyone, and when you make a little mistake you say sorry and everyone moves on. I just wanted to consider the wider issues around the topic. Here's my comment:


Matthew: to me your comments seem a bit indicative of a kind of cognitive dissonance that’s happening whenever this topic comes up. It seems like, whenever the issue is raised in the abstract, there’s near-universal agreement. Everyone seems to agree that:

a) there’s very few women involved in F/OSS b) that’s a bad thing c) it’s probably due to sexism, of some kind – few people seem to think it’s intentional and malicious, but still sexism by any reasonable definition

So far, so good, everyone’s recognizing a problem and pointing in the same direction. However, whenever anyone tries to raise any specific instance of the kind of sexism that, all the instances taken together, tends to exclude women from F/OSS, there seem to be people who say either ‘no, that’s not sexism, that’s just a guy speaking from experience / politically incorrect’ or ‘oh, come on, it’s just one little thing, don’t get so worked up!’, to paraphrase.

This just doesn’t add up. If you take that attitude at face value, apparently we all agree there’s sexism in F/OSS, it’s a bad thing both absolutely and in practical terms (it’s holding back F/OSS from being all it can be), and we should do something about it – but no actual practical instances of sexism really exist, or if they do, they’re so trivial that no-one should really get worked up about it.

How’s that going to work, then?

If we’re going to accept the big – yet paradoxically easy, because it’s abstract – proposition that sexism in F/OSS exists and should be tackled by people modifying their behaviour, we’re going to have to start actually listening when people start trying to point out exemplary instances of the kinds of behaviour that are problematic and need to be changed, rather than taking each example in isolation and trying to pick it apart or denigrate its individual significance.


To expand, there really is a lot of remarkably woolly thinking going on around this area. (Thinking that woolly would get very short shrift in any other context in the F/OSS world, I suspect). To line up and knock down common variations on the theme:

i) Person X (who made sexist remark Y) isn't sexist, so it's no big deal

This is usually true, so far as it goes. I've encountered quite a few people in the community who find interacting with women awkward, but very few who seem overtly, maliciously sexist - who don't think women are capable of F/OSS work and actively and consciously want to keep them away. However, that doesn't justify the conclusion. People who aren't sexist in this way say sexist things, and when that happens, the correct response isn't just to ignore it and accept the consequences because they're a 'good guy'. We don't do this in any other context, after all. If a smart person says a dumb thing, we don't just let it go and accept whatever dumb consequences result because we don't want to hurt their feelings. We call them out, and that's exactly the way it should be. Smart Person makes a booboo, community says 'you idiot, that's dumb', Smart Person says 'whoops, yeah, I must've been on crack', everyone moves on with their life. No drama, but the outcome is correct. Inadvertently making a sexist comment doesn't make you sexist, and no-one's saying it does. But by the same token, the correct response is not just to ignore it and carry on, but for others to point it out, and for the person who made the remark to say 'whoops, uh, sorry, didn't mean that'.

ii) Sexism isn't the reason girls aren't interested in F/OSS, so it's not a problem

This is wrong for two reasons. One, the proposition is just factually very hard to defend. It's pretty tough to support the suggestion that some facet of the biology which separates males from females means that females are...uh, my math foo is weak, but something like fifty times less likely to become interested in F/OSS (the statistic I'm working off is that 97-98% of F/OSS developers are male). It's far more credible to assert that it's not purely sexism strictly within the F/OSS world itself which produces this result, but wider sociological factors - but so what? Sexism in F/OSS may only be a part, but that doesn't make it not a problem. Even if we can't solve the entire problem ourselves, that doesn't exclude us from the responsibility to stop being part of the problem and start being part of the solution.

Second, even if one were to grant that there's some mysterious biological reason why female people are intrinsically less likely to be interested in F/OSS, that doesn't justify or excuse sexism at all. It's bad on its merits. It should never be accepted in any situation. The logic behind this assertion just doesn't add up at all. It seems to require that this mysterious biological imperative is the sole reason that only a tiny minority of women become involved in F/OSS, and therefore any instance of what appears to be sexism isn't really sexism, or somehow - mysteriously - is sexism but doesn't actually discourage any women from becoming involved in F/OSS at all. Huh? I don't see where that works.

iii) Female Person X wasn't offended by the remark, so it wasn't sexist

Uh, nope. Doesn't follow. A sexist remark is not one that offends 100% of all women everywhere. That's all you need to say, really, about this objection.

iv) I just don't want to talk about this! can't we go back to talking about 'important technical stuff'?

It's easy to have a degree of sympathy with this position. I'm kind of there myself; I know how much effort and energy it takes to be an activist on this kind of issue, and I just don't have it myself. I'm not as diligent as I probably should be on picking this kind of thing up all the time. But...

there's several problems here, again. One, I'm frankly suspicious of the unspoken attitudes inherent in the person who cares enough to make a comment saying they don't care about an issue. That's...a little inconsistent, right there. You care so little about sexism that you registered to post a comment about how little you care? hum.

Second, I think this comment often means 'look, I'm really quite happy with the very male status quo, it's all working out quite nicely for me. I'm happy to concede, in vague abstract terms, that perhaps there could be rather more women involved in F/OSS, but if it involves - heaven forfend! - me having to lift a finger or even just read about something a bit uncomfortable, well, that's just not going to work for me, so please stop with your efforts to do something about the problem, I find it icky.' I hope it's self-evident why this is, frankly, the most spineless position of the lot.

Finally, this position also sometimes manifests itself as 'I don't want to read about this in Technical Form X', where X is a mailing list or planet or IRC channel or whatever. To which I say, well where the hell is anyone supposed to talk about it? Are we supposed to create Planet Sexism Is Bad, Yo! and put all the posts there, where only people who are already interested will read them, and all the people whose attitudes we actually need to change can go right ahead and completely ignore them? What's that going to achieve, then? How far would the civil rights movement have got if Martin Luther King had just sat around in a room with a bunch of other black people?

Martin Luther King: Man, racism sucks. Second Guy: Yup. Yup, it sure does. What are we going to do about it? MLK: We're going to sit here and talk to each other about how much it sucks, that's what we're going to do! Second Guy: Uh huh, that'll work well.

Unless the issue is talked about in the places where it's causing trouble, nothing is going to improve.

Final edit: in case all the above isn't enough, just remember that if you reckon it's all a bunch of silliness and everyone should just stop moaning and get on with it, you're lining yourself up with good ol' Sam Varghese. And let's face it, that's never a comfortable place to be. If I ever want to know what I think about something, I just look at what Sam thinks, and I can be pretty sure the diametric opposite is a good place to stand...

My personal experience is as an even more unusual minority in F/OSS than being a woman - I'm gay. (I've mentioned this before but I don't really make a point of it, so some people probably don't know). I can't even recall anyone else openly gay in the F/OSS community at all - I'm sure there are a few, but it's a very very small number. I've not really been much bothered by this, but I think it's because I've never seen F/OSS as a social venue for me, really. I like to do the work and contribute and I have fun interacting with many people in the F/OSS world on that level, but I don't really make my personal friends from this world, the people I'd hang out with on weekends. I think it does serve that function for quite a lot of the straight male people involved, and I think it's very unlikely to serve that function very well for lots of people - gay people and women included - as current attitudes stand. That's an exclusionary factor, for those who like to combine their hobbies, enthusiasms or work life with their social life.

News: Blocker bug review on Friday, video acceleration updates

Just some miscellaneous news items. This Friday, 2009-09-25, will be the third Fedora 12 Beta blocker bug review meeting: it's at 15:00 UTC in Freenode #fedora-bugzappers . Please do come along if you're interested in helping make Fedora 12 Beta release go as smoothly as possible. If you have any bugs you think are important to be fixed before the Beta is released, please do come along and propose them. All are welcome!

Video acceleration stuff: I sent an updated libva to my repository a couple of days ago. An updated mplayer-accelerated should be coming soon. I've been working with the RPM Fusion mplayer maintainer; hopefully we'll be able to update the Fusion package to be a recent mplayer snapshot and include VDPAU support, soon, now libvdpau is in Fedora. This will likely be only for F12. We're also planning to roll the libva support into the package for F13.

Gwenole has emailed me to let me know about gnash-vaapi, which is a modified version of gnash that supports VAAPI acceleration of h264-encoded Flash video (so, for e.g., YouTube HD videos). It only works with certain libva implementations, though, not including the Poulsbo one at present, which makes it tricky for me to test. It should work with NVIDIA via vdpau-video, apparently. I may try to get it packaged anyway.

Test Day update: Xfce Test Day Thursday 2009-09-24

Time for another Test Day update! This coming Thursday, 2009-09-24, will be Xfce Test Day. The Xfce spin of Fedora is popular among Xfce enthusiasts, and the maintainer Kevin Fenzi is always keen to have it as polished as possible for each Fedora release, so please come out to Freenode IRC #fedora-test-day and help us make sure Xfce is in good shape for Fedora 12. There will be a live image available for testing, and testing is pretty straightforward: all we need you to do is use Xfce just as you usually would, and check to make sure everything's working properly. We're also interested in seeing if the official Xfce browser, Midori, is good enough to be the default browser on the Xfce live spin. Thanks to all who come along to help test!

Sound / PulseAudio Test Day on Wednesday 2009-09-16

So after the excitement of last week's X.org Test Week, we have another big ticket Test Day coming up on Wednesday: sound Test Day. We will be doing some broad-based tests to try and make sure that typical audio use works properly on as wide as possible a range of hardware in Fedora 12. And yes, the Test Day includes PulseAudio. Here's your chance to test out whether the current Fedora 12 code and PulseAudio work properly with your sound card, and complain - appropriately, of course! - if they don't. As usual, we will be providing live images for testing, so you'll be able to test without installing anything permanently or affecting your regular Fedora (or other operating system!) at all. Testing will be very easy, just a case of booting the live image and testing a few simple sound-related operations. Just testing whether sound works at all out of the box will be very helpful. So please come along on Wednesday and help us test! Even if you can't make it on Wednesday, it will be very useful if you can run the tests and provide your feedback in the results table, either before or after the Test Day happens. The Test Day itself will be taking place all day in #fedora-test-day on Freenode IRC. If you're not sure how to use IRC, see this guide.

As with the graphics events, this Test Day may well be of interest to those outside the Fedora community as well: all distributions use much the same set of sound drivers, and most use PulseAudio these days, so identifying and fixing problems in Fedora Rawhide is very likely to help out other distributions too. Fedora Rawhide has the very latest PulseAudio code, so if you come along and help us test, you'll help make sure that code is nice and polished when it lands in your distribution of choice.

Please note that the wiki page isn't finalized yet; it's obviously missing the live images (which I haven't yet generated) and the test cases. These should appear tomorrow, but I wanted to get the word out about the event as soon as possible.

I'd like to say a big thanks to Lennart Poettering for being very positive about the idea of doing a Test Day, and getting involved with the organization. He'll be present on IRC for the Test Day, too (his IRC nickname is 'mezcalero'). Please don't use this as a venue for broad, unhelpful complaints or flames about PulseAudio: we're genuinely interested in fixing any specific issues that arise in PA as a result of the Test Day, but flamewars about its intrinsic value will only clutter up the channel, annoy everyone involved, and waste time. Thanks!

Intel Test Day tomorrow

Quick reminder - rounding out Graphics Test Week, tomorrow is Intel graphics Test Day. As with the Radeon and Nouveau events, we'll be in #fedora-test-day on Freenode IRC all day long, and live images are already available from the Wiki page so you can get down to testing. Please do spare some of your time to help test if you have an Intel graphics card! Thanks.

Nouveau Test Day tomorrow

A quick reminder that tomorrow (or today, for many people!) is Nouveau Test Day in #fedora-test-day on Freenode IRC. If you have an NVIDIA graphics card, please swing by, grab the live images from the Test Day page, and help test out nouveau for Fedora 12!