[00:00] <idnar> uh, so, I don't want to be That Guy, but this new font seems really terrible
[00:00] <idnar> kerning and hinting / smoothing seem totally out of whack, at least at the resolutions I have available to me
[00:02] <wgrant> idnar: Using a recent version of Firefox?
[00:02] <mwhudson> idnar: a change landed to improve things on ff overnight
[00:03] <wgrant> Oh look, it's deployable.
[00:03] <wgrant> Let's deploy it.
[00:03] <idnar> wgrant: I'm running latest Aurora on OS X, and a recentish Aurora on Linux
[00:04] <wgrant> Yeah.
[00:04] <wgrant> Firefox 6 and Chromium anything seem to fall back to a sensible font OK.
[00:04] <wgrant> But Firefox 7 and later don't.
[00:04] <wgrant> Not sure if it's a bug.
[00:05] <idnar> hmm, why would they fall back?
[00:05] <wgrant> We are only providing the medium weight, when we use bold and italic and bold-italic variants.
[00:05] <idnar> ah, I see
[00:05] <wgrant> The terrible font you see is probably the medium variant being scaled to bold.
[00:05] <idnar> yeha, the non-medium weights are really awful
[00:06] <idnar> the kerning seems off even for medium weight, though
[00:06] <wgrant> Hmm. How does font.ubuntu.com look?
[00:07] <wgrant> I don't have an OS X machine, but it looks mostly fine on Windows for me.
[00:07] <wgrant> As fine as Windows browser rendering tends to be.
[00:07] <idnar> font.ubuntu.com looks fine, but I can't seem to go smaller than 16 there
[00:07] <wgrant> What about the body text?
[00:08] <wgrant> Below the selectors.
[00:08] <wgrant> That's the same font, but smaller.
[00:08] <idnar> that seems to have the same problem
[00:09] <idnar> on my laptop which has a higher display resolution (as in dpi), it looks better, but spacing still seems weird
[00:09] <idnar> small clip from my desktop machine (running Debian / Aurora): http://bucket.mithrandi.net/launchpad-font.png
[00:11] <wgrant> Urgh, that's awful.
[00:12]  * wgrant hacks around in dconf, since GNOME 3 has removed the DPI option.
[00:13] <lifeless> do they hardcode 72dpi full stop, or actually trust the X metrics now ?
[00:13] <lifeless> cause 72dpi on a 144dpi screen is awful
[00:13] <wgrant> They used to hardcode 96dpi.
[00:14] <wgrant> It *seems* to be smaller now.
[00:14] <idnar> GNOME on my system uses the X metrics
[00:14] <idnar> er
[00:14] <idnar> wait
[00:14] <idnar> damnit.
[00:14] <wgrant> You might think that, but you could well be wrong :)
[00:15] <wgrant> For ages it completely ignored them.
[00:15] <idnar> resolution is 85dpi in GNOME, but xdpyinfo reports 95x94 dpi
[00:15] <wgrant> Then it used them for a while.
[00:15] <wgrant> Then it stopped.
[00:15] <wgrant> What if you zoom up one level in Firefox?
[00:16] <idnar> it looks different, but still not great
[00:16] <idnar> I think I need to restart Firefox for my font rendering to be completely fixed now that I've fiddled with the dpi settings
[00:17] <idnar> fwiw, in Chrome 15 (dev channel), I seem to get Ubuntu used for bold / italic text, same as in Firefox
[00:17] <StevenK> wgrant: Is this the fix that nigelb did that isn't deployed?
[00:17] <wgrant> Odd, Chrome 13 on both Windows and Linux works OK.
[00:17] <wgrant> StevenK: Yes.
[00:19] <idnar> hmmm, Firefox seems to be ignoring GNOME font rendering settings completely
[00:19] <idnar> so that made no difference
[00:20] <idnar> ignoring them for the content rendering, I mean (the UI chrome *is* affected)
[00:20] <wgrant> people.canonical.com/~wgrant/fonts-seem-ok.png is what it looks like for me when it's that small.
[00:20] <wgrant> Bah, Firefox now has Chromium disease.
[00:20] <wgrant> Who needs URI schemes anyway.
[00:20] <idnar> heh
[00:21] <idnar> it tries to put the scheme back in for you when you copy, but fails a lot of the time
[00:21] <wgrant> Yeah, the second time I copied it worked.
[00:25] <lifeless> users don't understand uris

[00:26] <wgrant> idnar: Is Chrome's rendering any better, ignoring the weight issue?
[00:29] <idnar> it seems about the same as Firefox (on my laptop, running OS X)
[00:30] <idnar> Firefox: http://mithrandi.smugmug.com/Other/Scratch/i-nFJGWXP/0/O/Screen-Shot-2011-09-09-at.png
[00:30] <idnar> Chrome: http://mithrandi.smugmug.com/Other/Scratch/i-nsHg9cR/0/O/Screen-Shot-2011-09-09-at.png
[00:30] <idnar> you can see that it's not rendered the same way, but I'm not sure it's any better or worse
[00:31] <wgrant> It seems yto be almost correctly kerned and hinted, but it's really really heavy.
[00:31] <wgrant> How odd.
[00:31] <wgrant> Since that page has all the variants.
[00:33] <wgrant> idnar: lenny, wheezy or sid?
[00:33] <wgrant> Er.
[00:33] <wgrant> s/lenny/squeeze/
[00:33] <idnar> sid on my desktop
[00:36] <wgrant> I don't have a recent desktop Debian VM around. I shall install one and try stuff out.
[00:37] <exarkun> lifeless: I have an aversion to destroying data.  I suppose I don't really need them, but I would feel better hiding them than removing them.
[00:49] <wgrant> idnar: http://people.canonical.com/~wgrant/lp-fonts/firefox-win/lp.png http://people.canonical.com/~wgrant/lp-fonts/firefox-win/font.ubuntu.com.png are Firefox 6 on Windows. Not fantastic, but I never really liked Windows font rendering anyway.
[00:49] <wgrant> Wheezy is still installing...
[00:53] <wgrant> Also, you can see the latest font changes on qastaging.launchpad.net.
[01:46] <lifeless> exarkun: do you want to hide the release as well ?
[01:46] <lifeless> exarkun: if not, could you expand on why ?
[01:46] <exarkun> Sure, hiding the release sounds good.
[01:47] <lifeless> I believe there is a checkbox for that
[01:47] <lifeless> I would expect it to hide the downloads associated with the release too
[01:47]  * exarkun looks around
[01:49] <exarkun> I see an "Active" checkbox for a milestone
[01:49] <lifeless> mmm no, thats related to whether bugs can be targeted
[01:50]  * lifeless has a poke
[01:50] <lifeless> nope
[01:51] <lifeless> exarkun: looks like I'm wrong and we don't support this
[01:51] <exarkun> okay
[01:51] <exarkun> thanks for checking
[01:51] <lifeless> exarkun: I suggest filing a bug noting that you can't hide releases/downloads - including a reason why you want to hide-but-not-delete them.
[01:52] <exarkun> I'm not sure I can give a reason for wanting to hide but not delete them.
[01:52] <lifeless> if its hidden we have to answer questions like 'who can see hidden $thing' and 'who can unhide them again'
[01:53] <exarkun> Yea, understandable.
[08:01] <mrevell> Hallo
[08:42] <wgrant> We're about to go down for a couple of minutes for a database upgrade.
[08:45] <wgrant> And we're back.
[10:07] <geser> wgrant: Hi, an user in #ubuntu-packaging is asking why his packages in his PPA are in pending (not published yet)
[10:07] <geser> I'm guessing it's about https://launchpad.net/~ambrop7/+archive/badvpn/+packages
[10:09] <ambro718> Hi. How long is it supposed for a PPA package to get over "pending publication"? My package finished building 7 hours ago and it's still pending.
[10:10] <wgrant> Hmm, that's not good. Let me check some logs.
[10:10] <ambro718> this is the package in question https://launchpad.net/~ambrop7/+archive/badvpn/+packages
[10:10] <wgrant> ambro718: You don't see any warnings at the top of https://launchpad.net/~ambrop7/+archive/badvpn?
[10:11] <ambro718> wgrant: warnings? no.
[10:11] <ambro718> I did once accidentally delete the PPA but then re-enabled it; could that have left it in a strange state?
[10:12] <bigjools> yes
[10:12] <wgrant> It should still work, but yes, it will bein a strange state.
[10:12] <wgrant> ambro718: https://launchpad.net/~ambrop7/+archive/badvpn/+edit, check that the Publish flag is enabled.
[10:12] <wgrant> ambro718: But don't rely on undeleting PPAs.
[10:12] <wgrant> That it works at all is a bug.
[10:13] <ambro718> wgrant: thanks' I checked that flag
[10:24] <ambro718> why is my PPA not building for ARM? https://launchpad.net/ubuntu/+ppas says that it's supported, and my control file has Architecture: any
[10:25] <bigjools> ambro718: because you need special permission to use ARM
[10:25] <ambro718> bigjools: ah, ok. Don't really need it yet, just asking.
[10:26] <ambro718> how is PPA building for ARM? does it use actual ARM hardware, or emulation?
[10:26] <bigjools> actual hardware
[10:27] <ambro718> and which hardware is fast enough for that?
[10:27] <lifeless> hahhahaahaha
[10:27] <bigjools> none of them? :)
[10:48] <geser> when I try to view the "inline" package details on a source package in LP, I get a "Failed to fetch package details". Is this known?
[10:49] <geser> with Firebug I see a request for +listing-archive-extra which gives a 404
[10:49] <geser> e.g https://launchpad.net/ubuntu/+source/ubuntu-dev-tools/+sourcepub/1931343/+listing-archive-extra
[10:50] <geser> when trying to expand 0.129 on https://launchpad.net/ubuntu/+source/ubuntu-dev-tools/+index
[10:50] <lifeless> geser: I believe there is a bug
[11:38] <wgrant> geser: It works for me. DO you have an archive_context_url set in LP.cache down near the bottom of the page?
[11:41] <bigjools> works for me too
[11:42] <bigjools> and woo, it closed bugs.  My code works!
[11:45] <wgrant> bigjools: Excellent.
[11:47] <nigelb> bigjools: \o/
[11:48] <Daviey> Is it viable to try to replicate a local instance of just the buildd part of soyuz to test some build failures? (They are working in sbuild/pbuilder)
[11:48] <bigjools> it's not easy but possible.
[11:48] <bigjools> https://dev.launchpad.net/Soyuz/HowToUseSoyuzLocally
[11:51] <Daviey> Hmm, thanks bigjools
[11:56] <StevenK> Daviey: What sort of failures?
[11:56] <StevenK> If it works under sbuild, it ought to build ...
[11:57] <bigjools> could be a Xen bug
[11:58] <wgrant> StevenK: Well, if it works under our 7 year old fork of sbuild...
[11:58] <StevenK> Yes, we should fix that.
[11:58] <StevenK> bigjools kept threatning me with it, but now I'm IMMUNE
[11:58] <geser> wgrant: http://paste.ubuntu.com/685895/ for the LP.cache line from that page
[11:59] <geser> wgrant: does it load for you when you aren't logged-in?
[11:59] <geser> I retried it after logging in and it works but not when I'm not logout again
[11:59] <wgrant> geser: Ahhh.
[12:00] <wgrant> Yes, that would be it. The cache has some odd behaviour for anonymous users. Could you file a bug, please?
[12:01] <Daviey> StevenK: bug 831073
[12:01] <Daviey> StevenK: it's like it's using a different pwd to what sbuild and pbuilder does.
[12:01] <Daviey> (note, this is multiple upstream tarball package - so my thought is that the buildd's are not handling this correctly)
[12:03] <wgrant> And the award for most useless named component orig tarballs goes to...
[12:03] <bigjools> StevenK: it's in Perl.  You know Perl.  Ergo ...
[12:03] <Daviey> heh
[12:04] <StevenK> That's like saying "There's a bug in the kernel. It's in C. You know C. Ergo ..."
[12:04] <bigjools> Perl - for masochists everywhere.
[12:04] <StevenK> Which is an utter fallacy
[12:04]  * bigjools wins this round of trolling
[12:04] <StevenK> Just because I know Perl doesn't mean I want to spend two weeks stabbing myself with sbuild's particular idiocy
[12:05] <Daviey> StevenK: Hmm, the kernel comment is infact valid.  Kernel 4.0 will be written in Go, yaknow.
[12:05] <bigjools> and on that happy note, I head to lunch
[12:05] <StevenK> Daviey: I was already holding daggers to hold sbuild back. Tell me does it hurt when I do this ...
[12:06] <Spads> StevenK: "Never admit to anyone that you know how to write sendmail rules or TROFF macros" -- Marshall "Kirk" McKusick
[12:06] <Daviey> :)
[12:06] <StevenK> Haha
[12:06] <StevenK> Spads: So we should add regexs and Perl to that list?
[12:07] <StevenK> They fall under the same line noise as sendmail rules.
[12:07] <Spads> StevenK: http://stackoverflow.com/questions/1732348/regex-match-open-tags-except-xhtml-self-contained-tags/1732454#1732454
[12:08] <StevenK> Bwahahahaa
[12:09] <Spads> Best Of The Web.
[12:12] <StevenK> Daviey: Did it only fail on amd64?
[12:12] <StevenK> Daviey: Perhaps it is wanting some file that is only built in an arch-indep target
[12:13] <geser> wgrant: bug 845544, do you need any other data?
[12:13] <wgrant> geser: That's fine, thanks.
[12:14] <wgrant> Daviey: I wonder if you need to build-depend on automake.
[12:14] <wgrant> Daviey: I can reproduce not quite the same error, but a similar one, by removing automake -- you depend on automake1.7, but automake was providing aclocal here.
[12:15] <wgrant> And you see the build log complaining about aclocal being missing.
[12:15] <wgrant> Daviey: One thing you could try is building in the same chroot that LP uses.
[12:15] <wgrant> Daviey: You can see the URL at https://launchpad.net/api/devel/ubuntu/oneiric/i386
[12:41] <Daviey> StevenK: no, failed for all
[12:42] <Daviey> wgrant: Yeah, will try that.
[12:42] <Daviey> thanks
[15:31] <jsjgruber> #ubuntu-sso
[23:49] <rodrigo> hi! I couldn't find I working way in the docs about how to add a tag to a bug report using a link, can someone help me with it? THANKS!!
[23:50] <rodrigo> something like +filebug?field.tag=doc
[23:52] <wgrant> rodrigo: field.tags=doc
[23:52] <rodrigo> wgrant, jojojo! I can't believe it...
[23:53] <rodrigo> wgrant, THANKS!
[23:54] <rodrigo> wgrant, and for the description field?
[23:55] <wgrant> rodrigo: It should be field.description, but it seems it's actually field.comment.
[23:56] <rodrigo> wgrant, THANKS A LOT! I coudn't believe that I was stuck because of a typo...
[23:57] <wgrant> Heh.
[23:57] <rodrigo> wgrant, thanks again!.
[23:58] <wgrant> np