[00:00] <sladen> DBO: a slightly different, but related case
[00:00] <DBO> =/
[00:00] <DBO> pretty...
[00:00] <DBO> i guess I'll just make Docky embed everything, mwahahahaha.... wait
[00:05] <sladen> https://stage.maemo.org/svn/maemo/projects/haf/trunk/gtk+/docs/dnd_internals.txt  grep for "we actually have to figure out which window
[00:07] <DBO> sladen, how did you find that?
[00:11] <sladen> DBO: I'm not sure what search phrase; possibly "window manager ignore xdnd" or "xdnd masking input masking"
[00:11] <sladen> DBO: I actually have a related case for a transparent keyboard so I'm quite interested in find the answer
[00:11] <DBO> sladen, your google-fu is strong
[00:11] <DBO> well if it makes any difference to you, I am doing this all in Mono
[00:11] <DBO> so we will see if I get it working =)
[00:16] <Roey> mako:  greetings
[00:16] <Roey> mako:  hmm, wrong maco :)
[00:17] <Roey> but hello nonetheless!
[00:22] <sladen> DBO: there's going to be interesting situation when the target window, docky window and icon window overlap, if two out of the three are trying to proxy to the "correct" bottom window
[00:23] <sladen> DBO: but as long as you (in Docky) manage not to proxy it to the clipped mouse icon window, it should work
[00:23] <DBO> yeah
[00:23] <DBO> i am really curious how I am supposed to figure out how to proxy to the right window
[00:23] <DBO> lets say there are 4 or 5 windows in a stack
[00:24] <DBO> which one do I figure out is on top?
[00:25] <sladen> testing the x,y against the rectangle of the topmost, then the second most, then the third most ... until you get a hit
[00:26] <DBO> sladen, i am not sure I follow you here
[00:27] <DBO> you must excuse me, I have been programming for far less time than I should have by now
[00:27] <DBO> I can get a ref to the root window
[00:27] <DBO> are all other windows one of its children?
[00:27] <DBO> or rather, how do I find the topmost window?
[00:31] <sladen> DBO: take a bunch of pieces of paper and spread them (overlaping) on the table.  Close you eyes and put your finger down;  which piece of paper did you hit?
[00:32] <KDesk> hi
[00:32] <DBO> sladen, i know how to do that logically
[00:32] <DBO> just not in code...
[00:32] <sladen> DBO: hopefully you can check that you hit this piece of paper as you're inside it's four edges/four corners
[00:33] <DBO> i can check and see if I am in bounds
[00:33] <sladen> DBO: yup, you can get a list of all of the TOPLEVEL windows that are a a child of the ROOT window
[00:34] <DBO> right but what if two of those toplevel windows are overlapping?
[00:34] <sladen> then, if you're doing thwat Gtk is doing for efficiency you can subscribe to changes and keep you're internal list up to date
[00:34] <sladen> or if you're lazy, you can just query for the full list each time a DnD comes in
[00:34] <KDesk> For which architecture are the deb packages in Ubuntu optimized for i386, 686 or what?
[00:35] <sladen> KDesk: they're optimised for recent chips, but will run on i486 IIRC
[00:35] <DBO> i will be lazy
[00:36] <sladen> DBO: if you put all the windows in order from top to bottom, and start at the top and work downwards, you'll check the topmost one first
[00:37] <DBO> yes i know that, but how does one put them in order
[00:37] <DBO> that is the question I have been trying to ask
[00:37] <KDesk> sladen: ah, I thought there where not optimized, or only for 386, is possible to know this info from a binary package?
[00:38] <ebroder> Is there anyone from backporters who could take a look at LP bug #323546?
[00:38] <sladen> for w in sorted(get_root_window()->children) { if w.bounds.text(event->mouse.position) proxy_dnd_to(w); }
[00:39] <DBO> sladen, sorted?
[00:39] <DBO> I am not trying to be difficult
[00:39] <sladen> DBO: psuedo-code
[00:40] <DBO> i know
[00:40] <DBO> but thats the thing I am saying
[00:40] <DBO> I see nothing in gdk's api that lets you determine which window is "topmost"
[00:41] <sladen> in Python (a fabulous language, btw), it would return a copy of the array, sorted by calling __cmp__() on each object
[00:41] <sladen> which window is topmost, depends on where you are looking on the screen
[00:41] <Roey> sorted(collectionOfObjects), sladen?
[00:42] <Roey> sladen:  doesn't sorted() call __cmp__ on every element in the list passed to it?
[00:43] <sladen> KDesk: don't think so, but you can check the build-log that built that binary package.   If you were running a supercomputercluster you'd probably be worried about optimisation;  however the default configuration options are probably about as good as you can get for Desktop yse
[00:44] <KDesk> sladen: ah, ok, thanks a lot! A last question, where can I find the build log, in launchpad?
[00:44] <sladen> Roey: so you might do  sorted(window_list, cmp=lambda a,b: x.height < y.height)
[00:45] <Roey> ahhhh ok
[00:45] <Roey> I had thought that sorted() calls __cmp__
[00:45] <Roey> but, you're right, I've used sorted() in the way you described to sort IP address stringts
[00:45] <Roey> strings
[00:45] <Roey> so yeah
[00:51] <sladen> KDesk: probably easier if you say which package you're after.  eg. https://launchpad.net/ubuntu/+source/xbill/2.1-7/+build/593853
[00:53] <KDesk> sladen: oh, thanks! I was only curious about the optimization. :)
[00:58] <cjwatson> ScottK: yes, yes - maybe I shouldn't have mentioned Debian. The point is that there's nothing wrong with having a package in a private revision control system and just uploading the tar/diff that results from it, as long as it's then your responsibility to maintain it.
[00:58] <ScottK> cjwatson: Certainly.
[01:40] <DBO> sladen, it works!
[01:40]  * DBO dances the dance of joy
[02:41] <jelmer> slangasek, hi
[03:47] <sladen> DBO: rock!
[03:47] <DBO> sladen, how can I thank you in the source, real name, sladen, email?
[04:26] <bettercoder> Hey guys! I think I am cool enough to be a mainstream dev! I made a program that prints hello world! pretty sweet, eh??!!!!?!??!?!?!??!?!?!?!??
[04:27] <bettercoder> also, is ubuntu vulnerable to the y2k bug/
[04:27] <bettercoder> hello???????????????!!!!!!!!!!!!!!!!!!????????????????????!!!!!!!!!!!!!!?????????????
[04:27] <bettercoder> WWWWWWWWWWWWHHHHHHHHHHHYYYYYYYYY WOOOON"TTT ANNNYYONNNE ANNNNSSSSSSWEEEEEEEEERRRR????????????
[04:28] <patxbot> stfuckup
[04:28] <patxbot> jeese
[04:28] <bettercoder> whats does stfuckup mean?
[04:28] <patxbot> it means be quiet
[04:29] <bettercoder> ok, thats a good idea
[04:29] <patxbot> stop spamming
[04:29] <bettercoder> So
[04:29] <bettercoder> hw do i b a dev?
[04:29] <patxbot> ty
[04:29] <bettercoder> huh??huh??
[04:29] <patxbot> you develop brains *first* cos i do not have any of them...
[05:54] <jsmidt> Is there anyone here who understands PPAs well?
[05:54] <jsmidt> I am trying to build Python 3.0 for hardy.
[05:54] <jsmidt> https://launchpad.net/~python-ubuntu/+archive/ppa
[05:54] <jsmidt> It build depends on python-sphinx
[05:55] <jsmidt> I have successfully build this in the PPA for Hardy
[05:55] <jsmidt> Yet Python 3.0 still complians the build dependency isn't met
[05:55] <jsmidt> How do you get a package in the PPA to "see" the other packages?
[05:56] <jsmidt> https://launchpad.net/%7Epython-ubuntu/+archive/ppa/+build/854896/+files/buildlog_ubuntu-hardy-i386.python3.0_3.0-0~ppa7~hardy_FAILEDTOBUILD.txt.gz
[05:56] <jsmidt> Thanks in advance.
[06:12] <ScottK> jsmidt: PPAs have nothing to do with Ubuntu.  They use Ubuntu, but are done by Launchpad.  Try PPA questions in #launchpad.
[06:12] <jsmidt> ScottK, thanks
[06:13] <jsmidt> (Though I'm sure you are technically right, very few people would say PPAs have nothing to do with Ubuntu.  Much in Ubuntu is done through using PPAs)
[06:23] <ScottK> People use PPAs for testing, but "Ubuntu" the distro does not come from PPAs.
[06:23] <jsmidt> ScottK, yes I know, I know.
[06:25] <jsmidt> ScottK, I wasn't trying to argue.
[06:25] <ScottK> OK.  I think it's an important point.  I am tired of dealing with bugs filed against Ubuntu packages that are really from some random PPA.
[06:37] <jsmidt> ScottK, I'm sure that is very frustrating.  And so now that I think about it, my issue is probably annoying for the same reason.  :)
[11:33] <slangasek> pitti: please consider those Hotkey pages locked by me currently, for spec drafting :)
[11:34] <slangasek> pitti: but thanks for the information - bryce and I were struggling with exactly this question on the plane; I'll review what you've added to see if it makes sense out of what I see with thinkpad_acpi
[11:35] <slangasek> jelmer: hi, thanks for the fixed branch - what's the right way for me to replace the current ~ubuntu-core-dev branch with yours?
[11:36] <ogra> slangasek, hey, arrived safely ?
[11:36] <slangasek> ogra: yep, all 16 of us
[11:37] <ogra> great :)
[11:41] <slangasek> pitti: btw, hal-info isn't currently in bzr?  Should I push an ~ubuntu-core-dev branch based off of the jaunty package-import?  (I have a patch for it to fix a ThinkPad hotkey)
[14:13] <ion_> http://golf.naurunappula.com/nn/0/151/007/336376.jpg
[14:13] <ion_> Sorry, wrong channel. :-(
[14:30] <tjaalton> lool: libdrm is maintained in git, so probably the "Rename Vcs-*" bit is wrong? (thanks for adding lpia, I didn't know -1.2 was out)
[14:52] <hyperair> could someone look at bug #268141 please? debdiff's attached for both intrepid and jaunty.
[14:55] <directhex> erm... is requestsync being insane again?
[14:57] <lool> tjaalton: Oh sorry, I checked the git tree and didn't see the ubuntu branch
[14:57] <lool> tjaalton: Which branch is it?
[14:57] <lool> tjaalton: Uh I'm blind
[14:57] <Rocket2DMn> TheMuso, i'm looking at bug 308539 which states that the problem is that libft-perl is uninstallable - https://launchpad.net/ubuntu/intrepid/+source/defoma/0.11.10-0.2ubuntu1
[14:57] <lool> tjaalton: I checked it out extra yesterday evening to check
[14:58] <lool> tjaalton: But I missed the branch
[14:58] <lool> tjaalton: Sorry about that
[14:58] <lool> tjaalton: Yeah, not only is the rename wrong but I should have committed there
[14:58] <Rocket2DMn> TheMuso, can you tell me how to proceed with the triage on that bug?  Can we somehow link it to that page
[15:02] <Rocket2DMn> TheMuso, please ignore my previous comments, sorry
[15:04] <tjaalton> lool: no problem :)
[15:05] <TheMuso> Rocket2DMn: uh... ok. :)
[15:05] <Rocket2DMn> heh, sorry man
[15:06] <TheMuso> np
[15:26] <lool> tjaalton: Pushed 2.4.4-0ubuntu4 and a revert of the XS-Original-Vcs-* change to git
[17:19] <TheMuso> c
[17:27] <pitti> fta: that was a bug in gvfs, hopefully worked around in the latest version
[17:28] <fta> pitti, eh.. what for?
[17:28] <pitti> slangasek: locked pages> oh, sorry; but I only really added information, so I guess it's ok
[17:29] <pitti> slangasek: hal-info isn't in bzr; I commit stuff directly into git upstream; but if it helps you, feel free to bzrize it
[17:29] <pitti> fta: you pinged me earlier about getting file:// urls when reporting apport crashes
[17:29] <fta> pitti, oh, ok. thanks
[17:30] <fta> pitti, i ended up filling bug 323409
[17:31] <pitti> fta: FYI, bug 314263 is the breakage you mention
[17:32] <fta> pitti, yes, sure. i encountered 314263 while fighting against 323409.
[17:32] <fta> now waiting for Keybuk to have a look at 323409
[17:56] <slangasek> pitti: ah, well, if it doesn't help you, I probably won't bother bzring hal-info since LP needs to be taught about the project first
[17:56] <slangasek> pitti: anyway, should I upload my ThinkVantage change, or email it to you?
[18:06] <allquixotic> Hi, I found a bug in Jaunty (a regression from 8.10) which causes hald to crash during input hotplugging when I have a Logitech ChillStream video game controller plugged in to USB. The controller worked, and hald didn't crash, under 8.10. What sort of priority should I file a bug with? What details should I include?
[19:42] <calc> hello
[21:42] <bluefoxicy> okay this has been bugging me forever
[21:42] <bluefoxicy> can somebody please extend fakeroot to debian package INSTALLATION?
[21:42] <ion_> Why?
[21:42] <bluefoxicy> I should not be able to make a .deb that gets run, asks a question, gets a "no" and aborts installation
[21:42] <bluefoxicy> and leaves a rootkit behind on my system
[21:43] <bluefoxicy> current method of package installation:  glorified tarball.  Unpack, run pre-install.sh, does whatever it wants to your system AS ROOT
[21:43] <ion_> It’s okay if it leaves behind a keylogger on your user account?
[21:43] <bluefoxicy> installing a non-root service that runs as a restricted user and can't affect your system (say...wesnoth's dedicated server?)?  It can still gain root access
[21:43] <ion_> It’s not a problem of the package manager if a user installs a malicious package.
[21:44] <bluefoxicy> ion_:  I have had to re-install before (later, modify the deb) because of a bug in the java deb
[21:44] <bluefoxicy> if you don't check "I Agree" on the license question, it makes a permanent change to your system and aborts installation
[21:44] <bluefoxicy> pre-rm and post-rm don't undo it
[21:44] <elmo> bluefoxicy: it's not a permanent change
[21:44] <elmo> it's a debconf question
[21:45] <bluefoxicy> elmo:  yes, and if you say no, and then attempt installation again, it doesn't ask;it just says, "oh, no, you didn't want to agree to the license, so screw you"
[21:45] <elmo> bluefoxicy: ok, but that doesn't make it permanent
[21:45] <elmo> the debconf DB is entirely under your control
[21:45] <bluefoxicy> elmo:  nothing's technically "permanent."
[21:46] <ion_> 0) That can be fixed in the packaging, 1) how is fakeroot relevant here?
[21:46] <bluefoxicy> it could replace /bin/sh with malware, and I could technically replace the malware with a proper /bin/sh
[21:46] <bluefoxicy> ion_:  0) It's a package bug; 1) I don't believe the system should be changed, at all, until the package is finished installing.
[21:47] <bluefoxicy> in Gentoo, when a package is built, emerge has a sandbox environment that forbids real writes outside the build and image directory
[21:47] <bluefoxicy> post-installation file access pulls the files into the image directory, and i.e. updates to samba.conf or /etc/passwd get written to copies in the image directory
[21:47] <soren> I don't think fakeroot is the right tool for something like htat.
[21:47] <bluefoxicy> the final step is to "merge" that to /
[21:48] <bluefoxicy> at any point before then, your system has been completely unchanged
[21:48] <bluefoxicy> soren:  yeah that's whyI said extend... fakeroot doesn't do anything nearly useful for that
[21:48] <bluefoxicy> at one point I almost wrote a completely different package manager because i don't like things playing around with full root access running arbitrary code just to install
[21:49] <soren> I don't think fakeroot is the right application to extend to achieve something like that.
[21:49] <bluefoxicy> I don't have to audit the final installed files, config changes, etc; I have to audit the entire logic of the installation system
[21:49] <bluefoxicy> soren:  maybe so.
[21:49] <elmo> how do you audit the final installed files?
[21:49] <elmo> JOOI
[21:50] <bluefoxicy> elmo:  make all changes to an isolated image that gets  copied over / to finalize installation; then stop it at that point, and examine that tree?
[21:50] <Notch-1> hi, i'm trying to boot intrepid from a loop device, using root=/dev/loop0 (or root=700) boot parameter, and a little initrd script that mount the device containing the image, and then "losetup /dev/loop0 /path-to-the-image"... but it still can't find init... can anybody help me?
[21:50] <bluefoxicy> SUID binaries, strange changes to /etc/passwd, debconf settings, strange binaries that shouldn't exist, funny scripts generated by obfuscated logic in the pre-install scripts of the package...
[21:50] <bluefoxicy> all lain out clear
[21:52] <bluefoxicy> anyway
[21:52] <bluefoxicy> endrant
[21:59] <tjaalton> lool: excellent, thanks
[22:09] <Notch-1> please guys, anybody familiar with booting loop images?
[22:20] <ScottK> Notch-1: User support in #ubuntu
[22:21] <Notch-1> ScottK: thank you