[02:59] <Andrew54thousand> I am not sure if this is an ubuntu bug, or somewhere else, but does anyone know if there have been changes in support of ATSC tuner cards from 10.10 to 12.04...having some issues with a KWorld ATSC 115 and I am trying to pinpoint where the issue is...not sure if its an ubuntu issue, kernel issue or what
[03:00] <Andrew54thousand> the drivers load fine and the signal gets picked up, but it isn't locking on channels.  I have tried myth, tvheadend and w_scan and none are able to locate a service
[03:06] <Andrew54thousand> definitely not a hardware issue, i plug the old hard drive in and it works flawlessly...anyone know a log or somewhere to even look at why this is happening?
[08:21] <hjd> Hm, I have some crashlogs in /var/crash, so I thought I should report them using ubuntu-bug. I went through the steps, but it didn't open Launchpad nor required me to log in. I see that it created some .upload and .uploaded files. Did it report it anonymously or did something go wrong?
[13:07] <elgaton> Hello everyone - does anyone know how is it possible to get the "null character" used in the crypt() implementations (I'll need to use it for fixing a weird bug)?
[13:11] <penguin42> elgaton: I doubt it's possible with the crypt() function itself since it doesn't take a length, so the only length it would get is from the input
[13:11] <penguin42> (and I'd hope it doesn't just read the first 8 bytes....)
[13:12] <elgaton> penguin42: OK, thanks
[13:12] <penguin42> whats' the bug?
[13:13] <elgaton> Bug #943507
[13:13] <ubot2> Launchpad bug 943507 in pam-mysql "libpam-mysql lets you log in with any password when crypt=1 is set and the password field contains an empty string in the user record." [Undecided,Confirmed] https://launchpad.net/bugs/943507
[13:14] <elgaton> (It has been made public a while ago - stumbled on the report and decided to fix it)
[13:16] <penguin42> elgaton: How about replacing that with   strcmp(row[0], row[0]?crypt(passwd, row[0]:""))
[13:16] <penguin42> oh shuffle that ) before the :
[13:17] <elgaton> penguin42: Won't work: if the salt (row[0]) is empty then strcmp compares the empty salt with itself.
[13:17] <elgaton> Oh, sorry - did not read the second part
[13:18] <elgaton> OK, won't work either.
[13:18] <penguin42> hmm yes you're right
[13:19] <penguin42> strcmp(row[0], row[0]?crypt(passwd, row[0]):passwd))
[13:19] <penguin42> i.e. if the salt is empty compare the empty salt with the entered password
[13:19] <penguin42> (which would probably better to be an explicit if then else
[13:20] <elgaton> Would require a strcmp() in any case on row[0] since otherwise the memory address of row[0] would be checked for equality.
[13:21] <penguin42> oh sorry, I'll wake up in a winute
[13:21] <penguin42> strcmp(row[0], row[0]809?crypt(passwd, row[0]):passwd))
[13:22] <penguin42> gah, synergy
[13:22] <penguin42> strcmp(row[0], row[0][0]?crypt(passwd,row[0]):passwd)
[13:23] <penguin42> is what I mean, so you're either comparing row[0] against the output of crypt (if row[0][0] is non nil) or you're comparing row[0] against passwd directly when row[0] is actually empty
[13:27] <elgaton> Would still need an if/then/else as the crypt() output is directly used in the strcmp() function (it is not stored in row[0]).
[13:28] <penguin42> I don't quite see why; in that line I'm not using the output of crypt() when row[0] is empty
[13:30] <elgaton> Sorry again - the code is a bit criptic. Anyway, the bug would still be present if row[0] is one byte long - maybe I could write this:
[13:30] <elgaton> vresult = (strlen(row[0]) < 2) ? strcmp(row[0], passwd) : strcmp(row[0], crypt(passwd, row[0]));
[13:31] <penguin42> elgaton: Actually, I think you're right - it might be better to always do the crypt, and then check to see if crypt returns an empty string, if it returns empty then check row[0] is empty
[13:31] <penguin42> sorry, check both row[0] and passwd are empty
[13:35] <elgaton> Would require a malloc() though
[13:35] <penguin42> why?
[13:37] <elgaton> The code I wrote before checks if the salted, hashed password does not meet the POSIX requirements (i.e. is less than two characters long). If it does not meet them, it performs the comparison on the "raw" password, otherwise it calls crypt() as normal.
[13:37] <penguin42> { char* cryptres; cryptres=crypt(passwd, row[0]); if (cryptres[0]) { vresult=strcmp(row[0], cryptres); } else { vresult=strcmp(row[0], passwd); }
[13:38] <elgaton> Ah, OK, I thought the "cryptres" pointer needed initialization at first
[13:40] <penguin42> although I'd worry I don't understand crypt() semantics well enough - in mine if you can find something that got crypt() to return empty that then got directly compared against the key it would be bad
[13:43] <elgaton> That's what I thought too. Unfortunately, the POSIX specs do not define error behaviour. What they mandate is that the salt 1) should be at least two characters long and 2) should not contain an "implementation-dependant" null character. In case the salt does not respect these prerequisites, the return value is undetermined.
[13:44] <penguin42> ah right
[13:47] <elgaton> As I planned to send the patch upstream too, I was looking for a way to determine that character. Seems there is no way other than brute-force testing.
[13:48] <elgaton> (For reference, here are the specs: <http://pubs.opengroup.org/onlinepubs/9699919799/functions/crypt.html>)
[13:51] <penguin42> elgaton: Hang on, when you pass an empty salt what is crypt returning?
[13:52] <penguin42> is it returning NULL?
[13:52] <elgaton> elgaton: An empty string (on the glibc implementation).
[13:53] <penguin42> interesting that doesn't match that spec - it says it should return a null pointer on error
[13:55] <elgaton> If you look at the "Errors" section you'll see that crypt() can only fail if it is not implemented on that particular implementation.
[14:00] <penguin42> so a bad salt isn't classed as an error?
[14:01] <elgaton> No, it isn't (errno stays at 0).
[14:04] <penguin42> that's a shame, would have been too easy....
[14:08] <elgaton> Of course. So the choice is: 1) targetting glibc specifically, or 2) modifying the configure scripts to check what is returned when a null salt is passed.
[14:11] <penguin42> elgaton: Or as you say, belt and braces, check the salt is sane, and check the output of crypt
[14:12] <elgaton> penguin42: OK, I'll stick to that, seems the most universal approach.
[14:12] <elgaton> Thanks again
[14:12] <penguin42> yeh, nothing wrong with paranoid
[14:13] <elgaton> :)
[19:32] <dlentz> this person has about 20 bugs requesting packages to be lzma compressed? should they be combined?
[19:33] <dlentz> hold on a sec, lost the link..
[19:34] <dlentz> https://bugs.launchpad.net/~jerome-bouat/+bugs?field.searchtext=compression&search=Search&field.status%3Alist=NEW&field.status%3Alist=INCOMPLETE_WITH_RESPONSE&field.status%3Alist=INCOMPLETE_WITHOUT_RESPONSE&field.status%3Alist=CONFIRMED&field.status%3Alist=TRIAGED&field.status%3Alist=INPROGRESS&field.status%3Alist=FIXCOMMITTED&field.assignee=&field.bug_reporter=&field.omit_dupes=on&field.has_patch=&field.has_no_package=&orderby=-
[19:34] <dlentz> id&start=0
[19:35] <dlentz> err, https://bugs.launchpad.net/~jerome-bouat/+bugs?field.searchtext=compression&search=Search&field.status%3Alist=NEW&field.status%3Alist=INCOMPLETE_WITH_RESPONSE&field.status%3Alist=INCOMPLETE_WITHOUT_RESPONSE&field.status%3Alist=CONFIRMED&field.status%3Alist=TRIAGED&field.status%3Alist=INPROGRESS&field.status%3Alist=FIXCOMMITTED&field.assignee=&field.bug_reporter=&field.omit_dupes=on&field.has_patch=&field.has_no_package=&orde
[19:35] <dlentz> rby=&
[19:35]  * dlentz sighs
[19:35] <wylde> dlentz: bug #something , the bot should spit out the link :)
[19:35] <dlentz> but i'm not showing one specific bug..
[19:36] <wylde> dlentz: I know you're talking about mutiples, yeah
[19:37] <dlentz> ok, this will give you the idea: https://bugs.launchpad.net/~jerome-bouat/+bugs
[19:38] <wylde> dlentz: I didn't look too deeply but they certainly give the impression that they should be duplicates.
[19:40] <dlentz> wylde, thanks. i'll wait a few hours to see if there are other opinions, and if not, i'll dupe htem
[19:41] <wylde> dlentz: yeah, and they all look like wishlist items as well. This guy really wants to push lzma eh?
[22:02] <steemed> Anyone want to fix a bug for me? I'm pretty sure it's a simple fix because it just involves dependencies
[22:02] <steemed> https://bugs.launchpad.net/ubuntu/+source/ubuntuone-control-panel/+bug/995262
[22:02] <ubot2> Launchpad bug 995262 in ubuntuone-control-panel "ubuntuone-control-panel-common requires ubuntuone-installer" [Undecided,New]
[22:04] <dlentz> steemed, it is a simple fix (quick edit of the control file), but i think the package maintainer needs to do it
[22:04] <steemed> Ah ok
[22:04] <dlentz> at least it's easy to work around
[22:04] <steemed> Mhm, thought so.
[22:05] <steemed> OK thanks, just wanted to point out my bug because it seemed fairly simple and quick to fix, but it just requires whoever manages the packages to fix
[22:05] <steemed> Now I know
[22:09] <dlentz> yw
[22:34] <penguin42> steemed: I marked it Triaged since you describe exactly what needs to be fixed, and set it Medium importance since it's fairly nasty in one sense but easy to work around
[22:34] <steemed> penguin42: thanks
[22:35] <steemed> penguin42: medium? I'm flattered :D
[22:35] <penguin42> steemed: Well, I was tossing a coin over whether it was Medium or low; I mean it's a core package and not being able to install it on some configs is nasty, but it's easy to work around
[22:36] <steemed> Yeah true
[22:38] <steemed> I do have to say that ubuntuone-installer package is genius to include in the default install so that it was able to squeeze the whole ISO on a CD
[22:38] <EvilResistance> penguin42:  which bug?
[22:38] <EvilResistance> you might want a secondary opinion on Medium vs. Low :P
[22:38] <steemed> bug 995262
[22:38] <ubot2> Launchpad bug 995262 in ubuntuone-control-panel "ubuntuone-control-panel-common requires ubuntuone-installer" [Medium,Triaged] https://launchpad.net/bugs/995262
[22:39] <EvilResistance> my question, then, and penguin42 can weigh in on this, how wide-spread is this issue actively?
[22:39] <EvilResistance> i.e. what percentage of Xubuntu or Kubuntu users install Ubuntu One?
[22:39] <penguin42> EvilResistance: I don't know, I don't use it, I just triaged it from the description
[22:40] <steemed> Not very, IMO. I found it out from a friend
[22:40] <EvilResistance> penguin42:  i'll see if i can't hash together a debdiff on it
[22:40]  * EvilResistance can probably do so, given he's done this for other packages
[22:40] <EvilResistance> having said this, if the percentage of users are *not* running into this issue, it may be prudent to set it to 'low'
[22:40] <EvilResistance> but the solution either needs the patch or the debdiff (in this case, the debdiff)
[22:41] <EvilResistance> steemed:  are you certain, though, that Ubuntu One doesn't need the GTK3 pacakges?
[22:41] <EvilResistance> last I checked, Ubuntu One uses the GTK interface
[22:41] <steemed> Ubuntu One as of 12.04 uses qt
[22:41] <steemed> So that it is portable across platforms
[22:42] <EvilResistance> mmkay, i'll hash together the debdiff
[22:42] <steemed> The reason the installer requires it is to reduce space taken on the ISO and use the already available libs
[22:42] <EvilResistance> penguin42:  i'll ask around on the xubuntu and kubuntu questions, but since i lurk in kubuntu, and havent seen many people wanting to install it, i'd say put it as "Low"
[22:42] <EvilResistance> (i'm on bug squad, and that's my personal opinion)
[22:42] <EvilResistance> but it coudl go either way
[22:43] <steemed> EvilResistance: ot
[22:44] <steemed> EvilResistance: It's posssible they are installing it and not complaining about how much is required
[22:44] <EvilResistance> true
[22:44] <steemed> Because well, people don't look
[22:44] <penguin42> EvilResistance: Nod, I'm on BC, but I tend to err towards Medium unless I'm sure; I reckon it's better for the guy who knows the package well to spot it and lower it rather than it being marked low and it not to get spotted
[22:45] <EvilResistance> i've *worked with* the package, but i'm not fully into the ubuntu-one package chain
[22:45] <steemed> Yeah, someone in the Ubuntu One Control Panel team needs to look at it
[22:45] <EvilResistance> although given i can probably propose a debdiff to fix it
[22:45] <penguin42> EvilResistance: ah ok, then you know more about the package than I do
[22:45] <EvilResistance> (since its usually debian/control thiat defines the additional requirements)
[22:46] <EvilResistance> i've already downloaded the source package
[22:46] <steemed> It was pointed out by someone who was looking through the package requirements on muon on Kubuntu 12.04
[22:47] <steemed> And posted on the forums and making fun of Ubuntu for requiring so many dependencies lol
[22:47] <steemed> But I looked at why it was doing it
[22:47] <steemed> And all the GTK dependencies were coming from the installer
[22:49] <steemed> It also prevents the installer from being removed without removing the control panel for ubuntu one
[22:51] <steemed> I would say low since not a lot of users complain, but then the medium is also something because of the hugeness of all the needed dependencies
[22:53] <penguin42> I think of the importances on a per-package basis
[23:05]  * steemed thinks about becoming a bug squad member
[23:06] <penguin42> please do, there are lots of bugs to be squaded
[23:06] <steemed> Squated? o.o
[23:06] <steemed> I don't want to poop bugs D:
[23:06] <penguin42> hehe
[23:07] <steemed> I got nothing much to do anyways. I'm at home, recovering from depression.
[23:08]  * penguin42 sympathises - although be a little careful, some bug reporting can be depressing
[23:08] <steemed> Why
[23:09] <penguin42> steemed: Well you can get some that are badly reported, or some where the guys who report it are really upset
[23:09] <steemed> lul
[23:09] <steemed> capslock?
[23:10] <wylde> or some that say ZOMG! IT JUST DUNT WERK!
[23:10] <steemed> Haha
[23:10] <steemed> MAI SUSPEND DUN WORK
[23:11] <wylde> there's TONS of suspend bugs :/
[23:11] <steemed> Yah, mine doesn't work while plugged in
[23:11] <steemed> It'll turn off the screen but that's it
[23:11] <steemed> I'm thinking it's a config that's screwed up somewhere
[23:17] <wylde> speaking of bugs, I was talking to the reporter of bug #995185 earlier. There were packages provided for keytouch up until Maverick, was the package dropped? Or just not packaged since? Or how would I find out?
[23:17] <ubot2> Launchpad bug 995185 in gnome-settings-daemon "Mutlimedia keys Disabled" [Undecided,New] https://launchpad.net/bugs/995185
[23:18] <penguin42> wylde: packages.ubuntu.com will show you the package state in all versions
[23:18] <penguin42> wylde: If it was removed I'd expect to see a bug for it saying to remove it
[23:18] <wylde> penguin42: yep, ok I'll re-check then
[23:18] <wylde> thank you
[23:18] <steemed> https://bugs.launchpad.net/ubuntu/+source/xubuntu-meta/+bug/995293
[23:18] <ubot2> Launchpad bug 995293 in xubuntu-meta "Gtk3 applications look awful under Xfce" [Undecided,New]
[23:18] <steemed> penguin42: ^ I'm already depressed with this one
[23:19]  * steemed cries in corner
[23:25] <wylde> I don't see a bug related to removing the package. I am curious because keytouch-editor has builds right up to precise, but keytouch and keytouch-data stopped at natty.
[23:38] <hggdh> wylde: rmadison can help (it will show, by default, all Ubuntu versions with this package); also, you can look at the debian/control file on the source package to find out what binary packages are produced by the build
[23:39] <wylde> ok, thanks hggdh
[23:40] <hggdh> wylde: to grab the source of a Ubuntu package: bzr branch lp:ubuntu/<version>/<package> where <version is the Ubuntu version you want, and <package> is the source package name
[23:42] <wylde> ahh ok :)
[23:44] <geofft> wylde: Looks like it was removed from Debian, see http://bugs.debian.org/632110
[23:45] <wylde> geofft: hmmm, ok. It struck me as odd that keytouch-editor is still in ubuntu, and has builds that are current.
[23:46] <wylde> any oversight perhaps?
[23:46] <geofft> well, keytouch-editor is still in Debian too, fwiw, although maybe that's a mistake
[23:46] <wylde> an*
[23:46] <geofft> You can see Debian removals (and other stuff) at http://packages.qa.debian.org/<package-name>
[23:47] <wylde> geofft: thanks I'll bookmark hat for future reference.
[23:47] <wylde> that*
[23:57] <trinikrono> i saw a email with texinfo bitesize bugs and i would i like to help, which one of the files should i download to see if is the bug indicated
[23:58] <trinikrono> an example being bug 984562
[23:58] <ubot2> Launchpad bug 984562 in texinfo "package install-info 4.13a.dfsg.1-8ubuntu2 failed to install/upgrade: ErrorMessage: subprocess installed post-installation script returned error exit status 2" [Undecided,Invalid] https://launchpad.net/bugs/984562
[23:58] <penguin42> well, that's already marked as invalid