[00:00] mok0: why cant you test it? test if it builds? === RAOF_298 is now known as RAOF_831 [00:00] RAOF_298: use ppa for it, instead? === RAOF_831 is now known as RAOF_61 [00:00] hello Hobbsee [00:00] handschuh: It builds fine, but not being a java programmer, I can't verify that the library works as advertised === RAOF_61 is now known as RAOF [00:01] hey ajmitch! [00:01] mok0: a ok, thats fine [00:01] Hobbsee: That seems somewhat against the spirit of the PPA, though [00:01] ? [00:01] mok0: thanks for reviewing [00:01] RAOF: well.... [00:02] RAOF: it's for testing packages, and distributing them if they're good? [00:02] As opposed to just testing whether they build or not? :) [00:02] I could, I suppose. Jaunty's now PPAable, isn't it. [00:03] well, if they don't build, then they don't get distributed. if they do build, people get them via the ubuntu repos anyway, so the ppa versions will be overwritten [00:03] yes, i believe so [00:08] in a man page, does = need to be \= ? === chuck_ is now known as zul [00:18] binarymutant: you mean escape the minus? No [00:18] s/minus/equal [00:19] thanks mok0 [00:29] erm [00:29] http://revu.ubuntuwire.com/details.py?upid=2998 <- dpatch is on the build depends, how could it fail to build ? [00:32] Why are you including both /usr/share/dpatch/dpatch.make and /usr/share/cdbs/1/rules/dpatch.make? [00:34] good question, copy&paste [00:35] well, at least the package got a review :P [00:37] I will fix the dpatch double inclusion once I get more things to fix [01:00] Can mk-sbuild-lv be used to create Debian schroots? [01:02] That's how I built my sid chroot, yes. [01:03] I may have had to specify a different mirror or something; it's been a while, and the disc that has those schroots is sitting in a box with (at least) a blown power supply. [01:05] * RAOF wonders whether it'd be considered polite to test the grub2 installs and works before uploading :) [01:06] RAOF: uh, yes. [01:06] It's a bootloader. You should try it yourself and in a virtual machine [01:08] RAOF: just put a link to your package on the forums [01:08] Juuust kiddin. [01:09] ajmitch: Good plan! [01:09] Cmon LP. Publish those binaries. [01:09] But rename it to openoffice.org3 [01:09] you'll soon hear of any probelms [01:10] Be sure and set the maintainer to ajmitch before you put it in your PPA though. [01:10] * StevenK grins [01:10] * ajmitch sets up procmail to bounce any mail about grub2 to ScottK [01:13] RAOF: thanks [01:15] ajmitch: evil [01:23] hello people [01:25] mHey emgent [01:26] heya master [01:26] :) [01:26] emgent: I still need one more +1 before you can call me that ;) [01:27] hahah [02:39] hey all , what is the command for knowing what patch sys the packages uses [02:39] = [02:40] I recommend checking what the build depends are [02:41] a quick look at debian/rules's includes is often indicative, too [02:41] yep [02:41] * ajmitch dies a little inside seeing questions about checkinstall [02:42] There's also the "what-patch" script shipped in... devtools? [02:42] ubuntu-dev-tools [02:42] persiabot is faster than dpkg -S :) [02:43] ajmitch: you mean... that's not how you make a deb? [02:43] ajmitch: do you recall someone coming into #ubuntu-devel a year ago trying to build a deb by using dh_make && dpkg-buildpackage in one fell swoop? [02:43] jdong: no, you use ar [02:44] * jdong sees Firefox 3.0.4 crack in jaunty and pre-backports [02:45] lol [02:50] anyone know of any good tutorials on dpatch? besides the wiki and morph's blog? [02:52] binarymutant, Generally it's just datch-edit-patch patchname, and including the right dpatch makefile fragment. [02:52] binarymutant: What aspect of dpatch would you like a tutorial for? I find it pretty self-explanitory. [02:53] zsh (at least) will even tab-autocomplete the patchname for you, if you're editing an existing patch. [02:58] ok... I have been goolgling about doing the import for quilt in debian/rules ... can anyones direct me as where to read about this? [03:00] well do I create 00list for dpatch? [03:00] binarymutant: 'emacs debian/patches/00list' [03:01] RAOF, so ya, because that would be a new file for me [03:02] RAOF: you mean vi? [03:02] jdong: No, because I'm a worshiper of the One True Operating System. [03:03] emacs? [03:03] Right. [03:03] * coppro uses neither! [03:04] found it :D [03:12] I am using a patch system and I am sure I did not touch the code, yet lintian complains about "patch-system-but-direct-changes-in-diff .pc/.version" any clues? [03:15] I suspect that's a quilt-related directory, but I'm not sure. [03:15] "Fetched 15.6MB in 1h59min35s (2169B/s)". Yay. [03:15] how could I check this? [03:16] LOL [03:16] RAOF: welcome to the first world [03:16] Well, you could see what's in that directory. But man quilt contains a reference to .pc. [03:16] I remember those days. [03:17] jdong: and people in the US dare to complain about a 250GB limit [03:18] ajmitch: well that's more because you are paying $80/mo for "unlimited" internet [03:18] jdong: Really USD$80/mo? [03:19] RAOF: some of the Comcast plans are around that price, yes. [03:19] I hope you get crazygood bandwidth, then. [03:19] RAOF: I know my neighbors on Comcast pay about $120 for high-speed internet and cable TV combined [03:19] yeah he's got some crazy 10 or 12mbit plan [03:19] which is kinda pointless when they "cap" you to 250GiB [03:20] Well, not really. Burst speed is important. [03:20] * RAOF looks at his modem's sync page, which suggests he's getting 24Mbit down. [03:21] Or would be, were caps not in effect. [03:21] RAOF: living right next door to the exchange? [03:22] No idea where my exchange is; were I next door I'd be expecting a better sync, though. [03:24] better than 24Mbps? [03:24] that is the upper limit of adsl 2 [03:24] s/2/2+/ [03:24] * RAOF was under the impression 28 was the upper limit, but that may be faulty memory. [03:26] * ajmitch has always seen 24 [03:27] of course the best I've seen at home is about 11 :) [03:34] i can only get 3mbps here i would love any of that bandwidth even just 10mbps :) [03:35] transferring OOo back and forth takes a long time on 3m/512k connection [03:36] Ouch! [03:37] Ouch indeed. [03:38] if i switched to cable modem i think i can get 6mbps maybe but then its shared bandwidth, so not sure what the effective rate would be [03:39] My experience with shared bandwidth is that it's best practice not to try to use it when your neighbours want to use it. [03:39] * ajmitch is on about 7Mbps/128Kbps [03:39] persia: heh [03:40] so uploads = slow [03:40] ajmitch: 128kbps would be very painful for OOo :) [03:40] if I were uploading OOo, I'd be tempted to copy it to my laptop & upload from work [03:48] I called my ISP earlier in the week and while I was on the phone, they said, "Oh. you have the 12/2 Mbit plan. We don't offer that any more. We can switch you to 25/5 for US$ 0.04 more per month." [03:48] I said, "OK". [03:48] ! [03:48] ScottK, FTW! [03:49] So sometimes it pays to call in. [03:59] 1993/win 20 [04:00] But what have you won lately? [04:07] RAOF: I'm guessing Do caches .desktop entries. Can I make it flush it? [04:12] StevenK: It only indexes .desktop entries on each start. Quitting and starting it again will fix that. [04:57] ScottK: In Australia it doesn't pay... with most major ISPs, they'll try to transfer you to the new set of plans which are vastly inferior to the one you've been on for 5 years... [04:59] And what kind of connection are you on to get 5Mbps upstream!? [04:59] I get some obscene upstream rate on this connection [05:00] * wgrant has 10000/256 [05:02] wgrant: It's fiber all the way to the house. [05:02] ScottK: Oh. I hate you. [05:02] That's actually the slowest speed they offer. [05:02] Well there's actual competition here now and it appears to be working. [05:03] wgrant: 10000/256 seems highly asymetrical... do you mean 1000/256 ? [05:03] jmarsden: No, 10000/256 [05:04] Wow, strange. I'm not sure I've seen a connection with a 40:1 ratio like that. [05:04] One would think not. [05:06] The 1GB plan (we're on the 12GB one) is 10000/128 [05:07] ScottK: where are you at? [05:07] * solarion is sad because he just mvoed away from a whole-city free fiber rollout. :( [05:07] solarion: Outside Baltimore, MD (USA). [05:07] ah [05:07] solarion: Not free at all. I have Verizon FIOS. [05:08] ScottK: Liberty Communications is doing the rollout in my case [05:08] The trick is I have the painfully more expensive and slower version you get if you are a [05:09] Basically, cable brought in broadband and phone. Instead of taking it on the chin, the little company rolled out IPTV services. But the bandwidth was limited, so you could only have 3 set-top boxes [05:09] 'business' and need static IP. [05:09] tehy're solving that problem now. :) [05:10] My experience with cable TV companies and ISP services is that they know a lot about cable TV. [05:10] I should say that the little company was the local phone company for about 3 small towns in Iowa [05:11] so when cable came to eat their lunch, they set out to eat the big, multi-state company's lunch. :) [05:11] Good for them. [05:11] they're pretty cool [05:12] it's very very very nice to have a small company for your isp. Customer service isn't just lip service. [05:12] * solarion has at&t now and misses his old company [05:12] When I called Verizon about a DNS problem last week, the first person I talked to said, "What's DNS". It was a long day. [05:12] yeah, that's been my experience with AT&T [05:12] Where in Iowa? [05:13] ScottK: I have cable, but the cable TV provider who is also the ISP is the second-largest national telco. [05:13] Eventually I got it sorted out. [05:13] http://www.lcom.net/switchmaxx/home.htm [05:13] http://www.lcom.net/switchmaxx/page.htm?page=Fiber <- their fiber rollout page [05:14] ScottK: AT&T's voice recognition system is complete crap [05:14] but at least it's not like OG&E's [05:14] even the customer service rep said it's a PITA [05:14] ooh, I see mad java ninjas posting to the MC list [05:15] * nixternal might be interested in working with the java team a bit, except for the maven bit [05:16] "maven bit"? [05:16] ya, seems the java team enjoys maven [05:16] ah [05:17] nixternal, Err, that's very much not the case. It's rather that the java team is trying to find a way to support upstreams that enjoy maven without actually letting maven break stuff in the normal way. [05:17] though once I switch the company's appliance OS from CentOS to Ubuntu, I may get to use Maven even more [05:18] nixternal, There's a plan in place to trick maven so it can't do things the way it expects, which may be interesting to you. [05:18] persia: it is hard not letting maven break stuff...that is why I am not using it with CentOS at work...we just use ant for building, and the developers for the good stuff [05:18] that does sound interesting [05:19] though I don't know how inclined I would be to working on Java stuff outside of work anyways...I know I am heading towards that Java meltdown :) [05:19] * nixternal looks for more C++ and Python projects to bring him back to earth :) [05:19] https://wiki.ubuntu.com/JavaTeam/Specs/MavenSupportSpec [05:19] what's the big difference between ubuntu server and desktop? [05:19] solarion, Default package selection, installer interface, and contents of the CD. [05:19] the server doesn't have a desktop environment :) [05:20] and kernel [05:20] though I don't know the differences there much [05:20] Oh, right, and kernel. [05:20] But still, I'd class kernel under "default package selection". [05:20] true [05:20] so it's probably not worht re-installing then [05:20] just install the new packages and go [05:21] * solarion is repurposing an old laptop [05:21] solarion, Depends on your goal. If you have an old laptop, and you want a mail/dns server, reinstall might be faster. [05:21] Then again, it's certainly less bandwidth to just install different packages. Depends on your constraint. [05:21] I suppose I want to reprtition [05:23] persia: the Java team going to be at UDS? If so, I will hang out with them a bit and see how that all goes...that looks like a killer spec to be honest [05:23] are you going to UDS btw? [05:23] In that case, reinstall is probably easier. Juggling paritions on a live system is a tricky game. [05:23] yeah [05:23] nixternal, Some of us will be there. Koon won't be able to make it. [05:23] though since i have a root partition and two others, it wouldn't be too bad [05:23] boot single and monkey around a bit [05:24] meh, is bedtime for bonzo [05:24] thanks [05:24] nixternal, Needs a *lot* of packaging work. More hands are very welcome. [05:24] groovy...I think it is time for me to go to bed..need to wake up in 4.5 hours :/ [05:24] See the bootstrapping dependencies list. [05:24] YAY SNOW IN CHICAGO TONIGHT! [05:30] * solarion is gonna miss snow [05:32] i installed qt4 in gnome, the designer, translator, and assistant, but the designer cant find the assistant when i try to open it, anyone know how to fix this? [06:45] good morning === persia_ is now known as persia [06:48] morning dholbach [06:49] hi nellery [06:55] morning o/ [07:37] good morning [07:38] I wonder if there is a way to specify a build-time dependency as either (a AND b) OR c. IOW, the dependency is fullfilled if either c is installed or both a and b simultaneously [07:41] Laibsch, Not simply. I think you'd need a package name to represent the contents of the parentheses. [07:42] To ask differently, with which would you prefer it to build? [07:42] (a and b) or (c) [07:43] depends on the release [07:43] I guess I'd prefer c, but that is not available in hardy [07:44] only intrepid and jaunty [07:51] general question here. I'm an app developer, and my app uses a library that's packaged at 0.23, and now 0.25 is out and it has a feature my app needs. I contacted the package maintainer but no reply, and I see no signs that it's being updated. what do I do? http://packages.ubuntu.com/source/jaunty/liblo [07:53] ianm_, liblo has *lots* of reverse dependencies, including a couple applications we routinely have difficulty building cleanly. [07:53] I'd personally be opposed to updating it except in coordination with Debian to avoid issues. [07:53] Debian is currently frozen for the upcoming Lenny release. [07:54] So, from a distro point of view, I'll say "wait". [07:54] For your specific personal use, you might want to update the app locally. [07:54] * persia looks for the wiki page about updating library packaging. [07:55] so after Lenny, they might update it, meaning it might get into Jaunty? or +1? [07:56] It's very likely to be updated once Lenny releases, and would then be in either Jaunty or Jaunty+1, depending on the timing of the update. [07:56] the feature I need in liblo is setting the broadcast flag on a socket when the dest address is 255.255.255.255. frustrating how hard it is to get that to end users [07:57] so Jaunty+1, isn't that like a year away? [07:58] * ianm_ swoons [07:58] I suspect it's more likely to be in Jaunty, but it does depend on timing. [07:58] is that 9.04 ? [07:58] Does it really need all of 0.25? Are the changes in 0.25 small enough that a patch is useful? [07:59] Yes. Jaunty will be released as 9.04 [07:59] Here it is: https://wiki.ubuntu.com/MeetingLogs/devweek0809/PackageUpdates [08:00] my app needs almost none of 0.25. just this: if (ip[0]==255 && ip[1]==255 && ip[2]==255 && ip[3]==255) { int opt = 1; setsockopt(a->socket, SOL_SOCKET, SO_BROADCAST, &opt, sizeof(int)); } [08:02] ianm_, That's easier to apply then, although it's still likely to be Jaunty before it would be released to end-users. [08:02] is it possible/OK to include a custom liblo in a PPA .deb? [08:03] I'd recommend using either a patched version or 0.25 locally for development. If it's not updated by 25th December, then submit a bug with a patch for the bit you need. [08:03] No, but your PPA can contain a custom liblo. [08:03] Of course, this may well break the other 30 applications that use liblo. [08:03] it can't be the only one to use it? [08:03] Nope. [08:03] can you static link this stuff? [08:04] been a while since I've done C :D [08:05] You can static link things, but that's likely to break if there's any reason to upgrade (e.g. security fix). Depending on how you static link, you may end up having something that usually can't be installed. === Tweenaks is now known as Treenaks [08:08] hm [08:15] maybe there's a way to find out the socket # and hack it [08:25] persia: thanks for the advice [10:25] if anyone has any time on their hands I would appreciate a review of my package, http://revu.ubuntuwire.com/details.py?package=charm . Thanks :) [10:32] binarymutant: you shuld state in the changelog what the paches do, and why they are needed [10:32] s/shuld/should [10:33] good point :) [10:36] handschuh, can you see anything else wrong? [10:36] binarymutant: nothing so far, but still looking [10:36] ty for taking the time [10:37] binarymutant: http://revu.ubuntuwire.com/report.py/legal?upid=4035 - have you contact to upstream? [10:38] handschuh, to fix license issues and to say how much I liked the app a while ago [10:39] handschuh, I'm not sure why it reports that, it says gpl in the copyright file, and the gpl is shipped with the package now [10:40] binarymutant: well, it is recommanded to add a license header to every source-file [10:40] I thought I should stay away from modifying the source [10:41] binarymutant: thats why you can contact upstream to fix it :-) [10:41] binarymutant: of course you should. Upstream has to add license headers. [10:42] For GPL, it says in the "how to use the GPL" section that license should be in every source file. I've rejected many packages for exactly that reason, and seen others rejected by archive-admins when they were submitted. [10:42] is that policy? Because I had already bugged them to include the license in the orig tar ball [10:42] That said, patching the files to include licensing is unacceptable. It must be done upstream. [10:43] so this app would get rejected by debian for that? kinda sucks :/ [10:43] persia: in debian/copyright it's -> This package is free software ... Is it a mistake when there it's $Package is free software ... ß [10:43] -ß +? === DktrKranz2 is now known as DktrKranz [10:44] sebner, Not generally a terribly important mistake. Best to use the text from the upstream license. [10:44] well the ljcharm.py has it in there but the wrapper script charm doesn't :/ [10:44] persia: okay, thx [10:45] binarymutant, Wrapper scripts are sometimes exceptions, and sometimes not. Depends on complexity, opinion of reviewers, and what the reviewing archive-admin ate for breakfast that day. [10:45] and the setup.py doesn't either... [10:45] binarymutant, For best results, get it included. [10:45] binarymutant: i am sure, upstream will be glad to receive this hint about the license [10:46] I'll try thanks for that info, hopefully she wont think i'm bugging her [10:47] binarymutant, There's no race. If you've passed her a patch recently, there's no harm waiting. There's lots of other stuff to do in the meantime :) [10:48] lol k [10:48] slytherin: do you have got some time to review a small java package @ http://revu.ubuntuwire.com/details.py?package=uiflite now? [10:48] is there anything else I could do to this package? [10:50] binarymutant: everything else looks fine to me ... (about dpatch, I am not sure) [10:51] cool, thanks again handschuh [10:51] binarymutant: have you checked the lintian of the package (binary) and the source? [10:51] yeah a bunch of times [10:52] binarymutant: ok great. [10:52] binarymutant: (as long as it was empty) [10:52] s/was/is [10:52] handschuh: not right now. I will try to find time in night, about 6 hours form now. [10:53] handschuh, the only error is the one about jaunty [10:53] slytherin: ok, no problem, as persia mentioned, there's a lot other stuff to do [10:53] binarymutant: thats perfectly fine [10:54] if someone has a few time to check my revu package ....http://revu.ubuntuwire.com/details.py?package=sqliteman === DrKranz is now known as DktrKranz [10:56] eMerzh: take a look at http://revu.ubuntuwire.com/revu1-incoming/sqliteman-0811170000/lintian [10:56] oups :) thanks [10:57] eMerzh: there is more [10:57] eMerzh: your CMakeLists.txt should be modifyed by a patch [10:58] mounin [10:58] eMerzh: debian/control: Copyright: there is no need to add the word "Copyright" before every name [10:59] persia, what about distutils should setup.py also include gpl? [10:59] ok, it's a patch bu i must understand why he apply the patch in the diff [10:59] ok [11:00] eMerzh: also, explain the path in the changelog (what does the patch do and why did you add it) [11:00] ok :) [11:01] thanks [11:02] eMerzh: also make sure the lintian of the binary package is empty (check with "lintian <.deb-file>") [11:02] binarymutant, I don't usually see a GPL'd setup.py, and don't remember something being rejected for it, but I typically avoid python packages, so that may influence my experience. [11:03] eMerzh: take a look at http://revu.ubuntuwire.com/report.py/legal?upid=4031 - there are some license issues that should be fixed by the upstream authors [11:04] Oh okay, then the package should be fine in the license department, I don't know why it says unknown for ljcharm.py since it's already in there. Very cool thanks === wolfger_ is now known as wolfger [11:07] handschuh, ok i will contact upstream autor... but i must wait a new release with correct licences to re-apply to revu? [11:09] eMerzh: you don't have to wait for the new upstream version [11:09] eMerzh: you could reupload and post a comment, that you have contacted upstream about the licenses [11:10] Ok, great... [11:18] eMerzh: should I make a summary as a comment to your package? [11:19] no , everything is noted...and work is in progress :p [11:19] eMerzh: great [11:19] thanks to you for your help [11:20] eMerzh: you're welcome [11:23] I'm looking for my first advocate for http://revu.ubuntuwire.com/details.py?package=metalink or comments if needed :) [11:44] bmm: check your legal-file at http://revu.ubuntuwire.com/report.py/legal?upid=4019 - it shows some files w/o copyright [11:46] porthose, re slidentd merge, IIRC you noticed an issue, have you it handy? [11:47] bmm: delete the commented lines of your debian/rules file [11:47] handschuh: I filed an upstream bug for the licenses http://sourceforge.net/tracker/index.php?func=detail&aid=2278178&group_id=148879&atid=772980 [11:47] Sourceforge bug 2278178 "License missing from some sources" [Pri: 3,Open] [11:47] handschuh: ok, will do. [11:48] bmm: you are not the upstream author? [11:48] handschuh: yes, I am :D [11:48] handschuh: so I should put out a new release? [11:48] * persia admires bmm's ability to maintain clear context separation [11:49] bmm: yes [11:49] handschuh: ok, I'll do that then :D [11:49] handschuh: thanks! [11:50] bmm: no problem [11:52] bmm, isn't metalink a bit too generic for a metalink file generator ? [11:52] bmm: your debian/rules looks very large ... you could look at cdbs (no manditory) [11:53] since, it's a technology name [11:53] anyne want to take bets on how much, if any, of the hardware on my new laptop is supported in intrepid? [11:54] directhex: brand? [11:54] joaopinto: well, it was meant as a "metalink " thing. [11:54] handschuh, dell [11:54] handschuh, You might be interested in reading about debhelper 7. It can make rules files shorter for common cases without switching to CDBS. [11:55] directhex, Which OS is supported on your laptop? [11:55] persia: didn't know that. thanks! [11:55] persia, formally, vista [11:56] * persia refuses to place a bet [11:56] bmm, I would name it something like metalink-generator, or metalink-crator, just metalink sounds more like a metalink client [11:56] ops, creator [11:56] directhex: my dells hardware is fully supportet so i bet 90% :-) [12:01] persia, spoilsport! [12:05] directhex: I would guess 90%+ [12:05] directhex: you might be having problem with card reader or fingerprint reader. [12:06] persia: Can you elaborate on the debhelper 7 making rules files shorter thing? [12:07] It's like CDBS, but less bad, I hear. [12:07] What's bad about cdbs? [12:08] CDBS is bad ? [12:08] slytherin, my bets are on a lack of graphics support or webcam support. AFAIK some nutter actually put the only-exists-in-an-obscure-git-tree wireless drivers onto the ubuntu kernel [12:08] * soren likes cdbs [12:08] CDBS is awkward to debug [12:08] * handschuh also likes cdbs [12:08] we use dh7 a lto in pkg-mono [12:08] It can be more difficult to get it to do special things. For lots of simple things it's great. [12:09] soren, It introduces a new command `dh` that does all the normal stuff for whichever rule is being called, as I understand it. I've not yet created a package with it, but those I've seen were rather readable. [12:09] most packages start life simple ;) [12:09] How is the legal check performed on REVU, is that something you can run at home? [12:09] let me grab a pkg-mono dh7 sample [12:09] joaopinto, Read the sendmail debian/rules. [12:10] persia: Oh, I see now. [12:10] http://svn.debian.org/wsvn/pkg-mono/gluezilla/trunk/debian/rules?op=file&rev=0&sc=0 is a "complex" dh7 example [12:10] That's rather convenient. [12:10] Indeed, and it does it without quite so much black magic. Still some magic, but less. [12:10] persia: Put a warning in front of that sendmail suggestion next time, please. That's just evil. [12:10] * soren doesn't mind black magic [12:10] :) [12:11] wgrant, Someone asked "what's wrong with CDBS". That's my standard answer [12:11] * persia is a big CDBS fan [12:11] It's just that CDBS can get unwieldy sometimes, which gets frustrating. [12:12] soren, I agree most of the time. I only stopped recommending CDBS for everything after being pointed at sendmail. There's limits to what one should try to do in CDBS. [12:13] That is quite horrible looking indeed. [12:13] persia, would debhelper turn sendmail rules any better ? [12:13] CDBS is very good for java packages. :-) [12:14] joaopinto, That's a false question. CDBS doesn't replace debhelper. CDBS calls debhelper for you. But yes, it would be tons easier to understand that rules file if it were done as a single makefile. [12:14] CDBS is very good for standard autoconf builds [12:14] persia, I mean, a non cdbs debhelper make file [12:14] Indeed, for those anything else is likely to be buggy, although dh7 is promising. [12:15] joaopinto, I thought so, which is why I answered in two parts :) [12:16] I just don't see why does CDBS get blamed for a complexity which is from the sendmail build/install process, not from using CDBS [12:16] slytherin, For the 90% case, I agree, although I suspect dh7 would work as well. It's the 10% that make non-CDBS the better choice. [12:17] joaopinto, The point is that CDBS more quickly becomes unreadable when dealing with very complex build systems. [12:18] Personally, I think any CDBS rules file longer than 24 lines should be redone differently, as it means that the CDBS framework isn't designed to handle your package. [12:18] well, non CDBS is harder to read simple building systems :P [12:18] Less than 24 lines, CDBS probably makes life easier. [12:18] .. for... [12:18] joaopinto, Is directhex's example harder to read? [12:19] persia, I am not refering to dh7, didn't tried it yet [12:19] Well, the shortest dh7 is: [12:19] joaopinto, Oh, then I agree with you. Like I said, I'm a big fan of CDBS. [12:19] #!/usr/bin/make -f [12:19] %: [12:19] dh $@ [12:19] hehe [12:20] sebner, Indeed, although not every package can do that. Still, that just means "default packaging : nothing to see here : move along". About the same as a CDBS rules file that only includes debhelper.mk. [12:20] persia: heh, yes. true =) [12:21] I'd much rather see that then the mess dh_make produces. That's just annoying, and almost everyone makes a mistake the first few times they try to determine what they need and what they don't. [12:22] okay, laptop report: audio works, wired network works, wireless network works, bluetooth works, graphics works (!), webcam works [12:22] it all bloody works out of the box [12:22] webcam resolution sucks though ;) [12:22] directhex, You lose. Time to recompile the kernel. [12:22] persia, to cause myself pain? o_o [12:23] directhex, Anyway, you probably ought add that to whereever the laptop support list went. Tell everyone else it's a good choice. [12:23] even the wifi killswitch works, incl. the LEDs [12:23] i wasn't anticipating this level of awesome [12:24] and the scroll-edges of the touchpad [12:24] directhex: does also the bt-led works properly? [12:25] handschuh, yes [12:25] okay, i lie, the horizontal scroll-edge doesn't work. that's the only busted thing i can find [12:25] directhex: you can change that in the mouse-settings [12:25] handschuh, you're right. that's everything then [12:26] handschuh, a laptop dell only started shipping a couple of weeks ago, full of bleeding edge tech, and it works 100% [12:26] directhex: good thing [12:26] directhex, If it's that new, why didn't you get it with Ubuntu? [12:27] directhex: I hope you don't have the Intel wireless AGN (with iwlagn driver) because then you have to keep an eye on your shutdowns :) [12:27] persia, too new to be sold with ubuntu [12:27] bmm, oh? [12:29] directhex: iwlagn in the current intrepid kernel seem pretty broken. Package capturing will get you allot of kernel messages and the occasional hang and shutdown with the driver and wireless still on can keep your laptop in limbo without shutting down :) [12:30] directhex: so if lsmod tells you you are using iwlagn, make sure you keep an eye on the shutdown process now and then ;) [12:30]