ari-tczew | ScottK: could you send MemoMsg to him? | 00:00 |
---|---|---|
udienz | ScottK: thanks for responding my email at ubuntu-motu-mentors | 00:03 |
=== yofel_ is now known as yofel | ||
=== freeflyi1g is now known as freeflying | ||
=== test__ is now known as anoteng | ||
hakermania | Hello everybody. | 08:25 |
hakermania | udienz: Hello, can you send me again the link talking about the patches, please? Thank you. | 08:27 |
hakermania | Dos anybody knows what are "*.in" files? | 10:11 |
geser | sort of templates | 10:12 |
micahcowan | They can be anything, but the most frequent situation are files that are processed by the "configure" script to produce the final file. | 10:12 |
micahcowan | configure looks for variables like @VAR@ and replaces them according to whatever settings it has. Makefile.in and config.h.in are typical. | 10:13 |
hakermania | Thank you, tomboy's source has a lot of them and I wondered.... | 10:14 |
micahcowan | Often times, a Makefile.in is itself a generated file (from Makefile.am). | 10:14 |
micahcowan | automake turns Makefile.am into Makefile.in (unlike configure, it doesn't just substitute, it adds a lot of material, and processes the user-provided values); then configure turns the Makefile.in into a Makefile, based on the users parameters to the configure script. | 10:15 |
tumbleweed | geser: thanks :) | 10:50 |
geser | tumbleweed: re CamelCase: I tried to mimic the LP API which also uses CamelCase when I wrote it to not be much more different than using the LP API directly | 10:52 |
geser | but as there are not many users of lpapicache (requestsync and ubuntu-build if I didn't miss one) it can be changed | 10:52 |
geser | and any ideas how to make it easier usable? | 10:53 |
tumbleweed | geser: yeah. requestsync also needs the compatible mail and lp interfaces, which is one reason for some of the oddities | 10:53 |
tumbleweed | there's nothing intrinsically wrong with the way it does anything, it's just different, so I looked at it and thought do I want to use this or just do things the normal way? | 10:54 |
geser | if you only need to access an object once then there is no big benefit in it | 10:55 |
tumbleweed | no, but it has handy functions like canUploadPackage etc for things that are tricky to do in pure launchpadlib | 10:55 |
geser | but the old requestsync code fetched e.g. an series or archive object in each function which needed it which made requestsync slow | 10:55 |
tumbleweed | I couldn't use things like that, because thye talk lpapicache wrapper clases as arguments, so you need to use lpapicache everywhere | 10:56 |
geser | and you already have LP objects or only strings? | 10:57 |
tumbleweed | lp objects, probably | 10:57 |
geser | this function could be extended to also use lp objects instead of the lpapicache classes if it would help you | 10:59 |
geser | or the logic could be moved to an own function and lpapicache uses it | 11:01 |
tumbleweed | yeah, I'm sure both would work, but libsupport might be a better home for things like that. Although no point moving them until necessary | 11:02 |
geser | I planned to use lpapicache also in other scripts but was hesitant to make them "depend" on u-d-t python modules for no benefit (many scripts can be used stand-alone) | 11:04 |
tumbleweed | aah, I hadn't really considered that one | 11:05 |
=== yofel_ is now known as yofel | ||
hakermania | Ok, I have a question about patching a system. I followed these rules: https://wiki.ubuntu.com/PackagingGuide/PatchSystems (where it says 'The easiest way'). After I make the patch (I simply edited the README file), the patch went to debian/patches. How to I apply the patch now so as the README file to change as I want to? I tried path -p0 < debian/patches/mypatch.patch but there are a lot of error messages! Or the package is ready fo | 12:23 |
ari-tczew | hakermania: which patch system does your package use? | 12:25 |
hakermania | what-patch outputs quilt. | 12:26 |
ari-tczew | hakermania: did you use command path -p0? patch is correct btw | 12:26 |
ari-tczew | but you shouldn't do this | 12:26 |
ari-tczew | just add filename to debian/patches/series | 12:26 |
cdbs | ScottK: I am sorry for that, if you read the previous upload of mine for the same package, you will understand | 12:26 |
cdbs | The same fix worked in the previous upload | 12:27 |
cdbs | and here it worked on my (slightly old) natty pbuilder | 12:27 |
ari-tczew | cdbs: it's not explanation | 12:27 |
cdbs | it appears a recent gcc change caused that FTBFS | 12:27 |
ari-tczew | cdbs: how often do you update pbuilder? | 12:27 |
cdbs | ari-tczew: around once a week or twice | 12:28 |
ari-tczew | cdbs: lol. do it everyday | 12:28 |
cdbs | if I had the proper connection :( | 12:28 |
ari-tczew | add to cron, maybe will helpful | 12:28 |
cdbs | this one frops very frequently, it takes me hours to upload small changes | 12:28 |
cdbs | *drops | 12:28 |
hakermania | ari-tczew: Ok, the patchs' name was automatically placed into the series file. My question is how do I apply the patch now ? Is the package ready for 'debuild -S' and if somebody try to install it, the patch will automatically take place or what? | 12:29 |
hakermania | I mean, how do i test if the patch really works? | 12:29 |
cdbs | I will need to consult someone on how to fix it, since a fix which worked a month ago isn't working now, wierd | 12:29 |
micahg | cdbs: right, there was a gcc update, doko posted to ubuntu-devel about it | 12:30 |
ari-tczew | hakermania: create a second directory, unpack source there and try to patch it. however, I prefer to just build package (binary) from *.dsc. | 12:30 |
cdbs | micahg: yes, and it appears I uploaded the same day :( | 12:31 |
cdbs | and I didn't get the FTBFS mail, oh its lying in spam | 12:31 |
hakermania | ari-tczew: I don't get you :/ | 12:32 |
cdbs | ScottK: fixing right now, for sure I will do it | 12:32 |
ari-tczew | hakermania: you don't need to use 'patch'. builder will apply patches | 12:33 |
hakermania | ari-tczew: So I can now for example use 'debuild -b' and test if the DEB package works with the changes of my patch? | 12:34 |
tumbleweed | hakermania: if your package uses quilt, quilt push -a will apply all patches, and quilt pop -a will unapply them all | 12:36 |
hakermania | tumbleweed: Cool, give me a sec to test these 2. | 12:36 |
tumbleweed | hakermania: I recommend spending some time getting familiar with quilt | 12:37 |
* ari-tczew is guessing that someone else will be better teacher. | 12:37 | |
hakermania | ari-tczew: :P | 12:38 |
hakermania | tumbleweed: Thx, I'll have an extended seach on the inet for quilt and its uses. | 12:38 |
ari-tczew | siretart: ping | 12:39 |
cdbs | hakermania: http://wiki.debian.org/UsingQuilt | 12:44 |
hakermania | cdbs: Your name has to do with patching, I've heard about it somewhere. Btw thanks for the link :) | 12:45 |
cdbs | Its much more than patching :) | 12:45 |
cdbs | cdbs is a full build system for packages | 12:45 |
tumbleweed | some of us don't like cdbs very much, but this cdbs is ok :P | 12:46 |
cdbs | :D | 12:46 |
cdbs | dh7 and debhelper were registered, so I had to settle for this | 12:46 |
geser | cdbs: passing -l libs with LDFLAGS doesn't work anymore since the ld --no-as-needed change. You have to put them after the object files. In many cases you can use either ..._LDADD or LIBS for it (look into the Makefile which variable could be used for it) | 12:50 |
cdbs | geser: I already uploaded that change to my PPA for a test build, will see what happens | 12:50 |
cdbs | I have to rely on my PPA for test builds nowadays, until the situation improves, and the weather becomes better | 12:51 |
cdbs | My connection is being disrupted by poor weather | 12:51 |
hakermania | cdbs: Patches should have extension .patc or .diff? | 12:52 |
hakermania | .patch* | 12:53 |
cdbs | hakermania: both are okay | 12:53 |
cdbs | some have no extensions at all :) | 12:53 |
cdbs | .patch is what I prefer personally for a new package | 12:53 |
geser | cdbs: I also see in debian/rules for courier on codehosting that it uses LDFLAGS=-lcrypt when calling configure. Please check the build log if configure finds everything it needs. | 12:53 |
cdbs | geser: that change applies when it builds the webadmin part, here I am making the change in courier/module.esmtp/Makefile | 12:54 |
ari-tczew | hakermania: I suggest to use the same extension as rest of patches in package (in debian/patches/) | 12:54 |
hakermania | ari-tczew: Only for typical reasons? | 12:54 |
kklimonda | geser: wouldn't it make more sense to prepare an actualy patch for packages that fail due to linking issues and then submit it to debian and upstream? tweaking debian/rules seems really hacky | 12:55 |
cdbs | geser: it appears the PPA rejected because I did debuild -sd :( will have to upload a whole 9 MB now | 12:55 |
ari-tczew | hakermania: for clear case | 12:55 |
hakermania | Ok | 12:56 |
geser | cdbs: so you didn't change debian/rules like the changelog entry says? | 12:56 |
cdbs | geser: no, I am now modifying upstream makefile in the pending upload, that's a better approach | 12:58 |
geser | cdbs: even if the LDFLAGS in debian/rules is not your change, it would be good to check if it's needed (or some feature missing) as using LDFLAGS that way doesn't work anymore | 12:59 |
cdbs | hmm, yes, will check that | 12:59 |
ari-tczew | geser: how often toolchain will be changing? | 12:59 |
cdbs | but that's a debian-specific change, so will have to file a bug there about it first | 12:59 |
ari-tczew | we don't have a time for 10x updating patches to fix FTBFS with binutils-gold. it's being annoying | 13:00 |
geser | ari-tczew: I hope not anymore for natty, the next change if any I'd expect for natty+1, but better ask doko if you want to know if any other changes are planned for natty | 13:00 |
ari-tczew | geser: what's the point to changing toolchain? | 13:01 |
geser | as far as I understand the reason for the recent changes is to lessen the needed dependencies as many program (and therefore packages) link to many unneeded libraries making library transitions harder than needed | 13:02 |
cdbs | yes, the toolchain changes have benefited a lot in that sense | 13:03 |
cdbs | hakermania: (just an FYI in case you don't know) All packages in Debian are synced very frequently to Ubuntu | 13:04 |
ari-tczew | cdbs: till DIF | 13:05 |
cdbs | ari-tczew: it can be done even after DIF, just that a sync must be requested | 13:05 |
ari-tczew | geser: does debian planning to approve similiar toolchain to ours? | 13:05 |
cdbs | ari-tczew: yes | 13:05 |
cdbs | ari-tczew: after squeeze release | 13:05 |
cdbs | its planned for wheezy | 13:05 |
ari-tczew | I'm asking due to curiosity about forwarding patches to fix building with binutils-gol | 13:05 |
ari-tczew | d | 13:06 |
geser | ari-tczew: the ld --no-add-needed (aka no DSO indirect linking) is planned for Wheezy, don't know about the ld --no-as-needed change | 13:06 |
geser | doko asked me forward my ld --no-as-needed changes to Debian and user-tag them accordingly | 13:07 |
ari-tczew | geser: what does it mean? | 13:10 |
geser | user-tagging? it's similar to the tags in LP and allows to search for bugs with a specific tag in the Debian BTS | 13:11 |
cdbs | Holy, my package upload to the PPA is failing as it did earlier | 13:17 |
cdbs | and I don't have any other way to test-build! | 13:18 |
cdbs | bug #687622 is responsible | 13:18 |
ubottu | Launchpad bug 687622 in update-manager (Ubuntu) "Unable to upgrade with update-manager" [Undecided,New] https://launchpad.net/bugs/687622 | 13:18 |
cdbs | I guess | 13:18 |
cdbs | err, bug #687662 | 13:18 |
ubottu | Launchpad bug 687662 in Soyuz "Upload processor attempts to verify hashes against expired files" [High,Fix committed] https://launchpad.net/bugs/687662 | 13:18 |
hakermania | What's the difference of launchpad bugs and bugzilla ? | 13:51 |
kklimonda | hakermania: ugh.. both let you report and triage bugs.. everything else is different | 13:54 |
geser | different software for managing bugs | 13:54 |
kklimonda | geser: what about my questions wrt patches for link failures? :) | 13:54 |
hakermania | But, bugzilla bugs rhen referred by users in lp, are called "upstream" | 13:54 |
hakermania | when* | 13:55 |
kklimonda | hakermania: different projects use different places for bug reporting. What we call upstream is (among other things) the bug tracking system that the project uses | 13:56 |
hakermania | kklimonda: Thx | 13:56 |
mr_pouit | kklimonda: (wrt patches for link failures) I agree with that. I find these kind of changes useless (and I had to redo all the work of the previous uploader when I wanted to forward upstream…) | 13:57 |
geser | kklimonda: missed it. yes, fixing the build system and forwaring the patch to Debian and upstream is better. | 13:57 |
=== emma is now known as em | ||
SpamapS | Daviey: meh... 15 month old disagrees with you that I should be in bed. ;) | 14:40 |
=== IVBela1 is now known as IVBela | ||
hakermania | How can i turn off the messages which tell who enters and quits the room? | 16:00 |
evaluate | hakermania, depends on your client | 16:00 |
hakermania | evaluate: pidgin :/ | 16:00 |
boulabiar | hakermania, on xchat, settings> hide join/Part messages | 16:01 |
evaluate | hakermania, hmm, not sure, I don't use pidgin for IRC. Try #pidgin ? | 16:01 |
hakermania | thx | 16:01 |
=== bilalakhtar_ is now known as cdbs | ||
=== paul__ is now known as Elbrus | ||
=== test__ is now known as anoteng | ||
Daviey | SpamapS: That was my guess :) | 21:15 |
OwaisL | Hi everyone. How much time does it usually take for REVU to reflect uploads? | 21:22 |
OwaisL | I just uploaded a couple of packges in last 30 mins but revu.ubuntuwire.com is not showing them. | 21:23 |
OwaisL | Hey, would anyone like to have a look at a package I just uploaded to Motu for inclusion in Natty? | 21:41 |
OwaisL | http://revu.ubuntuwire.com/p/gmailwatcher | 21:41 |
anoteng | anybody know what this error message from launchpad means (rejected PPA upload): Unhandled exception processing upload: 'NoneType' object has no attribute 'md5' | 22:29 |
Bachstelze | anoteng: looks like a bug in LP | 22:33 |
=== Daviey_ is now known as Daviey | ||
=== FlannelKing is now known as Flannel | ||
=== udienz1 is now known as udienz |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!