**** BEGIN LOGGING AT Mon Aug 31 02:59:59 2015 Aug 31 04:59:03 build #77 of mpc52xx is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/mpc52xx/builds/77 Aug 31 06:24:14 build #84 of ar71xx.nand is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx.nand/builds/84 Aug 31 07:26:56 build #84 of pxa is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/pxa/builds/84 Aug 31 09:56:17 I have sent two patches to openwrt-devel that restore the buildability of the AP121 target. I think they should get into CC as they are fixing regressions... am I right? if so what else shall I do for that? Aug 31 09:57:09 these are the two issues, with the patches attached to them: https://dev.openwrt.org/ticket/17885 https://dev.openwrt.org/ticket/14571 Aug 31 09:57:52 github pull requests would be so much nicer... but the main repo still seems to be the SVN one. Aug 31 10:05:04 attila_lendvai: you are doing it wrong - patches to main list should be submitted to mailing list Aug 31 10:05:10 and its missing a signed off Aug 31 10:05:19 https://dev.openwrt.org/wiki/SubmittingPatches Aug 31 10:05:59 plntyk: I've also submitted to the list, but no answer. it seems to be a constant stream of patches, I thought it's lost in the noise... will read up on signed-off Aug 31 10:06:54 attila_lendvai: i see them - but those are missing signed off - I dont now if somebody noticed them Aug 31 10:06:55 http://patchwork.ozlabs.org/patch/508527/ Aug 31 10:07:49 the second patch was missed by parser or so on the mailing list Aug 31 10:13:27 * attila_lendvai should have sent one patch per mail Aug 31 10:20:04 plntyk: shall I archive that patchwork entry and resend the two patches with a signed-off entry in two separate mails? Aug 31 10:26:08 resend in 2 separate mails with sign off should be enough - no need to archive that patchwork entry (it was parsed from mail list and is available in the mail list archive anyway) Aug 31 10:59:49 * attila_lendvai has resent those two patches Aug 31 11:21:53 attila_lendvai: if you lament the missing pull requests it means you are using git, so I suggest using git send-email for sending patches ;) Aug 31 11:22:10 inline patches are much easier to review than attached ones Aug 31 11:22:45 easier to review, and about 1000 times more likely to get mangled :) Aug 31 11:23:21 karlp: I have yet to see patch being mangled by "git send-email" Aug 31 11:23:30 *a patch Aug 31 11:24:19 pasting the patch into an email client is a recipe for desaster, there I agree with you ;) Aug 31 11:30:57 I was only commenting ont he inline portion :) gitsendemail exists because too few clients could do inline patches without mangling and special handholding. Aug 31 11:31:13 but I've yet to see a mangled patch attached ;) Aug 31 11:34:16 I've thought that there's some magic button in patchworks to apply the patches. I might be too young but using emails for this kind of job today is probably not one of the most efficient ways available... but I'll use git send-email next time. Aug 31 11:36:23 attila_lendvai: there isn't and maybe even shouldn't Aug 31 11:36:27 attila_lendvai: it might not be the most efficient, but it allows a wide range of clients to "read" them (and easy "offline" review) Aug 31 11:36:31 after applying a patch one should make sure it e.g. still compiles Aug 31 11:37:02 so you usually want to download a patch first anyway Aug 31 11:37:05 and apply it locally Aug 31 11:37:08 I mean a developer at least Aug 31 11:37:16 and since you already have it locally... Aug 31 11:37:28 i guess "git push" is just as easy as clickig patchwork button Aug 31 11:37:35 so it doesn't make a much difference Aug 31 11:37:36 buildbots can make sure of that. I'd apply several patches that look ok into a branch, let the buildbot do the compile, then test the firmware, then pull them over into the public devel branch. Aug 31 11:37:48 build #84 of ppc40x is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/ppc40x/builds/84 Aug 31 11:37:49 that would be nice for sure Aug 31 11:37:57 unfortunately it requires more work Aug 31 11:38:00 and builders Aug 31 11:38:18 and don't have such feature and probably couldn't afford builders Aug 31 11:38:21 attila_lendvai: proper buildtesting would require several hours per patch Aug 31 11:39:22 having few builders, prefferably each with its own set of devices, etc... Aug 31 11:39:29 that would be expensive :) Aug 31 11:39:36 in both: setup and maintanence Aug 31 11:39:42 KanjiMonster: several hours per a bunch of patches. I'd pull over several patches in one go. probably group them by target and likelihood of breaking something. but I don't want to be a PITA, just pondering... Aug 31 11:40:33 well, I can volunteer to test the AP121 platform if there's some kind of formal process for that Aug 31 11:40:43 attila_lendvai: if they touch e..g toolchain, you technically need to build *all* targets. and if they introduce new options disabled by default you can't test them automatically Aug 31 11:42:32 hrm, I obviously don't see the full complexity of this. what pushes me to think about this is that my two patches are pretty trivial, restore buildability, and I'd expect them to get quickly through whatever processes are set up. if not, then it means that someone is struggling somewhere. Aug 31 11:43:08 attila_lendvai: each target has its own maintainer, and we try to avoid stepping on others toes Aug 31 11:44:30 * rmilecki looking at KanjiMonster messing with rmilecki's brcm47xx and CMDHACK Aug 31 11:44:34 just kidding :P Aug 31 11:44:39 ;P Aug 31 11:46:21 well, the firmware didn't build for the AP121 target, maybe for years judging from this issue (https://dev.openwrt.org/ticket/14571). it's probably something rarely used, although these days it's <$10 per piece, so that will probably change. Aug 31 11:47:38 attila_lendvai: also remember that noone here is paid to work on openwrt; so the core devs do it in their free time Aug 31 11:49:12 I know. I'm almost exclusively working on opensource projects for the past 10 years, sometimes contracting for some projects to pay the bills or starting something with friends... Aug 31 12:09:42 Hi Chaps, having some problems with odhcpd, it appears to be ignoring prefix length on an interface. Can someone check my work? Aug 31 13:30:16 KevinDB: can you provide a proper bugreport then i will have a look tomorrow Aug 31 13:30:57 cyrusff: it's the one I just told you about ;) Aug 31 13:31:19 Ah Aug 31 13:31:32 Anyway details are always appreciated Aug 31 13:48:51 cyrusff: will do Aug 31 22:24:05 build #78 of mpc83xx is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/mpc83xx/builds/78 Sep 01 01:35:35 build #78 of ar71xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx/builds/78 **** ENDING LOGGING AT Tue Sep 01 02:59:58 2015