| === shivaram is now known as slingamn | ||
| cpaelzer | hi wgrant, is gdb on (emulated) riscv64 fully functional? | 06:49 |
|---|---|---|
| wgrant | cpaelzer: it's worked fine in my experience | 06:50 |
| cpaelzer | I was trying to debug a crash and can't see a backtrace at all - which might be part of the problem I try to analyze OR it might be that gdb support isn't complete on that platform | 06:50 |
| cpaelzer | ah ok, then I'll assume ti is part of the problem that I search | 06:50 |
| cpaelzer | thanks wgrant | 06:50 |
| cpaelzer | I wanted to make sure I won't go red herring fishing ... | 06:50 |
| wgrant | cpaelzer: Ah, I wouldn't be *that* sure it was reliable. | 06:51 |
| cpaelzer | hehe | 06:52 |
| cpaelzer | it is fine, hearing that it was seen working is what I wanted to hear | 06:52 |
| cpaelzer | but e.g. I get (so far) no backtraces at all | 06:52 |
| cpaelzer | so I got wondering if it might be a platform issue | 06:52 |
| cpaelzer | got a bt now | 06:54 |
| cpaelzer | need to find where it break, but that is problem-specific not riscv64 specific | 06:54 |
| Laney | seb128: https://launchpad.net/~laney/+archive/ubuntu/lo-test1/+build/19822873 indeed the no-change rebuild of 6.4.5 crashed the compiler cc vorlon - hopefully 6.4.4 doesn't have this problem ... | 09:03 |
| seb128 | 'fun' | 09:06 |
| seb128 | (not) | 09:06 |
| cpaelzer | Laney: seb128: there have been armhf gcc crashes since gcc-10 | 09:12 |
| cpaelzer | is this one reproducible at least - the othere ones were not | 09:13 |
| Laney | yeah it seems to be | 09:13 |
| cpaelzer | bug 1890435 and bug 1887557 (the latter being fixed) | 09:13 |
| ubottu | bug 1890435 in gcc-10 (Ubuntu) "gcc-10 breaks on armhf (flaky): internal compiler error: Segmentation fault" [Undecided,New] https://launchpad.net/bugs/1890435 | 09:13 |
| ubottu | bug 1887557 in gcc "GCC 10 is crashing with an internal compiler error when compiling Mesa" [Medium,Fix released] https://launchpad.net/bugs/1887557 | 09:13 |
| Whoopie | Hi, do we have support for wireguard in 20.04's VPN panel? Which package does provide it? | 09:13 |
| Laney | I don't think we got a successful armhf libreoffice build with gcc-10 yet | 09:14 |
| Laney | at least not afaik | 09:14 |
| cpaelzer | ok, so on the good side yours at least can be tracked and debugged - on the bad side thou how to compile until then :-/ | 09:14 |
| Laney | not sure if it's always the same object though | 09:14 |
| Laney | 24 hour turnaround time! | 09:14 |
| cpaelzer | outch | 09:14 |
| cpaelzer | yeah it might be just compiling so much that it fails at some random point each time | 09:14 |
| seb128 | to add even more fun, that is in the way of the icu transition | 09:15 |
| Unit193 | Whoopie: Sounds like something you may wish to ask in #ubuntu. | 09:17 |
| Whoopie | Unit193: ok | 09:17 |
| Laney | cpaelzer: I'll file it separately in case there are two bugs here, can be merged if needed later | 09:18 |
| cpaelzer | yep | 09:20 |
| Laney | ricotz: did you see a successful armhf build of LO with gcc-10 yet? can you point me to some fail logs pls? | 09:21 |
| Laney | how come it works @ Debian? | 09:22 |
| ricotz | Laney, is debian using gcc 10 by default yet? | 10:10 |
| Laney | ah | 10:10 |
| Laney | true, I didn't check that! | 10:10 |
| ricotz | https://launchpad.net/~libreoffice/+archive/ubuntu/libreoffice-prereleases/+build/19818811 | 10:10 |
| ricotz | sorry, it isn't there yet | 10:11 |
| Unit193 | ricotz: Yes. | 10:12 |
| Laney | https://buildd.debian.org/status/fetch.php?pkg=libreoffice&arch=armhf&ver=1%3A7.0.1%7Erc1-1&stamp=1597356576&raw=0 yeah it did use 10 | 10:12 |
| Laney | ricotz: have you got a build of that versin 7.0.1~rc1? | 10:13 |
| ricotz | Laney, the links points to a 7.0.1.1 ppa build | 10:13 |
| Laney | cool | 10:14 |
| Laney | keep an eye on that one then | 10:14 |
| Laney | ... in 12 hours ¬_¬ | 10:14 |
| ricotz | as expected the focal armhf build is happy | 10:14 |
| ricotz | with gcc 9 | 10:14 |
| Laney | yeah | 10:14 |
| ricotz | so yeah, using gcc 9 this specific build would be an option | 10:15 |
| Laney | I just kicked off builds of 6.4.4/6.4.5 with gcc-9 (for all arches though) in two silos | 10:15 |
| ricotz | ok | 10:15 |
| Laney | in case this autopkgtest problem is some other 6.4.5 regression | 10:16 |
| ricotz | my 6.4.6~rc2 builds were happy on focal | 10:16 |
| Laney | did you try the autopkgtest on armhf? | 10:16 |
| ricotz | https://launchpad.net/~libreoffice/+archive/ubuntu/ppa/+sourcepub/11525210/+listing-archive-extra | 10:16 |
| Laney | that's what we were seeing sadness with on groovy lately | 10:17 |
| ricotz | no, I don't an environment for that | 10:17 |
| Laney | ah yeah, this is where you need upload rights ;-) | 10:17 |
| Laney | or a raspberry pi :p | 10:18 |
| seb128 | Laney, http://autopkgtest.ubuntu.com/packages/libreoffice/focal/armhf has an happy 6.4.5/armhf/focal result | 10:19 |
| Laney | ok cool | 10:19 |
| Laney | didn't notice it got SRUed there | 10:19 |
| seb128 | I mentioned it the other day when we talked about new version regression, but that probably got lost in the middle of the discussion | 10:20 |
| Laney | likely | 10:20 |
| Laney | too many versions of different things going on | 10:20 |
| seb128 | I hope we get enough clarity to find a way forward with one of the options | 10:21 |
| seb128 | probably for monday at this point though | 10:21 |
| doko | ricotz, Laney: LO successfully built using gcc-10 in the test rebuild | 10:51 |
| Laney | k! | 10:56 |
| === Wryhder is now known as Lucas_Gray | ||
| xnox | vorlon: yes, and i didn't know about force-same-destination. Also not sure if i can copy things to groovy, i think as a core-dev i can only upload things to proposed. | 15:00 |
| Laney | I think that's true | 15:12 |
| ahasenack | vorlon: hi, just found an issue with motd-news-config, looks like it inherited "Priority: required" from src:base-files https://launchpad.net/ubuntu/groovy/amd64/motd-news-config/11ubuntu11 | 17:18 |
| ahasenack | is that something I can override in d/control for it? What should the new value be, optional? | 17:18 |
| * ahasenack sets optional | 17:34 | |
| vorlon | ahasenack: right, though the archive also has priority overrides which take precedence, so I'll also override it now. | 18:04 |
| ahasenack | thanks | 18:04 |
| ahasenack | vorlon: for all releases, i suppose? | 18:04 |
| vorlon | ahasenack: yes | 18:04 |
| ahasenack | I found this testing a release upgade from a focal desktop to groovy | 18:04 |
| ahasenack | ok, thanks | 18:04 |
| vorlon | I mean, I can only override it for series where the binary is already published | 18:04 |
| ahasenack | so it will have to be a reminder in the sru? | 18:04 |
| vorlon | if you set the priority in debian/control before SRUing, I *think* that will DTRT, but it would be good to have the reminder as well in the SRU bug | 18:05 |
| vorlon | best to make it part of the test case | 18:05 |
| ahasenack | ok, I will add it | 18:05 |
| ahasenack | I ahve 9 tests cases already, what is another one :) | 18:05 |
| === ebarretto_ is now known as ebarretto | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!