/srv/irclogs.ubuntu.com/2024/05/02/#ubuntu-devel.txt

=== dviererbe1 is now known as dviererbe
=== pushkarnk1 is now known as pushkarnk
ahasenackLocutusOfBorg: hi, is virtualbox (sync) 7.0.16-dfsg-3 meant to be deleted from noble unapproved, in favor of 7.0.16-dfsg-2ubuntu1.24.04.2?16:56
ahasenackhttps://launchpad.net/ubuntu/noble/+queue?queue_state=1&queue_text=virtualbox16:56
ahasenackops, that was -hwe16:58
ahasenackI meant 7.0.16-dfsg-3 vs 7.0.16-dfsg-2ubuntu1 for src:virtualbox in noble-unapproved16:59
Eickmeyerwaveform: Saw the Rpi testcases, and gave a cursory glance. This is purely for if one were to write to an NVMe and not a standard SD card testcase, right?17:05
waveformEickmeyer, basically duplicates the test cases so on Pi 4 we're testing SD and USB boot, and on Pi 5, SD, USB, and NVMe17:06
Eickmeyerwaveform: Oh! Well then, that makes perfects sense. I'll give it an ack and merge and let the other guys handle moving it to the ISO tracker. :)17:07
waveformthanks!17:07
waveform(hopefully I've picked reasonable IDs for the new cases)17:08
Eickmeyerwaveform: I mean, it looks like it. The merge was clean.17:10
EickmeyerThe IDs are usually just in sequential order in the order they were added.17:11
EickmeyerNothing *too* scientific there.17:11
Eickmeyerwaveform: I also like that you cleaned the model numbers to the specs since model numbers mean nothing to some people. :)17:14
ahasenackLocutusOfBorg: around?17:16
waveformheh, there's not many that can look at 902120 and say "that's a Pi Zero 2!"17:16
waveformon the IDs though, I suspect the numeric part at least is a key in some database somewhere, in which case it'd depend on someone adding new entries in the right order to match the new files I've defined17:17
Eickmeyerwaveform: It's not hard as long as the filename doesn't change. ;)17:18
EickmeyerI've done it, but I only have access to Ubuntu Studio and Edubuntu. For Edubuntu Rpi I just re-use the Pi ones you made. :)17:20
waveformheh, no worries -- incidentally, did you use the XML generator or just copy the outputs? (the XML generator may make it easier to customize things en masse)17:20
EickmeyerNah, just point it to the ID. The ISO tracker automatically fills in the rest.17:21
EickmeyerIn other words, it's the exact same test case, because there's no reason to change it.17:22
waveformah, that's a side of the tracker I've not seen yet (I just write test cases, and perform tests so far)17:23
EickmeyerYep. Where there's a capitalized variable (FAMILY, for instance), the tracker will swap that out with the flavor.17:24
waveformah, I wrote a little XML tool (tools/test_case_gen) that I use to handle the mass of duplication in the tests (e.g. I define the bluetooth test case once, then "include" it in the models of Pi that have a bluetooth module, likewise for wifi, ethernet, etc.)17:29
LocutusOfBorgahasenack, as you wish17:35
LocutusOfBorgthey are both equivalent17:35
ahasenackLocutusOfBorg: I'm just adding a comment to the bug,17:36
ahasenackLocutusOfBorg: done17:36
LocutusOfBorgahasenack, the patch is coming from #vbox-dev irc channel17:37
ahasenackis it committed?17:37
ahasenackwas anything else needed? I saw the same code further down17:37
ahasenackwhy isn't there a 7.0.16.1 out yet, given the criticality?17:38
ahasenack(not doubting you, and thanks for jumping on it, I'm just being careful with the details)17:38
ahasenackand since I don't see further movement upstream (or I just didn't find the bug), my worry is if this patch is enough/correct at this moment, or if further analysis is happening17:40
LocutusOfBorgit is correct17:40
LocutusOfBorgbut they don't release .1 updates, they jump to 7.0.18 and this takes time17:40
LocutusOfBorgso they advice to not update for now, and in some days they will release it17:40
ahasenackis there an upstream bug?17:41
LocutusOfBorgnot sure sorr17:41
ahasenackI found some, but the host is windows17:41
ahasenacklike https://www.virtualbox.org/ticket/2204517:41
-ubottu:#ubuntu-devel- www.virtualbox.org ticket 22045 in host support "VirtualBox 7.0.16 crashing host OS immediately with a BSOD by VBoxNetLwf.sys" [New]17:41
ahasenackfound a commit: https://www.virtualbox.org/changeset/104355/vbox17:42
ahasenackbugref:1058517:43
ahasenacklet's see17:43
LocutusOfBorghttps://www.virtualbox.org/ticket/2205217:43
-ubottu:#ubuntu-devel- www.virtualbox.org ticket 22052 in other "'VM guest crashed (Host: Windows 10, Guest: Debian GNU/Linux) due to '' rtR3WinUnhandledXcptFilter caught an exception'''" [New]17:43
ahasenackhttps://www.virtualbox.org/ticket/10585 doesn't seem to be it17:43
-ubottu:#ubuntu-devel- www.virtualbox.org ticket 10585 in other "The SIGALRM never fires" [Closed]17:43
ahasenackthat's on 7.0.14, not .1617:44
LocutusOfBorgsorry I already found the commit, but the paste failed17:44
ahasenackis there a git mirror by any chance?17:44
ahasenacksvn, ugh17:44
LocutusOfBorgahasenack, it is a tricky issue, it is due to USB  ASAN being enabled inside kernel17:44
LocutusOfBorgUBSAN whatever17:45
ahasenack"Author:17:45
ahasenackvboxsync" <-- I suppose they have an internal repository, closed up, and the external one just syncs from it17:45
LocutusOfBorgso, the bug was there since long time, but recent kernels made serious17:45
LocutusOfBorgyes they have17:45
ahasenackthe bugref in https://www.virtualbox.org/changeset/104355/vbox might also be pointing at an internal bug db?17:46
ahasenack"SrvIntNetR0: Fixed regression from r162583 that caused allocation of one segment instead of two ​bugref:10585"17:46
LocutusOfBorgyes correct17:47
ahasenackeven the mentioned changeset (r162583) is not available17:47
ahasenackclassic oracle, I think they do the same with mysql17:47
LocutusOfBorgalso correct, they map internal issues to external public ones17:47
LocutusOfBorg:/17:47
LocutusOfBorgand they sync from time to time17:47
ahasenackand no chatter about this in vbox-dev17:47
LocutusOfBorgthis way we can be "open source" without leveraging the internal customer defects17:48
ahasenackhttps://www.virtualbox.org/pipermail/vbox-dev/2024-April/thread.html april all messages17:48
ahasenackwould it be worth it sending an email there asking about details for this crash? Bug, if the patch from that commit is the only fix needed, etc?17:48
ahasenackalso ask why they have that big fat warning in the front page, but not in the 7.0.16 changelog, and haven't even pulled 7.0.16 from what we can tell17:49
ahasenackis that a crash in the kernel module loaded on the host? It actually crashes the host kernel and you have to reboot?17:49
ewluHi I'm looking to package a version of binutils with a different tuple name (i.e. riscv64-something_here-linux-gnu-ar) which also wouldn't override the existing binutils installation. I'm new to creating packages and I've been struggling to change the installation tuple name for a while. Can I get some help?18:25
ewluI've modified the debian/rules to target the new tuple name, messed with the CROSS_GNU_TYPE (since I see that's what is used in the install step), tried messing with the package name in debian/control. Is there somewhere else I should be modifying?18:27
Eickmeyerwaveform: Quick question: Is there a way we can get Edubuntu 24.04 added to `rpi-imager`?18:31
waveformEickmeyer, yes -- there's a few questions to answer though. Specifically: do you want a separate Edubuntu section? (we need to request that from rpi upstream, but it's probably preferable to having yet-another-image under the already packaged Ubuntu section), where should be icon be hosted? (ideally this should be a unique URL specifically for the imager, and the icon needs to be a 48x48px PNG or SVG), and finally how do you want to control18:34
waveformthe JSON? (I've got a script for that too, if you want :)18:34
waveform*the icon even18:34
Eickmeyerwaveform: Heh, the icon is easy; I can make that happen pretty quick. A separate Edubuntu section can be done, and JSONs are easy. :)18:36
waveformfor reference, https://git.launchpad.net/meta-release/tree/refresh_os_list?h=main is the script we use to maintain the JSON for the Ubuntu images (basic usage "point the url: field at the image and let it generate the rest")18:37
waveformand https://git.launchpad.net/meta-release/tree/raspi/os_list_imagingutility_ubuntu.json?h=main is what the JSON looks like18:38
Eickmeyerwaveform: Well, since we only have one, this should be easy.18:42
Eickmeyerwaveform: Think this is suitable? https://git.launchpad.net/~edubuntu-dev/+git/edubuntu-assets/plain/LogoAssets/edubuntu-rpi-imgr.png18:43
EickmeyerHmmm.. Actually, I have a better idea.18:44
waveformEickmeyer, best way to find out is to test it for yourself; generate the JSON with the icon URL as you'd like then serve your JSON with a local web-server (python3 -m http.server is fine), and run "rpi-imager --repo http://localhost:8000/os_list_imagingutility_edubuntu.json" for example -- you won't see the root menu this way but you should see how your entries will appear otherwise18:45
Eickmeyerhttps://www.edubuntu.org/assets/edubuntu-rpi-imgr.png18:46
waveformyeah, that icon certainly looks fine to me -- but I'd definitely try it out in the imager to check it looks okay there18:47
EickmeyerYep. I'm 100% confident in that one, it's a github host.18:47
jbichahttps://merges.ubuntu.com/main.html looks down, wish I could cheer it up19:14
vorlonjbicha: needs one of bdmurray, sil2100, or doko to fix it; all of them are currently off19:40
tarzeauwhen does debian archive sync start again?20:16
hallynhm, having a case, in a rather convoluted noble based build script, where dracut module install is failing on btrfs when it wants to do 'printf "%s\n" "$(cmdline)" > "${initdir}/etc/cmdline.d/00-btrfs.conf' , saying no such file.20:35
hallynif i edit /usr/lib/dracut/modules.d/90btrfs/module-setup.sh to create that dir on the line before,it passes.20:35
hallynseems very odd, didn't see anything in debian/changelog to explain that20:35
ahasenackdbungert: hi, around?21:41
dbungertahasenack: hey there21:41
ahasenackdbungert: I'm looking at the pam sru21:42
ahasenackdbungert: userdb is fine21:42
ahasenackdbungert: the other one about docs, I did a rebuild in noble, but I'm not getting the extra docs21:42
ahasenackdbungert: did you try a build in a ppa or something, and it worked?21:42
ahasenackdbungert: I'm also failing to debug why my local build doesn't have the docs21:42
ahasenackI do get "checking for DocBook XSL Stylesheets in XML catalog... found", which I can see in the noble archive build was "not found"21:43
dbungertahasenack: hmm, I thought I had, but I think I'm confusing with the oracular upload21:44
dbungertahasenack: yea, the 7ubuntu1 upload is happy and 5ubuntu5.1 is not, I'll investigate why.  Sorry about that.21:49
ahasenackok, I'll leave a comment in the bug21:49
dbungertahasenack: please reject pam 1.5.3-5ubuntu5.122:25
dbungertvorlon: would you reject pam 1.5.3-5ubuntu5.1 from noble?  I need to reupload.22:48
vorlondbungert: ack23:06

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!