18:00:13 <alexsaezm> #startmeeting Go SIG meeting 18:00:13 <zodbot> Meeting started Mon Jul 31 18:00:13 2023 UTC. 18:00:13 <zodbot> This meeting is logged and archived in a public location. 18:00:13 <zodbot> The chair is alexsaezm. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 18:00:13 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:00:13 <zodbot> The meeting name has been set to 'go_sig_meeting' 18:00:35 <alexsaezm> #topic Roll Call 18:00:41 <alexsaezm> Welcome everyone! 18:00:53 <alexsaezm> it's seems the bridge works this time (or at least I can see bot messages in matrix) 18:01:16 <alexsaezm> As usual, we can wait a little for everyone :) 18:03:45 <eclipseo> .hello2 18:03:46 <zodbot> eclipseo: eclipseo 'Robert-André Mauchin' <zebob.m@gmail.com> 18:04:14 <eclipseo> Hi people 😀 18:04:16 <MavJS> .hello 18:04:16 <zodbot> MavJS: (hello <an alias, 1 argument>) -- Alias for "hellomynameis $1". 18:04:22 <MavJS> .hello2 18:04:23 <zodbot> MavJS: mavjs 'Maverick Kaung' <mavjs01@gmail.com> 18:04:27 <MavJS> hello! :) 18:05:11 <alexsaezm> o/ 18:06:36 <alexsaezm> We have two issues tagged, for the meeting, the same ones as the last meeting. docker and ix86. Does anyone has a preference? (I don't have anything for those two) 18:07:49 <eclipseo> not much to ay about each 18:08:44 <alexsaezm> In that case, unless anyone else has something to add (we can wait) we can jump into the open floor 18:08:49 <MavJS> me neither :> 18:09:20 <eclipseo> ok for me 18:09:50 <alexsaezm> #topic Open floor 18:10:37 <alexsaezm> Go 1.21rc3 is on rawhide and seems that it made several package fail. Apart from that I have a "nice" list of FTBFS, so nothing fancy to add 18:11:01 <alexsaezm> (Tomorrow there will be a security go release apparently) 18:11:04 <eclipseo> could we put a lits of all Go FTBFS 18:11:25 <eclipseo> i, working on updating acme rignt now mut there's 50 new deps 18:11:42 <eclipseo> s/lits/list/ 18:11:48 <mikelo_m[m]> .hello mikelo2 18:11:49 <zodbot> mikelo_m[m]: mikelo2 'Mikel Olasagasti Uranga' <mikel@olasagasti.info> 18:12:06 <eclipseo> s/,/'m/, s/rignt/right/, s/mut/but/ 18:13:11 <mikelo_m[m]> eclipseo: I’ve that query, give 2mins 18:13:57 <alexsaezm> The F39 failures -> https://kojipkgs.fedoraproject.org/mass-rebuild/f39-failures.html 18:14:42 <eclipseo> thx 18:14:54 <mikelo_m[m]> https://bugzilla.redhat.com/buglist.cgi?bug_status=NEW&email1=go-sig%40lists.fedoraproject.org&emailassigned_to1=1&emailcc1=1&emailtype1=equals&list_id=12392650&query_format=advanced&short_desc=FTBFS&short_desc_type=allwordssubstr 18:15:11 <alexsaezm> that is way better ^ 18:15:50 <mikelo_m[m]> one of the packages failing with 1.21 is golang-github-pkg-errors, unmaintained upstream, that is used by quite a few packages 18:17:22 <mikelo_m[m]> golang-github-onsi-ginkgo-2 was affected also and upstream already fixed after reporting 18:17:44 <mikelo_m[m]> and a test with azure sdk that merged my PR upstream 18:18:49 <mikelo_m[m]> but... many of the ftbfs are previous to go-1.21 18:19:43 <alexsaezm> For this week I'm planning on going through the ones I own and fix them no matter if they are long FTBFS or not, because I have 109 bugs open :D 18:20:14 <mikelo_m[m]> newbie... 149 in my case :rofl: 18:20:18 <alexsaezm> hahah 18:20:26 <alexsaezm> wanna some more? 18:20:58 <mikelo_m[m]> I've been cleaning my plate the last few days, I'm close to 30 closed BZ iirc 18:21:25 <alexsaezm> O: 18:21:54 <eclipseo> don't have that much time to close that much! I'll look into it this wee-end more 18:22:36 <eclipseo> i'm looking at golang-github-pkg-errors forks but there is not much more activity 18:24:48 <alexsaezm> mikelo_m: do you have a build.log of golang-github-pkg-errors ? just to see if I can patch it 18:25:00 <alexsaezm> at least for this time 18:25:12 <mikelo_m[m]> https://koschei.fedoraproject.org/package/golang-github-pkg-errors 18:25:28 * alexsaezm always forgets about koschei when he needs it 18:26:06 <eclipseo> --- FAIL: TestStackTrace (0.00s)... (full message at <https://libera.ems.host/_matrix/media/v3/download/libera.chat/03ae0c18ad65690f6847ee2c98ab91766d2df4e2>) 18:26:57 <mikelo_m[m]> this list is not acurate but worth checking https://koschei.fedoraproject.org/affected-by/golang-bin?epoch1=0&version1=1.20.5&release1=1.fc39&epoch2=0&version2=1.21~rc3&release2=2.fc39&collection=f39 18:27:01 <alexsaezm> I can try to patch it, but it's not a solution of course 18:27:28 <mikelo_m[m]> the problem with that list is that many of the packages were in ftbfs mode already 18:28:20 <alexsaezm> podman is there?! 18:30:24 <alexsaezm> if they appear with the green check it means that they are not failing anymore but they did right? 18:30:29 <mikelo_m[m]> https://koschei.fedoraproject.org/package/podman yes 18:30:51 <alexsaezm> I don't understand why some of them are duplicated 18:31:00 <alexsaezm> tinygo and delve appear multiple times 18:31:19 <mikelo_m[m]> multiple times, where? 18:32:13 <alexsaezm> https://koschei.fedoraproject.org/affected-by/golang-bin?epoch1=0&version1=1.20.5&release1=1.fc39&epoch2=0&version2=1.21~rc3&release2=2.fc39&collection=f39 18:32:14 <alexsaezm> there 18:32:20 <alexsaezm> I see delve more than once 18:32:42 <alexsaezm> podman too 18:33:54 <mikelo_m[m]> I guess that it's because koschei keeps building packages and it may record how many times has been built against golang-1.21rc3 18:35:16 <daMaestro> hmm, yeah podman build failing on gvisor-tap-vsock it seems 18:36:17 <mikelo_m[m]> gvisor package is +18 months old, there may be some issue there 18:36:59 <daMaestro> and i love that it's gohacks that seems to be causing it "Package gohacks contains utilities for subverting the Go compiler." (in my naive reading of the build log output from go) 18:41:42 <alexsaezm> so I guess the plan now is to keep fixing this things :) 18:41:55 <alexsaezm> does anyone has something to add, or we can call it? 18:42:09 <MavJS> I have 1 thing, regarding my email on the golang list 18:42:30 <eclipseo> podman seems to be fixed in 4.6.0 18:42:40 <MavJS> my PR for gobuster was merged by Fab however a build was not submitted: https://src.fedoraproject.org/rpms/gobuster/pull-request/5 (could someone from go-sig submit a build?) 18:42:53 <MavJS> the other one is still waiting for a merge: https://src.fedoraproject.org/rpms/ffuf/pull-request/1 18:43:05 <alexsaezm> sure 18:43:08 <alexsaezm> let me do it 18:43:48 <MavJS> if that was for me: thank you :) 18:43:53 <alexsaezm> wait... it should be built 18:43:57 <alexsaezm> let me check... 18:44:05 <eclipseo> it won't build 18:44:12 <eclipseo> + %global goipath github.com/OJ/gobuster/v3 18:44:12 <eclipseo> + Version: 3.5.0 18:44:21 <alexsaezm> https://koji.fedoraproject.org/koji/buildinfo?buildID=2239402 18:44:30 <eclipseo> you chnaged the import patch 18:44:37 <eclipseo> s/patch/path/ 18:45:04 <eclipseo> ah its a binary ok 18:45:17 <alexsaezm> yeah, and it seems it was build during the mass rebuild 18:45:23 <alexsaezm> so rawhide has it already 18:45:39 <MavJS> ah, okay, since i didn't see a "build" in the PR like the rest, I thought it was 18:45:43 <eclipseo> do you want it on older path? 18:46:11 <MavJS> personally no preference 18:46:20 <MavJS> it's just a go binary tool 18:46:29 <eclipseo> s/path/branch/ 18:46:47 <MavJS> ah, there was one person wishing for it 18:47:42 <mikelo_m[m]> MavJS: I suggets you to write an email to fab if you want to be the co-maintainer of those two packages 18:47:47 <mikelo_m[m]> s/suggets/sugget/ 18:48:11 <MavJS> mikelo_m[m]: i did that today 18:48:31 <mikelo_m[m]> awesome 18:48:53 <MavJS> so i guess i still have to wait on ffuf merge from fab then ? :S 18:49:43 <eclipseo> well I can build the ther branches with PP 18:50:02 <MavJS> eclipseo: i didn't answer your question properly, if you could please build for F38 minimum for gobuster 18:50:38 <MavJS> i think current+rawhide branches are good enough, imho 18:50:51 <mikelo_m[m]> MavJS: PR is correct, but can you wait for a few more days for fab to answer your email? if not ping me and I'll do the PR + build 18:51:01 <MavJS> mikelo_m[m]: roger. will do. appreciate it. 18:52:57 <alexsaezm> anything else before we call it? 18:53:14 <MavJS> also, does anyone have a script/tool that can walk the dependencies to see what needs to be packaged? 18:53:21 <mikelo_m[m]> if someone uses restic, I'll appreciate feedback on 0.15.2 update https://bodhi.fedoraproject.org/updates/FEDORA-2023-6cd7536ba3 18:53:53 <MavJS> mikelo_m[m]: i use restic.. will take a look. 18:54:01 <mikelo_m[m]> MavJS: not me, but I also feel the need of something that parses go.mod and compares against Fedora versions. 18:54:40 <alexsaezm> so a tool that reads go.mod, checks the nvr there and check the packages in fedora? 18:54:45 <alexsaezm> hmmm that sounds interesting 18:54:53 <MavJS> i started working on that using https://pkg.go.dev/golang.org/x/mod/modfile#File 18:55:17 <mikelo_m[m]> some packages have +20 deps and checking versions manually can be tedious 18:55:20 <MavJS> not perfect.. and looking at the go2rpm parsing for name stuff 18:55:31 <MavJS> like removing go- from package name etc 18:55:32 <mikelo_m[m]> I usually check go.mod diffs between packages version and new version 18:56:00 <mikelo_m[m]> but for new packages or large updates can be handy 18:56:02 <MavJS> yeah, that's easy to do if you have around 10x first-level deps :> 18:56:18 <MavJS> looking at some "security" tools i gave up :p 18:56:31 <mikelo_m[m]> you usually only need the first level deps to build the package 18:56:37 <MavJS> i see 18:56:43 <MavJS> i'm still learning this stuff :) 18:56:47 <mikelo_m[m]> the problem is that we require everything in the chain because... Fedora I guess 18:56:56 <MavJS> heh :) 18:57:14 <mikelo_m[m]> when you build a package in Fedora it may require 700 packages, but if you just go with git clone && go build, it will take only those in the go.mod 18:57:26 <mikelo_m[m]> our problem is that we chain BuildRequires 18:57:47 <mikelo_m[m]> so if you have a chain of 7, you need to pull the deps of all the packages in the chain 18:57:50 <MavJS> do you folks also use copr to get the packages build into a project so you can build multiple packages ? 18:58:02 <alexsaezm> yeah 18:58:09 <mikelo_m[m]> I like to use copr for new packages and to test package updates 18:58:19 <MavJS> okay, i'll give that a go 18:58:33 <mikelo_m[m]> using copr cli can be useful to test new packages and it's deps easily 18:58:42 <MavJS> i might send some questions here or on the mailing list 18:58:47 <MavJS> thank you for all your answers :) 18:58:48 <mikelo_m[m]> iirc alexsaezm wrote a script based on some comments I gave him 18:59:07 <alexsaezm> https://github.com/alexsaezm/dotfiles/blob/main/scripts/.local/scripts/fedora-build-on-copr 18:59:10 <MavJS> <3 18:59:19 <mikelo_m[m]> alexsaezm++ 18:59:25 <alexsaezm> might not work as expected, I usually change a lot my scripts 18:59:32 <alexsaezm> but feel free to copy it and use it 18:59:44 <MavJS> appreciate the pointers :) 19:00:10 <mikelo_m[m]> the trick is to use PRs 19:00:15 <mikelo_m[m]> ouch, not PRs, but branches 19:01:07 <mikelo_m[m]> so fedpkg fork then git push to your fork, in a branch if possible, and then use that branch to be built by copr and then all the packages that require the package to be built against the update in copr 19:01:18 <mikelo_m[m]> super useful to ensure and update doesn't break things 19:03:57 <alexsaezm> yeah, if I recall correctly the script I pasted doesn't do that exactly, just try to build the currect branch on copr (f38 for example). If later I want to use my personal branch, I usually use the fepkg copr-build 19:04:15 <mikelo_m[m]> alexsaezm: what do you think about asking a Fedora Badge for the go-sig team? 19:04:26 <MavJS> that works for packages already in fedora, but for most what i see for new packages i was playing around with.. dependencies and its dependencies are not in fedora :> 19:04:46 <mikelo_m[m]> TIL fedpkg copr-build 19:05:01 <mikelo_m[m]> I usually use fedpkg scratch-build 19:05:22 <mikelo_m[m]> MavJS: for that I use a github repo and copr 19:05:44 <MavJS> mikelo_m[m]: ok, do you have an example repo? 19:05:46 <alexsaezm> I do too, but fedpkg copr-build user/project is really fast to use with already made dependencies, that's why I didn't include that functionality in the script 19:06:23 <alexsaezm> mikelo_m[m]: the folks that design the badges do an amazing work, it would be lovely to have a gopher with a fedora or something like that 19:06:28 <mikelo_m[m]> https://github.com/mikelolasagasti/rclone-specs plus https://copr.fedorainfracloud.org/coprs/mikelo2/rclone/ 19:06:38 <MavJS> sweet :) 19:07:09 <mikelo_m[m]> for packages that are in Fedora but you just want to update them, you can create a fork + branch and use copr's dist-git option 19:07:38 <mikelo_m[m]> for new packages, I've a hack-script that creates the output required for fedora-review 19:07:53 <MavJS> i see 19:08:13 <eclipseo> do you have a thing to rebuild all packages affected by a chane? 19:08:30 <mikelo_m[m]> yes 19:09:24 <mikelo_m[m]> https://mikel.olasagasti.info/tmp/fedora/chain.txt 19:10:25 <mikelo_m[m]> I use those commands for that, I don't have a script, but it's easy to script 19:11:11 <MavJS> neat :) 19:11:35 <alexsaezm> https://github.com/alexsaezm/dotfiles/blob/main/scripts/.local/scripts/fedora-builddependencies 19:11:40 <alexsaezm> I think it had a bug last time I used it 19:14:55 <mikelo_m[m]> #action Alex to check what is required for a go-sig Fedora Badge 19:16:17 <MavJS> did i miss some messages between "I think it had a bug last time I used it" and action for alex ? 19:16:31 <mikelo_m[m]> no 19:16:53 <MavJS> ah okay :) 19:16:56 <alexsaezm> I was referring to the github link I pasted, it's the same idea mikelo_m pasted, but in a script 19:17:39 <MavJS> yeah, that i got, but there was like 3 minutes of no messages, so i thought maybe the bridge died a bit :) 19:18:33 <mikelo_m[m]> any other topic? 19:18:47 <MavJS> not from me. thank you all for answering my questions patiently :) 19:18:47 <alexsaezm> not from my side 19:20:24 <eclipseo> sorry ts ate got to go 19:21:06 <eclipseo> nothing for me 19:21:35 <alexsaezm> then we can call it 19:21:39 <alexsaezm> thanks everyone!!! 19:21:49 <alexsaezm> #endmeeting