16:59:24 <jcajka> #startmeeting Go SIG meeting
16:59:24 <zodbot> Meeting started Mon Aug 31 16:59:24 2020 UTC.
16:59:24 <zodbot> This meeting is logged and archived in a public location.
16:59:24 <zodbot> The chair is jcajka. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:59:24 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:59:24 <zodbot> The meeting name has been set to 'go_sig_meeting'
16:59:31 * Eighth_Doctor waves
16:59:37 <jcajka> #topic Roll Call
16:59:55 <jcajka> oh... 30s early
17:01:05 <jcajka> #chair alexsaezm
17:01:05 <zodbot> Current chairs: alexsaezm jcajka
17:01:14 <jcajka> #chair Eighth_Doctor
17:01:14 <zodbot> Current chairs: Eighth_Doctor alexsaezm jcajka
17:02:12 <Eighth_Doctor> .hello ngompa
17:02:13 <zodbot> Eighth_Doctor: ngompa 'Neal Gompa' <ngompa13@gmail.com>
17:02:33 <eclipseo> .hello2
17:02:34 <zodbot> eclipseo: eclipseo 'Robert-André Mauchin' <zebob.m@gmail.com>
17:03:20 <jcajka> #chair eclipseo
17:03:20 <zodbot> Current chairs: Eighth_Doctor alexsaezm eclipseo jcajka
17:03:36 <jcajka> .hello2
17:03:37 <zodbot> jcajka: jcajka 'None' <jcajka@redhat.com>
17:03:50 <jcajka> still none :)
17:04:08 <nim> .hello2
17:04:09 <zodbot> nim: nim 'None' <nicolas.mailhot@laposte.net>
17:04:19 <alexsaezm> .hello2
17:04:20 <zodbot> alexsaezm: alexsaezm 'None' <asm@redhat.com>
17:04:35 <jcajka> #chair nim
17:04:35 <zodbot> Current chairs: Eighth_Doctor alexsaezm eclipseo jcajka nim
17:05:49 <jcajka> there are no tagged issues so moving to the open floor
17:05:56 <jcajka> #topic Open Floor
17:06:51 <eclipseo> still waiting on https://pagure.io/packaging-committee/issue/1007 to be approved
17:07:24 <eclipseo> no update until it's worked out
17:08:32 <eclipseo> otherwise I'm trying to write a Python script to compare commit in Fedora package with upstream from pkg.go.dev
17:09:07 <Eighth_Doctor> :)
17:09:08 <eclipseo> they don't have an API yet so I scrape the data with BeautifulSoup4
17:09:14 <Eighth_Doctor> :(
17:09:48 <eclipseo> But I'm kinda shit at coding so if any one wants to review that...
17:10:11 <jcajka> Not much from my side either, I will just mention there are planned new point releases of Go for tomorrow PST by upstream, fixing yet undisclosed security issue(s)
17:12:26 <nim> on my side I'll try to wrap the code removal FESCO requested in the redhat-rpm-config change, then finish the go module packaging implementation
17:12:48 <nim> back to work |)
17:13:24 <eclipseo> ok good
17:13:48 <nim> I saw someone got fed up enough with upstream to code test dep annotation in module files
17:13:58 <nim> not sure if it can be reused or not
17:14:14 <jcajka> eclipseo: I might be missing that in the webui of pkg.go.dev, but this just looks like github "mirror" with bit of data that they get from their GOPROXY(or parsing the code)(imports, imported by), does it have more to it?
17:14:35 <nim> I don’t really want to fork upstream code this way
17:14:49 <eclipseo> AH yeah theystill didn't do anything about that despite our numerous request
17:16:08 <eclipseo> jcajka: no it's that, but it's easier to parse one pkg.go.dev than the multiple forges used in Go
17:16:39 <jcajka> eclipseo: makes sense :)
17:16:42 <eclipseo> I've got Bitbucket, Gitea, Gitlabs
17:17:14 <nim> eclipseo, honestly, at this point I doubt they will add it without a major clash, it’s become a matter of pride to ignore the request
17:17:28 <jcajka> AFAIK it is all git, and then the hg pops up ;)
17:17:38 <jcajka> nim: have you seen the issue that the Caterpillar has hit with the bettercap on s390x, missing dep of devel package
17:18:24 <nim> jcajka, nope :( we have s90x users now ?
17:18:53 <jcajka> nim: seems that macros omit one dependency there
17:19:00 <nim> jcajka, but I've not read all my holliday mail yet
17:19:17 <jcajka> https://koji.fedoraproject.org/koji/buildinfo?buildID=1602962
17:19:23 <jcajka> today on IRC
17:21:29 <nim> jcajka, ok
17:22:20 <jcajka> nim: do you think that there is some workaround, like hard-coding the missing dep?
17:22:29 <jcajka> (in a spec)
17:22:51 <nim> jcajka, it’s always possible to add a dep
17:22:54 <Eighth_Doctor> jcajka: I think at one point I saw an SVN repo in pkg.go.dev :)
17:22:55 <nim> manually
17:23:56 <nim> what’s surprising is that those deps are computed by golist, and it’s the first time golist exhibits arch-specific behviour
17:24:42 <nim> unless it’s a transitive dep, and the problem is in another package
17:25:01 <jcajka> Eighth_Doctor: yeah, git, hg and svn....
17:25:02 <nim> that was updated in all arches except s390x
17:25:37 <eclipseo> nim not the first time I have used workaround before
17:26:04 <Eighth_Doctor> jcajka: I know carlwgeorge and I encountered a bzr repo dep too
17:26:19 <Eighth_Doctor> this is the downside to making the system so thin :(
17:26:23 <nim> I’d start by comparing the go package list in the s390x buildroot with the ones on other arches to check they match
17:26:53 <eclipseo> rg "different arch" -g "*.spec"  reveals 8 packages have this workaroud
17:27:06 <jcajka> nim: they should be same as in the missing package is pulled in the BR correctly, it is weird, from brief look I have even seen anything behind build tags, but it has been rather brief look, macros are not backwards compatible again... local prep is pain again
17:27:14 <nim> there has been a lot of flackiness in the s390x builders this summer, since the dc move
17:27:15 <eclipseo> golang-x-crypto for example
17:27:48 <jcajka> grrr... I have even seen -> I haven't seen
17:27:49 <nim> so maybe something is not up to date in s390x go stack
17:28:34 <nim> hopefuly this will all go away in module mode since we'll be using the upstream dep graph directly
17:30:00 <nim> so basically, if you want to find the root cause, start by comparing the buildroot between arches, if you want a quick fix, just add the dep manually
17:30:20 <jcajka> nim: BR looks ok
17:30:58 <nim> jcajka, sorry I don't have any other idea out of hand
17:31:28 <jcajka> what part of the code of the macros is responsible for the deps generation(for devel package), do you have link?
17:34:33 <nim> jcajka, the entry point is https://pagure.io/go-rpm-macros/blob/master/f/rpm/fileattrs/go.attr
17:34:54 <jcajka> nim: thanks
17:34:56 <nim> jcajka, but as you see it only calls go-rpm-integration with some flags
17:35:17 <jcajka> go-rpm-integration lives where?
17:36:13 <nim> here https://pagure.io/go-rpm-macros/blob/master/f/bin/go-rpm-integration#_423
17:36:32 <nim> and there you see requires are just a wrapper around golist
17:37:47 <jcajka> ack, thanks again
17:37:58 <nim> the next version removes the shell wrappering layer it’s easier to follow the code flow
17:39:07 <nim> but basically, in this version after you remove the wrappers it’s a direct golist call with some reformating of the output but no dep computation outside golist
17:41:35 <nim> sorry, without even reformatting now that golist has a template option
17:42:23 <jcajka> ok, thanks, sorry for hijacking the meeting, are there any other topics?
17:43:09 <alexsaezm> Nothing relevant from my side: Pushed today 1.13.15 and 1.14.7
17:43:26 <alexsaezm> (for F31 and F32)
17:45:04 <jcajka> alexsaezm: thanks
17:46:21 <eclipseo> no more topic from me
17:46:48 <jcajka> last call be it :)
17:47:04 <nim> nothing else here
17:48:25 <jcajka> #endmeeting