<zaytsev>
fijal: arigato: could you please confirm that we are set for this evening? two of our be developers are interested to join, but they both (like me) don't live in koeln, so it would be important for them to know if it works for you to take care of the logistics
<mattip>
zaytsev: you should be able to find a 14.1 compiler in the Visual Studio 2017 installer.
<mattip>
not just 15, it also installs 14
<mattip>
in any case, even 15 is preferable to the broken 14.0 there is now
<zaytsev>
fijal: arigato: also, i think we'd be on the safe side if i make a reservation @ WIPPN´BK, so would be great if you could tell me whether you'll make it at 19 at im zollhafen, so that i can reserve say from 19:30
<zaytsev>
mattip: it's not on the main page there, so i have no idea how to i find 2015 in 2017 installer. do you have a screenshot?
<zaytsev>
mattip: but i have installed 2017 long ago and it is supposed to be working. i have installed it using the installer. should i also manually set som VCVARS or something???
<zaytsev>
mattip: i can also upgrade 2017 to the latest version, this is the only thing that 2017 installer is offering me right now
<mattip>
look for "Visual Studio Build Tools 2017" inside the 2017 installer, there you can choose 15.4
<mattip>
the search order is 14.0, 14.1, 15
<mattip>
since pypy finds a broken 14.0, it never tries 15
<zaytsev>
mattip: it only offers me 15.9.4.0. no 15.4, no 14.1
<mattip>
removing the 14.0 will help
<zaytsev>
mattip: okay so where does 14.0 comes from?is this microsoft nonsense for python that i installed for you long ago? if yes, i assum you no longer need it and i should remove it and unset 14 variables?
<mattip>
+1
<mattip>
in my visual studio installer, I click on the "more" next to the "Visual Studio Build Tools 2017"
<mattip>
then modify
<mattip>
then I get a new windows with Visual C++ Build Tools
<mattip>
in a list on the right
forgottenone has quit [Ping timeout: 258 seconds]
<mattip>
there I can choose the VC++ 14.1 compiler
<mattip>
but maybe just remove the 14 one and kick off a build, the 15 compiler should work
<njs>
btw, azure pipelines has free 10x parallel windows machines maintained by "someone else", and supports jobs up to 6 hours long
<njs>
I think it might be impossible to hook up to bitbucket though
<mattip>
correct, github only
<zaytsev>
mattip: aha, i think i found it in the list of individual components
<zaytsev>
mattip: so 14.0 came from python tools
<mattip>
yeah, they refactored that all away, now you have to use the visual studio installer apparently
<zaytsev>
mattip: 14.1 came hidden with 2017 and on top of that i should have correctly functioning 15. so i should delete tools, and i should delete 14.1, and then maybe it will work
<zaytsev>
mattip: assholes. this is as counterintuitive as it gets, and then you are left trying to find out where did all this shit that you never conciously installed breaking your build comes from
<mattip>
although you do it in stages: delete tools, kick a build, if that fails delete 14.1, kick a build
<zaytsev>
mattip: alright, i now know what do you want from me. i will start the sutff and then let you know. need to have some breakfast and drive to work now
<njs>
but I suspect "bitbucket cloud" is only git repos
<zaytsev>
thank you
<mattip>
zaytsev: thank you for providing the service
<mattip>
on the one hand, everyone is using git/github so if we want to encourage users, we should too
<mattip>
like cpython
<mattip>
on the other hand, the monoculture is scary
<arigato>
zaytsev: yes, as far as I know we're still set for this evening. around the street called Im Zollhafen, where precisely?
themsay has joined #pypy
<zaytsev>
arigato: im zollhafen, 22 - could you or fijal just pm me your number? i will reply from miine, and then you could just call me so that i come out and pick you up
<arigato>
(re github, I'll strongly oppose moving pypy to git; at the moment we're kind of giving up using git for our vr development too)
<zaytsev>
ok must drive to work now, sorry
<arigato>
zaytsev: ok, fijal should do that
xorAxAx has quit [Read error: Connection reset by peer]
xorAxAx has joined #pypy
Ai9zO5AP has joined #pypy
<mattip>
arigato: agreed about mercurial
<mjacob>
+1 about staying with mercurial. we use hg at work with some new, experimental extensions (evolve and others) and it works so much better than git. also, tortoisehg is a very nice UI
<mattip>
it would be nice to find a better issue tracker/host than bitbucket
lritter_ has quit [Remote host closed the connection]
<mattip>
hmm, reading down from there, it doesn't seem like they will add the support
<mattip>
for mercurial in gitlab
dddddd has joined #pypy
<zaytsev>
mattip: we use bitbucket + git at work, and i'm very unhappy about bitbucket as a service. github is just so much better. we're cheap though and cannot really afford github at the moment, so we're staying with bitbucket
kipras has joined #pypy
<zaytsev>
mattip: in as far as gitlab'; -- drop database is concerned... i'm not so convinced by their professionalism either. at least there haven't been recorded catastrophic data loses by atlassian
<zaytsev>
mattip: maybe you could consider just mirroring pypy bitbucket as a git repository on github, but use bitbucket as master. this way you will be able to use the repo over git / svn and the issue tracker. what's annoying though is that you can't turn off pull requests @ github
<zaytsev>
fijal: could you please pm me your number when you're there and tell me if i should reserve a table at 19:30, should do so sometime soon i suspect
<mattip>
one of the core problems with changing anything is that as developers, we are fine with the minimalistic bitbucket
<mattip>
but I am sure our users are frustrated by the poor issue tracking, and general UX
jkbbwr has quit [Quit: Lost terminal]
<zaytsev>
mattip: oh, ui doesn't frustrate me at all. what frustrates me is that bitbucket is unreliable. we often get 5xx errors and they come and go. then they often have breakages which go unnoticed for half an hour, and then only you get an update on the status page that they are working on it. like in some two month webhooks were broken for several hours a day 5 times, 5 weeks in a row!!!11 and it was exactly the same problem
<zaytsev>
mattip: webhooks not working = our ci builds not triggering = our work is paralyzed. on top of that, bitbucket pipelines just get stuck sometimes, and run for hours, were they should have been automatically terminated, and you also get billed for it. shit like that...
<zaytsev>
mattip: i don't like differntial at all, but maybe that's my thing
xcm has quit [Read error: Connection reset by peer]