With my open source library hat on, CI should be the canary in the coalmine for my users, not a perfectly reproducible thing that proves it can be installed on an old version my users aren't using. I think for public stuff the redirect is exactly the right thing. I also "stack init" my projects on the CI to test with the latest versions, which is even worse for breaking consistency. I combine that with CRON jobs that kick off all my CI things nightly, so I am explicitly testing for the passage of time.
With my commercial developer hat on, there's no way I'd have redirected to an arbitrary stack! A very valid strategy for the right use case. Thanks for the binaries! On Thu, Dec 7, 2017 at 2:17 PM, Emanuel Borsboom <ma...@fpcomplete.com> wrote: > Ha, my unintelligible English here is probably due to staying up way too > late getting stack-1.6.1 out :) Let me try that again... > > It might also be good if our CI examples recommend the best practice of > specifying an exact Stack version by pointing directly to a specific Github > release rather than using the convenience redirects. What happened here is > one example of breakage if CI scripts don't specify a version; another would > be that a behaviour change in Stack breaks CI. > > Also, more importantly, the Windows 32-bit binaries are up now! > > > On Dec 7, 2017, at 6:13 AM, Emanuel Borsboom <ma...@fpcomplete.com> wrote: > > It might also be good if CI examples we recommend the best practice of > specifying an exact Stack version by pointing directly to a specific Github > release rather than using the convenience redirects. What happened here is > one example of breakage if CI scripts specify stick to a specific version; > another would be that a behaviour change in Stack breaks CI. > > On Dec 7, 2017, at 5:34 AM, Michael Snoyman <michael.snoy...@gmail.com> > wrote: > > We should probably change the way that stackage.org finds binaries. All of > the old binaries are present on Github still, it's just the lookup that > isn't working, since it only considered the most recent release. > > Yes, this is an ugly situation, we're still trying to play catch-up. > > On Thu, Dec 7, 2017 at 2:49 PM, Neil Mitchell <ndmitch...@gmail.com> wrote: >> >> Can I suggest that in future you leave the old binaries around until >> the new ones are available? My Windows CI automatically hits that URL, >> so it broke :( >> >> On Thu, Dec 7, 2017 at 12:28 PM, Michael Snoyman >> <michael.snoy...@gmail.com> wrote: >> > I believe there are plans to release it still. As the release >> > announcement >> > mentioned, we didn't complete the normal release procedure to rush the >> > fix >> > for GHC 8.2, and some binaries are still being compiled. >> > >> > On Thu, Dec 7, 2017 at 2:13 PM, Neil Mitchell <ndmitch...@gmail.com> >> > wrote: >> >> >> >> Hi, >> >> >> >> The page https://www.stackage.org/stack/ links to >> >> https://www.stackage.org/stack/windows-i386, which now gives a 404. >> >> Where did the 32bit binary disappear to? >> >> >> >> Thanks, Neil >> >> >> >> -- >> >> You received this message because you are subscribed to the Google >> >> Groups >> >> "haskell-stack" group. >> >> To unsubscribe from this group and stop receiving emails from it, send >> >> an >> >> email to haskell-stack+unsubscr...@googlegroups.com. >> >> To post to this group, send email to haskell-stack@googlegroups.com. >> >> To view this discussion on the web visit >> >> >> >> https://groups.google.com/d/msgid/haskell-stack/CAKcFpmJY27nB%3D%2BCiS0_6UZ56ePb-DH_0H7c-0VdommRmWqdD%3Dg%40mail.gmail.com. >> >> For more options, visit https://groups.google.com/d/optout. >> > >> > > > > > -- > You received this message because you are subscribed to the Google Groups > "haskell-stack" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to haskell-stack+unsubscr...@googlegroups.com. > To post to this group, send email to haskell-stack@googlegroups.com. > To view this discussion on the web visit > https://groups.google.com/d/msgid/haskell-stack/CAKA2Jg%2BzQRVzCGPgSp%3D%2BWRSoREyzZHgNSyCqozODLgtgUpK7vQ%40mail.gmail.com. > For more options, visit https://groups.google.com/d/optout. > > > -- You received this message because you are subscribed to the Google Groups "haskell-stack" group. To unsubscribe from this group and stop receiving emails from it, send an email to haskell-stack+unsubscr...@googlegroups.com. To post to this group, send email to haskell-stack@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/haskell-stack/CAKcFpmKhpd2LS5at90uwjDHvTDNAnAzmPBKt4jTVZnu7ED8ZVA%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.