Commit Graph

4132 Commits

Author SHA1 Message Date
Pietro Gagliardi a0ce3f1793 Try some more things 2019-04-07 10:57:11 -04:00
Pietro Gagliardi 3129cd7b46 OKay, let's use apt-cache to see how the packages here are set up. We'll go up the chain. 2019-04-07 10:19:44 -04:00
Pietro Gagliardi 416f4b4f70 Let's try installing libgtk-3-0:amd64 just for the heck of it 2019-04-07 09:06:49 -04:00
Pietro Gagliardi a564d66304 Oops, missed a spot. Good night for real. 2019-04-07 04:01:27 -04:00
Pietro Gagliardi d8781c961f One more, because I should have slept an hour ago. Not all those packages are lib32-able, and apparently libc6-i386 is a thing. 2019-04-07 03:57:26 -04:00
Pietro Gagliardi 08abc9972c Oh it's lib32 stuff. There used to be a metapackage for this why did they get rid of it 2019-04-07 03:50:07 -04:00
Pietro Gagliardi 16849b0a57 Okay, that had no effect, so follow the chain again: libgl1-mesa-dri:i386 -> libllvm6.0:i386 2019-04-07 03:40:16 -04:00
Pietro Gagliardi 8634048aba Okay, looks like for it to take effect I also need --no-upgrade. 2019-04-07 03:35:01 -04:00
Pietro Gagliardi f6022bdcc0 Okay, we have to run sudo apt-get update without --no-download. Try to force apt-get install to ignore updates. 2019-04-07 03:30:05 -04:00
Pietro Gagliardi e98a50b51c Shot in the dark; I'm starting to think the problem really is just outdated packages. 2019-04-07 03:23:44 -04:00
Pietro Gagliardi 1372d0d128 Follow the chain: libeg1-mesa:i386 -> libgl1-mesa-dri:i386 2019-04-07 03:15:42 -04:00
Pietro Gagliardi 92f0163563 Follow the chain: libwayland-egl1-mesa:i386 -> libegl1-mesa:i386 2019-04-07 03:06:24 -04:00
Pietro Gagliardi 574971c58d Yep that's what I thought; it tried to update the entire system instead :| Let's check the apt-cache policy. 2019-04-07 03:00:16 -04:00
Pietro Gagliardi b2d3d7f29d Okay, some people say apt-get upgrade is also needed. I'm doubtful, but let's try it. 2019-04-07 02:49:34 -04:00
Pietro Gagliardi 9c62325367 What is held 2019-04-07 02:45:37 -04:00
Pietro Gagliardi 2ee12f0bcb Okay we do want GCC 5 after all, and also ugh in general 2019-04-07 02:36:45 -04:00
Pietro Gagliardi e66257a2b5 Okay which version of GCC do we want 2019-04-07 02:33:53 -04:00
Pietro Gagliardi 6b2453ab4d Okay, that wasn't quite right; we don't want GCC 5, and I'm not sure why it doesn't want to install GTK+ 3 itself... 2019-04-07 02:30:47 -04:00
Pietro Gagliardi 6f26d869f2 Oops 2019-04-07 02:24:47 -04:00
Pietro Gagliardi 6e688b6476 Okay, that wasn't quite right... 2019-04-07 02:22:20 -04:00
Pietro Gagliardi 5452f8258f Let's try a Linux 386 build next! NumPy uses Docker; let's see if the Travis CI thing we did works here too instead. 2019-04-07 02:12:35 -04:00
Pietro Gagliardi 60687e6aff So apparently only macosx10.13 is valid?! Just use the default SDK. 2019-04-07 02:02:32 -04:00
Pietro Gagliardi efd7e8d07d Okay, so MinGW-w64 on Azure Pipelines is a non-starter. Let's find out what the macOS SDKs are *actually* called. 2019-04-07 01:57:47 -04:00
Pietro Gagliardi 23a0a041f0 ??? 2019-04-07 01:50:53 -04:00
Pietro Gagliardi f7c1515ae1 Roadblocks roadblocks 2019-04-07 01:43:31 -04:00
Pietro Gagliardi 5ded8398ba Bleh 2019-04-07 01:38:02 -04:00
Pietro Gagliardi 29b9e820da Some more guesswork. I have an idea... 2019-04-07 01:34:20 -04:00
Pietro Gagliardi d994dcb0b0 Hmmm 2019-04-07 01:29:44 -04:00
Pietro Gagliardi 4ec1a642a2 Getting closer... 2019-04-07 01:23:07 -04:00
Pietro Gagliardi 58cdf151a5 Ah, Write-Host was the key 2019-04-07 01:20:09 -04:00
Pietro Gagliardi 98e814f013 Grrr 2019-04-07 01:17:28 -04:00
Pietro Gagliardi e9792a2216 Oops again 2019-04-07 01:13:53 -04:00
Pietro Gagliardi 21591e4f6f Hm, Chocolatey is being annoying... 2019-04-07 01:11:18 -04:00
Pietro Gagliardi 651d502624 That didn't quite work 2019-04-07 00:43:41 -04:00
Pietro Gagliardi d3c553c424 Okay, things are not quite working right (and I forgot about shared builds not being available on MinGW-w64). Set up the next round of debugging early, and debug our PowerShell now. 2019-04-07 00:39:27 -04:00
Pietro Gagliardi a8f643c09c Sigh 2019-04-07 00:31:54 -04:00
Pietro Gagliardi 40100a17d1 Oops 2019-04-07 00:22:19 -04:00
Pietro Gagliardi 1500c750b6 Let's try MinGW-w64! 2019-04-07 00:19:28 -04:00
Pietro Gagliardi 21d4ad54b6 Part 2 of previous change 2019-04-06 23:56:05 -04:00
Pietro Gagliardi 26449f444f Executive decision: only run Ninja builds on CI. I only use these anyway, and this will let CI builds and releases go faster. If something breaks for other people, they will report bugs anyway, as they already have... 2019-04-06 23:53:40 -04:00
Pietro Gagliardi 31adbb419e Move the old Travis file out of the way and get rid of most of it. We still need some of its contents, but we will be getting rid of Travis entirely. 2019-04-06 21:13:16 -04:00
Pietro Gagliardi 6ba124ba83 Oops, let's not clobber Windows artifacts. 2019-04-06 20:58:56 -04:00
Pietro Gagliardi 8d90ae1dc9 And another oops, but I confirmed the archive commands look right, so let's go for real now. 2019-04-06 20:52:38 -04:00
Pietro Gagliardi 08096e13a4 Oops again 2019-04-06 20:45:14 -04:00
Pietro Gagliardi c1a0e2e3b6 Oops 2019-04-06 20:34:45 -04:00
Pietro Gagliardi f49253cbf8 Now try Windows artifacts. 2019-04-06 20:24:41 -04:00
Pietro Gagliardi 7e30a2afc1 Oops 2019-04-06 20:06:25 -04:00
Pietro Gagliardi 2584e82191 Moved the linux/darwn pipeline out of the way so we can start working on the Windows one. Also more TODOs. 2019-04-06 19:33:53 -04:00
Pietro Gagliardi 0c410022bb Oops 2019-04-06 19:28:33 -04:00
Pietro Gagliardi 988ca49e41 Split the Linux artifact rules into a template and try them on macOS and static builds too. 2019-04-06 19:26:35 -04:00