| 21 Aug 2021 |
|  sterni changed their display name from sterni[m] to sterni. | 22:21:47 |
| sterni | hm okay, I never tried compiling with useGccNg and crossSystem before | 22:25:21 |
| 22 Aug 2021 |
 John Ericson | sterni: alexfmpe: OK updated the status on the PR | 05:05:08 |
| John Ericson | Now that we are "into the normal libaries" passed libgcc and libc | 05:05:29 |
| John Ericson | * Now that we are "into the normal libaries" past libgcc and libc | 05:05:38 |
| John Ericson | It might be easier for others to work on | 05:05:53 |
 alexfmpe | So I could try adding say libssp? | 14:33:08 |
| John Ericson | yeah | 14:47:58 |
|  siraben changed their display name from siraben (he/him) to siraben. | 16:15:50 |
| John Ericson | alexfmpe: added libssp | 21:01:53 |
| alexfmpe | oh that was fast | 21:02:18 |
| alexfmpe | I was rebasing my mingw ghc PR on top of sterni 's gold rewiring | 21:02:46 |
| alexfmpe | and started adding those changes to non-8.10 ghc | 21:03:07 |
| alexfmpe | but it looks liike mingw ghc 8.8 was already non-building | 21:03:21 |
| alexfmpe | so I can't quite verify I'm not breaking it more | 21:03:30 |
| alexfmpe | I'm now building 8.8 without my changes to confirm the error is unrelated (it sure looks like it https://pastebin.com/7FHK1Q0s)
but if that still fails, and fixing 8.8 isn't simple, what's the usual way to proceed here?
be conservative about touching 8.8, since I might be breaking it more, or be conservative about divergence, and do the same changes that fixed 8.10? | 21:06:16 |
| John Ericson | alexfmpe: well i did it for aarch64 linux | 21:07:48 |
| John Ericson | a.k.a. easy stuff | 21:07:54 |
| John Ericson | mine has the position independent code issues | 21:08:08 |
| John Ericson | which need to be fixed even if one doesn't care about C++ (and they first manifested when building libstdc++) | 21:08:34 |
| alexfmpe | heh, you're getting hti by PIC a lot recently | 21:08:51 |
| John Ericson | hah | 21:08:59 |
| John Ericson | well this should be a lot simpler because normal platforms and standard compliant linkers | 21:09:18 |
| John Ericson | alexfmpe: and then comes the fun work of making it also build for mingw :D | 21:10:01 |
| 23 Aug 2021 |
| alexfmpe | In reply to @alexfmpe:matrix.org
and started adding those changes to non-8.10 ghc mingw 9.10 seems to also be broken, I get a few
In file included from compiler/cbits/genSym.c:1:0: error:
includes/Rts.h:29:10: error:
fatal error: windows.h: No such file or directory
29 | #include <windows.h>
| ^~~~~~~~~~~
|
29 | #include <windows.h>
| 15:17:01 |
| alexfmpe | In reply to @alexfmpe:matrix.org
and started adding those changes to non-8.10 ghc * mingw ghc 9.0 seems to also be broken, I get a few
In file included from compiler/cbits/genSym.c:1:0: error:
includes/Rts.h:29:10: error:
fatal error: windows.h: No such file or directory
29 | #include <windows.h>
| ^~~~~~~~~~~
|
29 | #include <windows.h>
| 15:17:10 |
| alexfmpe | I don't see anything obviously related to <windows.h> in 8.10.4.nix | 15:17:42 |
 Rick (Mindavi) | Case-sensitive? | 15:41:29 |
| Rick (Mindavi) | I could imagine that it's not shipped too, but I don't know | 15:42:57 |
| alexfmpe | In reply to @rick:matrix.ciphernetics.nl
Case-sensitive? I mean, there are serveral .isWindows in both ghc derivations | 15:44:28 |
