-
Notifications
You must be signed in to change notification settings - Fork 185
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add GNU/Hurd support #370
Add GNU/Hurd support #370
Conversation
Could we have a link to GNU/Hurd's documentation (for both urandom and the errno location). We should also add it to the list of targets in the CI |
Well, the Hurd doesn't really document it since it's just following GNU's interface. At best we can link to the implementation: |
Concerning the CI, the patches needed to build rustc and std are not finished committing in the various repositories. Actually getrandom is one of them :) so we have a chicken-and-egg issue if CI needs to be included before support is included |
For new targets, we don't require that Rust ships a However, provided that Rust supports your compiler target, we can still test them. See here. However, it looks like rust-lang/rust#115230 is merged. The change we needed from that was adding support for |
Yes it is possible, but as usual one needs to have the GNU/Hurd libc.so etc. at hand to be able to link. One can relatively easily build them from glibc's |
I was able to compile a C program for HURD (with Clang) on my Debian Box by just installing |
If Hurd is just implementing glibc interfaces, does it have support for |
It looks like We should have the docs link to the glibc man page: https://www.gnu.org/software/libc/manual/2.38/html_mono/libc.html#index-getrandom EDIT: It seems like the implementation also involves opening a file (I think), but it also seems to have some other complicated logic. I think to be safe we should use glibc's |
I think this PR is also blocked on libc support, see rust-lang/libc#3325 (comment) |
I guess your toolchain was then using linux's 32bit libc, so while this indeed exercises a hurd build, the actual linked result is a Linux binary, not a Hurd binary :) |
Ok, done so! |
(the mips failure is unrelated) |
Should be fixed by #372 |
I rebased on top of #372, and added std=core for now (building std depends on libc dep bump in rustc) |
The netbsd build failure is unrelated |
It seems like this current PR doesn't use the |
Yep, it's caused by cross-rs/cross#1345 |
Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Indeed, fixed so. |
Any news on this, is there still something missing? |
Thank you! The CI failure is not relevant to this PR, so we can ignore it. |
No description provided.