gccrs: Initial implementation of hotpatching rustc-lib - #4924
Polygonalr wants to merge 1 commit into
Conversation
|
You could try naming the file |
|
or maybe something even clearer since this is just for the kernel like |
There was a problem hiding this comment.
the Python code looks good to me but I'm really not good at Python :P for the README, I think it would be good to have a section on how to generate these patches and to re-iterate that these should only be present if they are NEEDED by the KERNEL to compile. this makes it clearer that we do intend to not patch core, but also want to start experimenting with the kernel and have to make some adjustments and concessions in that specific case
105bd97 to
3307946
Compare
|
I'm writing a script to update the vendored libraries you might want to integrate this with in some way. I'll have a PR up shortly. Edit: #4934 |
|
@powerboat9 Hmmm not too sure what's there to change in my PR besides updating the patch files manually when we do a core version bump |
3307946 to
6f7a75e
Compare
Addresses Rust-GCC#4835 libgrust/ChangeLog: * rustc-lib-patches/patches/0001-core-ffi-c-str.for-rfl-patch: Backported patch to support CStr parsing. * rustc-lib-patches/patches/0002-alloc-vec-extract-if.for-rfl-patch: Backported vec's implementation of extract_if. * rustc-lib-patches/rustc-lib-patch.py: Script to apply the patch files to rustc-lib. * rustc-lib-patches/README.md: the README file for this dir. Signed-off-by: Yap Zhi Heng <yapzhhg@gmail.com>
6f7a75e to
177ca08
Compare
This branch has not been deployed
3 participants
Closes #4835 (?)
ngl I feel icky about this implementation, making this PR to receive opinions on how to move #4835 forward, as it likely requires either backporting the feature to core, or bumping our target core version to 1.72, the latter being less than ideal.
Patch file for #4835 is not done yet, just patchingCStras an example first.I'm also not too sure about the stance of including .patch files in the gcc repo, since the root .gitignore ignores .patch files.
PR finalization edit: I kept both the implementations of
CStrandExtractIfminimal for now because I find it too tedious to test the compilation of the full suite ofimpl-s (especially so forCStr) when we are still unable to compilecore. When moreimpl-s are discovered to be required, we can either add them as additional patch files or modify the existing ones.