Replace access()#91
Conversation
|
the proper way to test this change would be to rebuild the whole toolchain, and then create a test app that calls Edit: well, I went ahead and rebuilt everything from scratch, and tested a sample psl1ght app, calling access() to some files and everything worked as expected. (maybe there was a shorter/faster way to test it 😅 ) |
|
I can't test for now, sorry. My laptop took a shit and I don't have a backup lappy |
|
The contents look fine, but I would suggest making this as a PR against newlib-1.20.0-PS3 and then updating the patch based on that. That makes it easier to track changes and is how development was made on newlib-1.19.0. |
|
@zeldin alright, I'm making a PR to newlib-1.20.0-PS3 with these changes, and then if everything looks good both PRs can be merged: ps3dev/newlib-1.20.0-PS3#1 @miigotu oops, that's too bad, I hope you can fix your laptop! thanks for the help 👍 |
|
@zeldin submodules have come a long way since we tried it last time. What do you think about trying them out again using the *-PS3 libs in a new branch off-master? |
|
@miigotu You mean building newlib from a submodule instead of dist archive + patch? Yes, that would be more convenient for development. Feel free to create the branch, I can try it out next week when my vacation starts... 😸 |
|
I'll try lol. But first I have to try and get this 12 year old laptop set up to use until my new one comes lol... |
|
Rather than merging this PR, I simply regenerated the patch from the newlib-1.20.0-PS3 repo. The two patches generated identical results when applied to the unpatched newlib sources, save a minor whitespace difference in syscalls.c. |
This is a proposed
newlibpatch with an emulated implementation ofaccess()access.cis based on @crystalct file: https://github.com/crystalct/PS3LibrariesUpdate/blob/master/src/access.c