Closed yutannihilation closed 1 year ago
Should track_rust_source
file be removed completely? Do we use its functions anywhere else?
Also, I believe it is worth updating the changelog.
Oh, it might be. I'll try it.
I'm still figuring out, but I think this doesn't introduce any user-visible changes. What do you think?
Hm, you are correct. We used pkgbuild before. The only visible change is the minimum required version :)
Yeah. One notable progress is that the developers now can just press the "Load all" menu of RStudio as long as no signatures are changed. But this is not our archivement...
Should
track_rust_source
file be removed completely? Do we use its functions anywhere else?
I checked. In short, there's not many functions that can be removed.
pretty_rel_path()
is used here; pretty_rel_single_path()
and on_error_return_default()
are used inside the function.
get_library_path()
is used here.
find_newer_files_than()
is used here get_file_info()
is used inside the function.
Sorry I missed your comment here. It feels like get_file_info()
should no longer be used?
No, I meant get_file_info()
is used in find_newer_files_than()
, which is used in register_extendr()
.
Ok, now I see. Perhaps it can later be simplified since we only want to check if wrappers are newer than dll or not.
Thanks for the hint! I replaced it with bare file.info()
.
Do you feel like it is ready for review? Or are there other things you need to address first?
Yes, I think so. I was wondering if I can reduce more code, but it seems this is all for now.
Hm, it blew up because of {purrr}
. Strange
DLL 'purrr' not found: maybe not installed for this architecture?
Yeah, I have no idea. I hope time will solve...
Fix #215
(Sorry, I mistakenly committed this directly to the main branch... I reverted)