Implement latest regex for fetch and cache#27734
Conversation
|
Alright everything seems to be working as expected now, except Nothing much we can do about that except wait it out, Manim's tarballs weren't effected as far as I am able to tell. |
11b9b11 to
948f63c
Compare
|
Yeah that’s truly an unfortunate situation… Everything was deleted, including the GitHub organization and their other social media platforms. |
948f63c to
da2d5e8
Compare
|
I've managed to part these changes out into separate commits now but continuing to work on the auto-update system may yet drive me mad. I think on a meta level I'd like to get rid of the distinction between the "@ALL" and "list of packages" modes of operation in the auto-updater. API induced delirium aside, that is out of scope for this PR though and we need to just get the auto-update workflow back up and running. |
da2d5e8 to
8ccb9b4
Compare
|
That's not great... |
1c7ba7a to
433085c
Compare
|
Okay I said I wasn't gonna unify the We just need to accumulate a package list, eliminate anything the pre-cache caught as up-to-date and shunt the finished list back into One minor optimization we can make here is providing the package list as termux-packages/scripts/bin/update-packages Lines 418 to 433 in fb6d543 |
433085c to
5ed2360
Compare
|
Alright, I'm finally out of spinning plates, so I can get back to this one. |
0254f19 to
b280948
Compare
0a657f1 to
069e335
Compare
"line 60: [: =: unary operator expected" and some minor build script cleanup %ci:no-build
no longer needed for these packages %ci:no-build
…dk-` %ci:no-build
…_VERSION_REGEXP where necessary %ci:no-build
…TE_VERSION_REGEXP %ci:no-build
0fec46d to
a74e22a
Compare
|
Let's see if the "days since Tom broke prod" counter gets to tick up another day. I've crossed my t's and dotted my i's. |
|
This PR completely ignores repology update issue, which I intend to fix |
|
Welp, there goes my counter I guess. |
I wasn't aware of a repology update issue, could you explain it? |
https://github.com/termux/termux-packages/actions/runs/21380867349/job/61547280850#step:4:20 |
|
Oh ok, I have seen that before but I didn't know that was what you were talking about, I didn't realize that was causing a measurable problem because, even though it has been happening for a long time, repology-based auto updates continued to succeed despite that, for example 7c279bb but I agree it will be good for it to get fixed |
|
I can't replicate that issue locally, nor have I seen it on the CI. Absolutely do look into it though, if it's happening it should be fixed. |
It's not caused by this PR, it's something that has been happening for a long time before this PR |
less has its own update function that isnt repology I am pretty sure its 22d58b7 that broke, cc @thunder-coding I do intend to fix but I am busy |
|
Ah, I had temporarily changed that back to That was running into the 15 minute timeout. |
|
@truboxl I think I am fixing it in this while also fixing something else, review would be appreciated, especially if you don't like this way of fixing it and would prefer a different way |
I need to just throw this up as a PR for further testing at this point.
The main issue with the auto-updater after the gitlab method refactor seems to have been that we never taught it how to use
latest-regexin the_fetch_and_cache_tags/@allcodepath.This PR implements that and some related cleanup.
The changes should be broken out into multiple commits,
I don't have the time to do so right now,
and I need to see if this behaves as it should on the CI.