On Fri, Dec 15, 2023 at 10:53:31AM +0800, Maytham Alsudany wrote: > I'm looking for a sponsor to upload the packages I've prepared for gtea and its > dependencies. > > gtea I am waiting for your comments below for this. > ├── golang-code-gitea-sdk Uploaded after removing test-dep from Depends. > │ └── golang-github-go-fed-httpsig You added a patch to disable a chunk of unit-tests. While skipping tests is OK, it is good to usually have "some" reasoning behind doing so. For this particular case, only sha1 test was failing and only this part of the test should have been removed instead of disabling the entire chunk. This was failing because sha1 was probably not supposed to be included in the package but it was implemented anyway due to a bug in the go compiler itself; the tests were not accordingly patched at this tag. Since go/crypto still supports sha1, I let it pass for now. This could have been also made to pass via removing "crypto.SHA1" in hashToDef dict. However, removing sha1 support entirely requires more changes -- which are fairly simple to do -- these can be patched accordingly once go/crypto removes support for SHA1 altogether. I uploaded this after fixing up the tests. > └── golang-gitea-noerw-unidiff-comments > └── golang-github-seletskiy-tplutil tplutil is <100 lines of code, of which even less is used in unidiff-comments. This is used only in writer/changeset_writer.go. Now, final gtea package only uses "unidiff.ReadChangeset" function from this package. I did not check myself but I suppose it maybe possible to exclude writer/ from unidiff-comments and get gtea building. If this makes sense, could you consider to try this out? Also, let me know if you disagree with the above approach. Best, Nilesh
Attachment:
signature.asc
Description: PGP signature