Documentation fixes - #5111
Conversation
|
Not sure what to make of it, but there are a lots of holes caused by things I've "fixed" in docurium. For example, our filter groups are different. I have hashsig, which doesn't show up at all now. I think that's the docparser-cleanups branch I have (passing the tempdir as the include path), I'll try to confirm. Not trying to be dense, I'm having a hard time remembering what's fixing what where how 😜. On the other hand, git_odb_backend (whose "ours" version should look better after this PR gets in), which I ought to be still mangling because the typedef in "types.h" gets parsed after the struct "sys/odb_backend.h", overwriting the comments). |
You mean why there's no compile errors due to unknown prototypes/functions? Most users include "git2.h" directly which transitively includes all of our external headers, including "indexer.h". But docurium wouldn't do that, hence why it will probably fail to generate proper documentation here. |
pks-t
left a comment
There was a problem hiding this comment.
Looks good to me, thanks for all your nice documentation work! Not merging yet to give others a chance to chime in.
Docurium seems to choke on this header because it can't see both git_indexer_progress & git_indexer_progress_cb, let's add the include.
|
Rebased, should be ready to 🚢. |
|
Looks good to me. Thanks, @tiennou! |
I've made some runs of my custom docurium, and here are the results.
I'm not sure how commit 1 is possible but you never know ^^. Note that I'm adding documentation for things which are just out of one-pass-declaration-parsing, so either we're fine with them as a first approximation, or I'll need help to clarify (a.k.a. help is welcome).