Open simonsan opened 2 years ago
That's great research right there and I can assure you that nothing has been missed. This oddity was intentional knowing that I could write a tool that just suits me, with my preferences hard-coded.
However, I see (and happily so) that cargo smart-release
outgrows its original audience (me) and that comes with changing requirements. I think it's fair to not be surprising and as a first step, pass on 'kind' names unchanged with the option of passing additional filters on the command-line or via configuration files.
For now the fix is only available on main
, and I will create a new release once Byron/gitoxide#316 is fixed.
Thanks for giving it a shot!
[TRACE] Skipping unknown git-conventional kind "chore" and all 2 message(s) in it.
Maybe I'm doing something wrong, but I can't get
cargo-smart-release
to "find" conventional commit types likechore
orfeat(x)
. Is there something I've overlooked or doing wrong?After a bit of research it seems that the
conventional commit type
is being parsed from the message, butcargo changelog
doesn't want them to be written to the changelog.I see that in the changelog generation it's not being matched for e.g.
chore
: https://github.com/Byron/gitoxide/blob/fbe2c93b13fc05414dc5c7d80f4197d279b6b81b/cargo-smart-release/src/changelog/section/segment.rs#L26 while in https://github.com/Byron/gitoxide/blob/fbe2c93b13fc05414dc5c7d80f4197d279b6b81b/cargo-smart-release/src/commit/message.rs#L122 these clearly exist.I'm asking myself if it would be sensible to do the mapping 1:1 in regards to conventional commits and pass a filter to
cargo changelog
where someone could state that they might not want to havechores
printed to the changelog file or have their own mapping parsed from atoml
configuration file.As said above, if I'm overlooking something, please correct me. :)
Thanks for the nice tooling!