Bastien Guerry <[email protected]> writes:

> IIUC, your suggestion is that, if the attached patch is well formatted
> and has [PATCH] in its suject then it should be promoted as a patch?

I've done that now, https://tracker.orgmode.org/next/ shows the new
output with more patches.

On the other hand, I've tighten the rule so that a patch either has a
[PATCH] or a well-formatted patch (with git format-patch).

So inline diffs and quick git diff > x.patch with no explicit [PATCH] 
in the subject are not promoted as patch by BARK, they are just code
for comment in a thread.

See https://tracker.orgmode.org/next/bark-manual.html

> I agree that "false negatives are worse than false positives" and that
> explains the previous behavior, catching as much patches as possible.
> But it also makes sense to leave room for quick patches that are still
> visible within bug/request reports, but not promoted as "standalone"
> patches: let's find the right trade-off here.

I'm waiting for Jacob and Ihor to comment on
https://tracker.orgmode.org/next/ but I think it's more useful. Fewer
false positives, and perhaps more importantly, it's easier to remember
what is detected as a patch.

Thanks!

-- 
 Bastien

Reply via email to