Re: [PATCH v10 2/3] interpret-trailers: add own-identity option

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Junio C Hamano <gitster@xxxxxxxxx> 于2021年3月19日周五 上午3:20写道:
>
> "ZheNing Hu via GitGitGadget" <gitgitgadget@xxxxxxxxx> writes:
>
> > From: ZheNing Hu <adlternative@xxxxxxxxx>
> >
> > Beacuse `git commit --trailer="Signed-off-by: \
> > $(git config user.name) <$(git config user.email)>"`
> > is difficult for users to add their own identities,
> > so teach interpret-trailers a new option `--own-identity`
> > which allow those trailers with no value add the user’s own
> > identity. This will help the use of `commit --trailer` as
> > easy as `--signoff`.
>
> I have a suspicion that this is too narrowly focused to be useful in
> practice, and I find that the proposed "--own-identity" is quite a
> mouthful.
>

Well, the original meaning of this `--trailer`and`--own-identity` option is to
imitate the behavior of `--signoff` to create other trailers, For the
time being,
it can only "imitation", it does not yet support the ability to provide multiple
identities instead of "own-identity".

If `--own-identity` is mouthful, is there a better name?

> > +--own-identity::
> > +     Used with `--trailer`. Those trailers without value with the
> > +     `--own-identity` option all will add the user's own identity.
>
> So, the assumption here is that the name of the trailer tag alone,
> without the ':' separator, can identify which trailer the user is
> talking about, and it can be distinguished from the tag name plus ':'
> and nothing else which is calling for a trailer entry with an empty
> string as its value?
>

Yes. `--own-ident` is for trailers without a separator.

> OK.
>
> The reason why this looks too narrowly focused on oneself alone to
> be useful to me is because I often need to add various -by trailers
> to incoming patches, and have a script to do exactly that (which
> does not use interpret-trailers, as I do not think
> interpret-trailers can accept a patch email as its input, and the
> script predates interpret-trailers) but it will be useless if that
> script were limited to add -by for myself.
>

I understand it. But I think `--own-ident` is more inclined to add
identities for users. If users want to add other people’s identities,
they need to use another method.

> Wouldn't it be a lot more useful if
>
>         git commit --trailer="Helped-by:@Ch.*Couder"
>
> is expanded (note: I am not married to the syntax, but only for
> illustration purposes, I am using "a value prefixed by @ triggers
> the 'name expansion'" convention in this example.  People can come
> up with better convention) by finding an author or a committer whose
> name matches the given pattern from "git log"?  Then, instead of
>
>         git commit --own-identity --trailer=Signed-off-by
>
> I can say
>
>         git commit --trailer=Signed-off-by:@gitster
>
> and I can even add more than one, e.g.
>
>         git commit --trailer=Helped-by:@peff --trailer=Signed-off-by:@gitster
>

I agree that this idea can be extended again from "own-identity "
mode to "multiple-identity mode",‘@’ seems to have a lot of
 important meanings in git, but I don’t know how to design a solution
 that is more suitable than yours for the time being.

 Another doubt  is what happens if one person corresponds to
 multiple mailboxes or one mailbox corresponds to multiple people?

I think `--own-ident` may be a somewhat "narrow" option that can be
kept. Also add a new option for interpret-trailers to resolve the identity
of other people?

Thanks for suggestions.




[Index of Archives]     [Linux Kernel Development]     [Gcc Help]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [V4L]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]     [Fedora Users]

  Powered by Linux