threads / bug / 54860

[BUG] In git-log, --name-status flag overrides the --patch flag

Subject: [BUG] In git-log, --name-status flag overrides the --patch flag

## tl;dr

4 messages between Dec 20, 2020 and Dec 23, 2020.

replies: 3people: 3as markdown or json

Érico Rolim· Dec 20, 2020, 04:23 UTC · lore

Hi!

When using the command

  git log --patch --name-status

It seems the name-status flag somehow overrides the patch flag, by which I mean that I get the same output as simply running

  git log --name-status

It would be nice for the combination of these two flags to work, as it would allow one to view both a summary of changed files as well as the changes themselves, at the same time.

Happy holidays, Érico

Đoàn Trần Công Danh· Dec 21, 2020, 13:16 UTC · re: Érico Rolim · lore

Re: [BUG] In git-log, --name-status flag overrides the --patch flag

On 2020-12-20 01:23:05-0300, Érico Rolim <erico.erc@gmail.com> wrote:

Show 14 quoted lines

> Hi!
> 
> When using the command
> 
>   git log --patch --name-status
> 
> It seems the name-status flag somehow overrides the patch flag, by which I mean
> that I get the same output as simply running
> 
>   git log --name-status
> 
> It would be nice for the combination of these two flags to work, as it would
> allow one to view both a summary of changed files as well as the changes
> themselves, at the same time.

I'm not arguing if this your expectation is make sense or not, however, the Documentation says something about "Show only"...

Anyway, --patch-with-raw provides more information than --patch --name-status combined. I guess it's too much information?

-- 
Danh

Junio C Hamano· Dec 21, 2020, 19:09 UTC · re: Đoàn Trần Công Danh · lore

Re: [BUG] In git-log, --name-status flag overrides the --patch flag

Đoàn Trần Công Danh  <congdanhqx@gmail.com> writes:

Show 18 quoted lines

> On 2020-12-20 01:23:05-0300, Érico Rolim <erico.erc@gmail.com> wrote:
>> Hi!
>> 
>> When using the command
>> 
>>   git log --patch --name-status
>> 
>> It seems the name-status flag somehow overrides the patch flag, by which I mean
>> that I get the same output as simply running
>> 
>>   git log --name-status
>> 
>> It would be nice for the combination of these two flags to work, as it would
>> allow one to view both a summary of changed files as well as the changes
>> themselves, at the same time.
>
> I'm not arguing if this your expectation is make sense or not,
> however, the Documentation says something about "Show only"...

Perhaps the documentation would need to be clarified? I suspect this is not very useful combination [*1*] and if this were years ago, I would suggest making the command line error out, instead of silently ignore one and keep the other one in effect.

[Footnote]

*1* "--patch --stat --summmary" was designed as the standard "both a
    summary as well as the changes".

Đoàn Trần Công Danh· Dec 23, 2020, 23:54 UTC · re: Junio C Hamano · lore

Re: [BUG] In git-log, --name-status flag overrides the --patch flag

On 2020-12-21 11:09:58-0800, Junio C Hamano <gitster@pobox.com> wrote:

Show 22 quoted lines

> Đoàn Trần Công Danh  <congdanhqx@gmail.com> writes:
> 
> > On 2020-12-20 01:23:05-0300, Érico Rolim <erico.erc@gmail.com> wrote:
> >> Hi!
> >> 
> >> When using the command
> >> 
> >>   git log --patch --name-status
> >> 
> >> It seems the name-status flag somehow overrides the patch flag, by which I mean
> >> that I get the same output as simply running
> >> 
> >>   git log --name-status
> >> 
> >> It would be nice for the combination of these two flags to work, as it would
> >> allow one to view both a summary of changed files as well as the changes
> >> themselves, at the same time.
> >
> > I'm not arguing if this your expectation is make sense or not,
> > however, the Documentation says something about "Show only"...
> 
> Perhaps the documentation would need to be clarified?

Honestly, I think documentation is fine, "Show only"... should mean only.

Show 5 quoted lines

>
> I suspect
> this is not very useful combination [*1*] and if this were years
> ago, I would suggest making the command line error out, instead of
> silently ignore one and keep the other one in effect.

Although, I don't have strong opinion on this combination, I would say yes, and we should error out for other combination like --name-status --patch too. And Érico, in a reply to me, said he would prefer seeing an error instead of silencely discard one option. I'll look into the implementation in this weekend.

Show 6 quoted lines

> 
> 
> [Footnote]
> 
> *1* "--patch --stat --summmary" was designed as the standard "both a
>     summary as well as the changes".

-- 
Danh