Re: [PATCH v2 5/5] builtin/config: do not spawn pager when printing color codes
- From
Patrick Steinhardt <ps@pks.im>
- Date
- Sep 18, 2025, 06:03 UTC
- Message-ID
- <aMuglu_TtQPht1xP@pks.im>
- In-Reply-To
- <xmqqikhjhbgo.fsf@gitster.g>
On Mon, Sep 15, 2025 at 10:28:23AM -0700, Junio C Hamano wrote:
Show 33 quoted lines
> Patrick Steinhardt <ps@pks.im> writes:
>
> > With `git config get --type=color` the user asks us to parse a specific
> > configuration key and turn the value into an ANSI color escape sequence.
> > The printed string can then for example be used as part of shell scripts
> > to reuse the same colors as Git.
> >
> > Right now though we set up the auto-pager, which means that the string
> > may instead be written to the pager command. This is of course quite
> > nonsensical; there shouldn't be any use case where the color code should
> > end up in the pager instead of in the TTY.
> >
> > Fix this by disabling the pager in case the user is asking us to print
> > color sequences.
>
> I am of two minds. Part of me obviously agrees that it is more
> straight forward with this change. But it may
>
> An interactive user experimenting while writing their own script
> might say something like
>
> $ git config --type=color --default="reverse red" n.n
>
> If the command emitted directly to the terminal, then everything
> they type from then on will be bloody red, but the pager protects
> them from such an accident. Instead, they are forced to say
>
> $ C=$(git config get --type=color --default="reverse red" n.n)
> $ R=$(git config get --type=color --default="reset" n.n)
> $ echo "So$C Bloody ${R}Red"
>
> but these are likely what they would be writing in their script
> anyway, so...True. That being said, I'm mostly trying to emulate the old behaviour that we had in `git config --get-color`. We have the following condition there:
/*
* The following actions may produce more than one line of output and
* should therefore be paged.
*/
if (actions & (ACTION_LIST | ACTION_GET_ALL | ACTION_GET_REGEXP | ACTION_GET_URLMATCH))
setup_auto_pager("config", 1);`ACTION_GET_COLOR` is not part of this condition, so we wouldn't set up the auto pager there, either. So I think it's sensible to match that behaviour so that the new command really is a drop-in replacement for the old one.
I should probably clarify the commit message.
Patrick