mbox series

[0/3] Override isatty with GIT_FORCE_TTY

Message ID 20210706210317.706313-1-me@avishay.dev (mailing list archive)
Headers show
Series Override isatty with GIT_FORCE_TTY | expand

Message

Avishay Matayev July 6, 2021, 9:03 p.m. UTC
This is the resulting patch series from the discussion in the mailing list[1]
and the Fugitive github issue[2].

The goal is to allow to override git decisiveness about not doing
something when some fd is not a pty, which can hinder Fugitive.

I didn't annotate the location of the commits as I wasn't sure what was
the correct one.. As it is not specific to a single mechanisms, but
rather it is something global, is it fine as it is or should I use
something as `global, `tree:` or `tree-wide:`?

I also didn't document this feature, or at least I wasn't sure where
to document it (other than on these commits), I would compare it to the
GIT_SMART_HTTP as it is something pretty niche and it isn't (and IMO
shouldn't be) documented to the regular user.

Thank you,
Avishay

1. https://lore.kernel.org/git/CAJ-0OswsrnAuCwU6U=S2i1qKkg=66U-8RHSGqD2kh9T_30Yw9w@mail.gmail.com/
2. https://github.com/tpope/vim-fugitive/issues/1772

Avishay Matayev (3):
  Change isatty's parameter to a *_FILENO define
  Allow isatty to be overriden with GIT_FORCE_TTY
  Add tests for GIT_FORCE_TTY

 builtin/am.c               |  2 +-
 builtin/bisect--helper.c   |  4 ++--
 builtin/blame.c            |  2 +-
 builtin/bundle.c           |  2 +-
 builtin/checkout.c         |  2 +-
 builtin/commit-graph.c     |  4 ++--
 builtin/commit.c           |  2 +-
 builtin/fsck.c             |  2 +-
 builtin/gc.c               |  2 +-
 builtin/log.c              |  2 +-
 builtin/merge.c            |  4 ++--
 builtin/multi-pack-index.c |  2 +-
 builtin/pack-objects.c     |  2 +-
 builtin/pack-redundant.c   |  2 +-
 builtin/prune-packed.c     |  2 +-
 builtin/prune.c            |  2 +-
 builtin/rebase.c           |  2 +-
 builtin/repack.c           |  2 +-
 builtin/send-pack.c        |  2 +-
 builtin/shortlog.c         |  4 ++--
 builtin/sparse-checkout.c  |  2 +-
 builtin/unpack-objects.c   |  2 +-
 cache.h                    |  3 +++
 color.c                    |  2 +-
 column.c                   |  2 +-
 compat/mingw.c             |  6 +++---
 compat/mingw.h             |  2 +-
 compat/winansi.c           | 16 ++++++++--------
 config.c                   |  5 +++++
 date.c                     |  2 +-
 editor.c                   |  2 +-
 pager.c                    |  4 ++--
 preload-index.c            |  2 +-
 read-cache.c               |  2 +-
 remote-curl.c              |  2 +-
 sequencer.c                |  4 ++--
 sideband.c                 |  2 +-
 t/t9904-git-force-tty.sh   | 38 ++++++++++++++++++++++++++++++++++++++
 transport.c                |  6 +++---
 transport.h                |  2 +-
 wt-status.c                |  2 +-
 41 files changed, 101 insertions(+), 55 deletions(-)
 create mode 100755 t/t9904-git-force-tty.sh

Comments

Avishay Matayev July 30, 2021, 12:04 p.m. UTC | #1
On Wed, 7 Jul 2021 at 00:03, Avishay Matayev <me@avishay.dev> wrote:
>
> This is the resulting patch series from the discussion in the mailing list[1]
> and the Fugitive github issue[2].
>
> The goal is to allow to override git decisiveness about not doing
> something when some fd is not a pty, which can hinder Fugitive.
>
> I didn't annotate the location of the commits as I wasn't sure what was
> the correct one.. As it is not specific to a single mechanisms, but
> rather it is something global, is it fine as it is or should I use
> something as `global, `tree:` or `tree-wide:`?
>
> I also didn't document this feature, or at least I wasn't sure where
> to document it (other than on these commits), I would compare it to the
> GIT_SMART_HTTP as it is something pretty niche and it isn't (and IMO
> shouldn't be) documented to the regular user.
>
> Thank you,
> Avishay
>
> 1. https://lore.kernel.org/git/CAJ-0OswsrnAuCwU6U=S2i1qKkg=66U-8RHSGqD2kh9T_30Yw9w@mail.gmail.com/
> 2. https://github.com/tpope/vim-fugitive/issues/1772
>
> Avishay Matayev (3):
>   Change isatty's parameter to a *_FILENO define
>   Allow isatty to be overriden with GIT_FORCE_TTY
>   Add tests for GIT_FORCE_TTY
>
>  builtin/am.c               |  2 +-
>  builtin/bisect--helper.c   |  4 ++--
>  builtin/blame.c            |  2 +-
>  builtin/bundle.c           |  2 +-
>  builtin/checkout.c         |  2 +-
>  builtin/commit-graph.c     |  4 ++--
>  builtin/commit.c           |  2 +-
>  builtin/fsck.c             |  2 +-
>  builtin/gc.c               |  2 +-
>  builtin/log.c              |  2 +-
>  builtin/merge.c            |  4 ++--
>  builtin/multi-pack-index.c |  2 +-
>  builtin/pack-objects.c     |  2 +-
>  builtin/pack-redundant.c   |  2 +-
>  builtin/prune-packed.c     |  2 +-
>  builtin/prune.c            |  2 +-
>  builtin/rebase.c           |  2 +-
>  builtin/repack.c           |  2 +-
>  builtin/send-pack.c        |  2 +-
>  builtin/shortlog.c         |  4 ++--
>  builtin/sparse-checkout.c  |  2 +-
>  builtin/unpack-objects.c   |  2 +-
>  cache.h                    |  3 +++
>  color.c                    |  2 +-
>  column.c                   |  2 +-
>  compat/mingw.c             |  6 +++---
>  compat/mingw.h             |  2 +-
>  compat/winansi.c           | 16 ++++++++--------
>  config.c                   |  5 +++++
>  date.c                     |  2 +-
>  editor.c                   |  2 +-
>  pager.c                    |  4 ++--
>  preload-index.c            |  2 +-
>  read-cache.c               |  2 +-
>  remote-curl.c              |  2 +-
>  sequencer.c                |  4 ++--
>  sideband.c                 |  2 +-
>  t/t9904-git-force-tty.sh   | 38 ++++++++++++++++++++++++++++++++++++++
>  transport.c                |  6 +++---
>  transport.h                |  2 +-
>  wt-status.c                |  2 +-
>  41 files changed, 101 insertions(+), 55 deletions(-)
>  create mode 100755 t/t9904-git-force-tty.sh
>
> --
> 2.25.1
>
Bumping this thread, I'll appreciate any constructive feedback.