git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: Discussion for interactive --patch commands to get --unified support

From
Leon Michalak <leonmichalak6@gmail.com>
Date
May 2, 2025, 16:14 UTC
Message-ID
<CAP9jKjG+khoUmRpVJ8om-bs_qjB=VFCj3p3h0VSaBHVnqPLNVg@mail.gmail.com>
In-Reply-To
<bf7f6606-e719-4c3d-b7ab-ef7351f66f37@gmail.com>

Inheriting the diff.context setting is what scratches my itch the most, although also being able to set the context in the command list of `add -i` sounds interesting too. Personally, I don't think I would use the command line overrides too much myself as most of the time (like with diff) I'd like to set the option and forget it but it does have a certain consistency to it.

Slightly off-topic to the discussion, but does anyone have advice on how to deal with providing a sentinel value for something like context? I'd expect to pass `--unified` to the underlying diff command *only* if the user specifically has overridden it via command line option or a diff.context config, just like diff.algorithm has done, however diff.algorithm is a `char *` so the value can be NULL which is a good sentinel value. My thinking is then the underlying command can just deal with the value as it sees it, such as giving a default if not provided or making sure it's a minimum of 0 etc. Otherwise the level above has to deal with it which then probably involves `git_diff_ui_config` and other validations which I don't even think is it's responsibility and would probably duplicate logic unncessarily?

I may be completely off in my assumptions here being new to the codebase, so if anyone has any thoughts I'd greatly appreciate any comments!

On Fri, 2 May 2025 at 15:39, Phillip Wood <phillip.wood123@gmail.com> wrote:
Show 18 quoted lines
>
> On 29/04/2025 10:16, Leon Michalak wrote:
> >
> > (https://stackoverflow.com/questions/6711670/git-show-more-context-when-using-git-add-i-or-git-add-e)
> > which mentions you can do `GIT_DIFF_OPTS=-u<number> git add -p` which
> > does work however isn't very user friendly or convenient.
> This is a question for others on the list rather than Leon - is it
> intentional that the plumbing diff commands respect GIT_DIFF_OPTS? If a
> script that wants to create a diff with a certain number of context
> lines runs `git diff-index -U <context>` is it helpful for that to be
> overridden if GIT_DIFF_OPTS happens to be set in the environment?
> Looking at the history it seems that environment variable used to be the
> only way to override the default context setting but that's not the case
> now.
>
> Best Wishes
>
> Phillip
Previous: Phillip WoodNext: Leon Michalak
Message 8 of 12 in “Discussion for interactive --patch commands to get --unified support”
  1. Leon MichalakApr 29, 2025
  2. Junio C HamanoApr 29, 2025
  3. Jeff KingApr 29, 2025
  4. Leon MichalakApr 30, 2025
  5. Junio C HamanoApr 30, 2025
  6. Phillip WoodMay 2, 2025
  7. Phillip WoodMay 2, 2025
  8. Leon MichalakMay 2, 2025
  9. Leon MichalakMay 2, 2025
  10. Junio C HamanoMay 2, 2025
  11. Leon MichalakMay 2, 2025
  12. Junio C HamanoMay 2, 2025

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.