Re: [PATCH 0/13] unraveling the mysteries of color variables

From: Jeff King <peff@peff.net>
Date: Sep 16, 2025, 20:12 UTC
Message-ID: <20250916201235.GA612811@coredump.intra.peff.net>
In-Reply-To: <20250916201036.GA612463@coredump.intra.peff.net>

On Tue, Sep 16, 2025 at 04:10:37PM -0400, Jeff King wrote:

Show 13 quoted lines

>   [01/13]: color: use GIT_COLOR_* instead of numeric constants
>   [02/13]: color: return enum from git_config_colorbool()
>   [03/13]: grep: don't treat grep_opt.color as a strict bool
>   [04/13]: diff: simplify color_moved check when flushing
>   [05/13]: diff: don't use diff_options.use_color as a strict bool
>   [06/13]: diff: pass o->use_color directly to fill_metainfo()
>   [07/13]: diff: stop passing ecbdata->use_color as boolean
>   [08/13]: pretty: use format_commit_context.auto_color as colorbool
>   [09/13]: color: use git_colorbool enum to type to store colorbools
>   [10/13]: color: return bool from want_color()
>   [11/13]: add-interactive: retain colorbool values longer
>   [12/13]: config: store want_color() result in a separate bool
>   [13/13]: color: convert git_colorbool into a struct

I forgot to mention that these should be applied on top of jk/add-i-color. Patch 11 touches the same spot in a way that creates lots of conflicts (the problems there pre-date that recent topic, but it was expanded from one color variable to two).

-Peff