Re: [PATCHv3 1/4] color.h: document and modernize header

From: Junio C Hamano <gitster@pobox.com>
Date: Jan 8, 2018, 19:14 UTC
Message-ID: <xmqqr2r088p3.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <20180104224007.214548-2-sbeller@google.com>

Stefan Beller <sbeller@google.com> writes:

Show 9 quoted lines

>  /*
> - * Set the color buffer (which must be COLOR_MAXLEN bytes)
> - * to the raw color bytes; this is useful for initializing
> - * default color variables.
> + * NEEDSWWORK: document this function or refactor grep.c to stop using this
> + * function.
>   */
> -void color_set(char *dst, const char *color_bytes);
> +extern void color_set(char *dst, const char *color_bytes);

The original that is removed by the patch documents the function well enough; as long as the NEEDSWORK comment is followed through in a later step in the series, it's alright, though ;-)

Show 12 quoted lines

> -int git_config_colorbool(const char *var, const char *value);
> -int want_color(int var);
> -int color_parse(const char *value, char *dst);
> -int color_parse_mem(const char *value, int len, char *dst);
> +/*
> + * Parse a config option, which can be a boolean or one of
> + * "never", "auto", "always". Return a constant of
> + * GIT_COLOR_NEVER for "never" or negative boolean,
> + * GIT_COLOR_ALWAYS for "always" or a positive boolean,
> + * and GIT_COLOR_AUTO for "auto".
> + */
> +extern int git_config_colorbool(const char *var, const char *value);

"never" and "always" not being part of usual boolean vocabulary makes it a bit awkward to explain.

Show 7 quoted lines

> +/*
> + * Output the formatted string in the specified color (and then reset to normal
> + * color so subsequent output is uncolored). Omits the color encapsulation if
> + * `color` is NULL. The `color_fprintf_ln` prints a new line after resetting
> + * the color. The `color_print_strbuf` prints the given pre-formatted strbuf
> + * instead, up to its first NUL character.
> + */

Obviously, it does not have to be part of this step nor series, but the above observation makes us realize that color_print_strbuf() would probably be an unreasonably narrow interface. It is not too much to ask the caller to dereference and pass only the .buf component of the strbuf to an alternative helper that takes "const char *" and by doing so would allow us to let other callers that do not have a strbuf but just a plain string use it, too.

Looks good.