[PATCH] handle arbitrary ints in git_config_maybe_bool

[PATCH] handle arbitrary ints in git_config_maybe_bool

[Jeff King]

This function recently gained the ability to recognize the
documented "0" and "1" values as false/true. However, unlike
regular git_config_bool, it did not treat arbitrary numbers
as true. While this is undocumented and probably ridiculous
for somebody to rely on, it is safer to behave exactly as
git_config_bool would. Because git_config_maybe_bool can be
used to retrofit new non-bool values onto existing bool
options, not behaving in exactly the same way is technically
a regression.

Signed-off-by: Jeff King <peff@peff.net>
---
This was posted earlier as part of the command-specific pager topic; you
ended up splitting part of that out into jk/maint-decorate-01-bool. This
should logically go on top of that (b2be2f6).

It probably doesn't make a difference in the real world, but I think it
is safer (as described above), and the code is a little cleaner. I
should have just done it this way in the first place.

 config.c |    8 +++-----
 1 files changed, 3 insertions(+), 5 deletions(-)

diff --git a/config.c b/config.c
index 32c0b2c..d73b090 100644
--- a/config.c
+++ b/config.c
@@ -429,13 +429,11 @@ static int git_config_maybe_bool_text(const char *name, const char *value)
 
 int git_config_maybe_bool(const char *name, const char *value)
 {
-	int v = git_config_maybe_bool_text(name, value);
+	long v = git_config_maybe_bool_text(name, value);
 	if (0 <= v)
 		return v;
-	if (!strcmp(value, "0"))
-		return 0;
-	if (!strcmp(value, "1"))
-		return 1;
+	if (git_parse_long(value, &v))
+		return !!v;
 	return -1;
 }
 
-- 
1.7.3.4.761.g98ad5

Re: [PATCH] handle arbitrary ints in git_config_maybe_bool

[Junio C Hamano]

Jeff King <peff@peff.net> writes:

> This function recently gained the ability to recognize the
> documented "0" and "1" values as false/true. However, unlike
> regular git_config_bool, it did not treat arbitrary numbers
> as true. While this is undocumented and probably ridiculous
> for somebody to rely on, it is safer to behave exactly as
> git_config_bool would. Because git_config_maybe_bool can be
> used to retrofit new non-bool values onto existing bool
> options, not behaving in exactly the same way is technically
> a regression.
>
> Signed-off-by: Jeff King <peff@peff.net>
> ---
> This was posted earlier as part of the command-specific pager topic; you
> ended up splitting part of that out into jk/maint-decorate-01-bool. This
> should logically go on top of that (b2be2f6).
>
> It probably doesn't make a difference in the real world, but I think it
> is safer (as described above), and the code is a little cleaner. I
> should have just done it this way in the first place.

I'd reword "arbitrary numbers" to "arbitrary non-zero numbers".
Other than that, thanks for your attention to the details.

Will apply.

>  config.c |    8 +++-----
>  1 files changed, 3 insertions(+), 5 deletions(-)
>
> diff --git a/config.c b/config.c
> index 32c0b2c..d73b090 100644
> --- a/config.c
> +++ b/config.c
> @@ -429,13 +429,11 @@ static int git_config_maybe_bool_text(const char *name, const char *value)
>  
>  int git_config_maybe_bool(const char *name, const char *value)
>  {
> -	int v = git_config_maybe_bool_text(name, value);
> +	long v = git_config_maybe_bool_text(name, value);
>  	if (0 <= v)
>  		return v;
> -	if (!strcmp(value, "0"))
> -		return 0;
> -	if (!strcmp(value, "1"))
> -		return 1;
> +	if (git_parse_long(value, &v))
> +		return !!v;
>  	return -1;
>  }
>  
> -- 
> 1.7.3.4.761.g98ad5