Re: [PATCH v2] Fix safe_create_leading_directories() for Windows

[Junio C Hamano <gitster@pobox.com>]

Junio C Hamano <gitster@pobox.com> writes:

> Junio C Hamano <gitster@pobox.com> writes:
>
>> This has a bit of conflict with another topic in flight; I think I
>> resolved it correctly, but please double check.  The following is
>> how it would apply on top of 'pu'.
>>
>>  sha1_file.c | 9 +++++----
>>  1 file changed, 5 insertions(+), 4 deletions(-)
>>
>> diff --git a/sha1_file.c b/sha1_file.c
>> index 131ca97..9e686eb 100644
>> --- a/sha1_file.c
>> +++ b/sha1_file.c
>> @@ -113,11 +113,12 @@ int safe_create_leading_directories(char *path)
>>  
>>  	while (!retval && next_component) {
>>  		struct stat st;
>> -		char *slash = strchr(next_component, '/');
>> -
>> -		if (!slash)
>> +		char *slash = next_component;
>> +		while (!is_dir_sep(*slash))
>
> Gaah; we need to check for the end of string here, i.e.
>
> 		while (*slash && !is_dir_sep(*slash))
>
> will be what I'll queue on 'pu' for today.
>
>> +			++slash;
>> +		if (!*slash)
>>  			return 0;
>> -		while (*(slash + 1) == '/')
>> +		while (is_dir_sep(*(slash + 1)))
>>  			slash++;
>>  		next_component = slash + 1;
>>  		if (!*next_component)

Another thing I noticed (but I won't fix it up myself today, as I am
deep into today's integration cycle already) is that we temporarily
replace the slash with NUL and then restore them before we return,
but the restoration is done with the slash.  If we were to go in the
direction of this patch, you may want to update that one to use
whatever dir-sep was used in the input string.