Alex Riesen <raa.lkml@xxxxxxxxx> writes:
>> But I wonder if this is worth it (not this "more complete", but your
>> patch). We convert 16 or so instances of initialization for "no strdup"
>> list, but there are about the same number of "strdup" instances still
>> spelled out:
>>
>> $ git grep -e 'struct string_list [^ ]* = {.*'
>> $ git grep -e 'struct string_list [^ ]* = {.*1'
>>
>> Wouldn't it be more sensible to use this instead?
>>
>> #define STRING_LIST_INIT(pleasedup) { NULL, 0, 0, (pleasedup) }
>
> This begs for using strdup(string-to-dup) in the macro argument, which
> will not compile with ancient compilers which don't allow code in an
> initializer.
Err, one of us must be confused.
I was suggesting to rewrite things like these, found in builtin/remote.c
(add and rm):
struct string_list track = { NULL, 0, 0, 0 };
struct string_list branches = { NULL, 0, 0, 1 };
like so:
#define STRING_LIST_INIT(pleasedup) { NULL, 0, 0, (pleasedup) }
struct string_list track = STRING_LIST_INIT(0);
struct string_list branches = STRING_LIST_INIT(1);
I don't see strdup in macro argument anywhere.
The original patch in this thread proposed to do this instead:
#define STRING_LIST_INIT { NULL, 0, 0, 0 }
struct string_list track = STRING_LIST_INIT;
struct string_list branches = { NULL, 0, 0, 1 };
converting about half of initializers to the macro, and leaving the other
half spelled out.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html