This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [RFC patch]: Adjust the use of 'long' type in dwarf2.h header
2011/2/21 Mark Kettenis <mark.kettenis@xs4all.nl>:
>> Date: Mon, 21 Feb 2011 16:25:35 +0100
>> From: Kai Tietz <ktietz70@googlemail.com>
>>
>> 2011/2/21 Kai Tietz <ktietz70@googlemail.com>:
>> > 2011/2/21 Pierre Muller <pierre.muller@ics-cnrs.unistra.fr>:
>> >>> Yes, I saw this too and it is a thorn here. But the issue about a
>> >>> transformation function (like dwarf_vma) is that we have memory leaks
>> >>> - as this function is to be used on some printf's more then once - as
>> >>> static buffer would be overriden. Well printfs can be splitted into
>> >>> smaller parts, but this then would make localization even harder.
>> >>
>> >> ?In GDB code, this is solved in utils.c source by the use of the
>> >> get_cell function, which rotates over 16 char arrays of size 50 each
>> >> currently, in dwarf.c 4 arrays of 16 char (as long as longest address is
>> >> 64-bit)
>> >> for this dwarf_vma function would probably be enough, no?
>> >>
>> >>
>> >> Pierre Muller
>> >> GDB pascal language maintainer
>> >
>> > Yes, this sounds ok. I think maximum here are right now 3 dwarf_vma
>> > prints within one printf. So it should be ok. I'll add here such a
>> > rotating function. I don't assume we need here to handle
>> > multi-threading, so rotation code can be pretty simple.
>> >
>> > Kai
>> >
>>
>> So this version uses new function dwarf_vma_print function for
>> printf-messages, which are getting localized.
>
> dwarf_vma_print is a really bad name for that function since it
> doesn't print anything.
>
well, it doesn't output, but it uses sprintf. So what's about
dwarf_vmatoa instead?