Possible errors in call to output_operand_lossage?

Philipp Thomas pthomas@suse.de
Sat Feb 16 14:15:00 GMT 2002

On Fri, 15 Feb 2002 11:32:43 -0800, you wrote:

>On Fri, Feb 15, 2002 at 04:41:28PM +0100, Philipp Thomas wrote:
>> -	  output_operand_lossage ("invalid %%-code");
>> +	  output_operand_lossage ("invalid %%%%-code");
>The original was correct; likewise with 81 other instances.
>That was the point of the change, no?  ;-)

Now you got me really confused ;-) Seems like I misunderstood the code
completely :(

These messages were output verbatim, i.e. the above would have been
output as

   invalid %%-code

Now the string passed to o_o_l *will* go through vasprintf and would
thus be output as

   invalid %-code

if the percentage signs weren't doubled.

Your comment implies that the original use was incorrect as printf
style was assumed by whoever wrote the code and that things like %%R
where meant to be output as %R. If so, only those messages that use a
single '%' would have to be changed, right?


More information about the Gcc-bugs mailing list