bug-gawk
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: 32-bit profiling counts?


From: Andrew J. Schorr
Subject: Re: 32-bit profiling counts?
Date: Mon, 8 Jun 2020 16:40:50 -0400
User-agent: Mutt/1.5.21 (2010-09-15)

On Mon, Jun 08, 2020 at 01:22:13PM -0600, arnold@skeeve.com wrote:
> I went with long long. :-)  The patch will get pushed to the repo
> shortly.

That was brave. :-) But unless I'm confused, you missed at least one
printf format here:

profile.c:1046:                         fprintf(prof_fp, " # %ld", 
ip1->exec_count);

Also, here's another thought: if you're going to dive in and make this
change, might it be safer to apply the attached patch?
This way, we leave the other uses of dl unchanged and switch
only exec_count to use long long. I'm not sure it matters, but it seems
a bit less risky. And why incur a potential performance hit for the
other uses of dl? It passes "make check" and "make valgrind".

Regards,
Andy

Attachment: exec_count_long_long.patch
Description: Text document


reply via email to

[Prev in Thread] Current Thread [Next in Thread]