I wanted to be able to output to a file other than gmon.out, and this
environment variable did the trick. However the manual didn't tell me about it
-- I had to go rummaging through the code. I will be following up with a patch.
--
Summary: update gprof manual with note about GMON_OU
--- Additional Comments From benasselstine at gmail dot com 2007-03-26
20:49 ---
Created an attachment (id=1652)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=1652&action=view)
add note about GMON_OUT_PREFIX to the manual (first iteration)
Here's my first try at a patch that fi
I am compiling and linking C for an ATmega2560. When I use the default options,
the resulting code works fine. When I try to remove unused trampoline entries
by placing -W,--relax in the ld options, the resulting code seems to have bad
subroutine calls.
Not all subroutine calls are bad, nor are a
--
What|Removed |Added
CC||eweddington at cso dot atmel
||dot com
http://sourceware.org/bu
--- Additional Comments From wilson at specifix dot com 2007-03-27 01:23
---
This looks right to me.
The bug is curious, but apparently we only handle or.andcm and and.orcm because
these are the only parallel compares that gcc is smart enough to emit. So when
we added support for these
