-fomit-frame-pointer lossage FYI

-fomit-frame-pointer lossage FYI

[Mike Mattie]

[-- Attachment #1.1: Type: text/plain, Size: 815 bytes --]

Hello,

I was going to report a problem with laggy scrolling in emacs through the usual channels. While rebuilding
emacs to narrow down the problem I discovered that my package manager was building with -fomit-frame-pointer.
I like debugging so I killed the flag. To my joy it eliminated my scrolling problem entirely. The profiling
graphs differed significantly between emacs compiled with and without the flag. It was not simply the hot-spots
which were similar but it looked like the call-paths changed.

I am using gcc 4.2.2, emacs 23.0.60, on i686-pc-gnu-linux. This is not a problem report, just a FYI that
-fomit-frame-pointer may cause serious problems with emacs. If this is not already known to the emacs developers
then I hope this tidbit helps other users/developers.

Cheers,
Mike mattie

[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-devel

Re: -fomit-frame-pointer lossage FYI

[Richard Stallman]

    I was going to report a problem with laggy scrolling in emacs
    through the u= sual channels. While rebuilding emacs to narrow
    down the problem I discovered that my package manager was b=
    uilding with -fomit-frame-pointer.  I like debugging so I killed
    the flag. To my joy it eliminated my scrolling= problem entirely.

That probably means a real bug in either Emacs or GCC.
Unfortunately there is no way to know which except by
tracking down the specific cause.

Could you try to do this?

The first step is to try compiling various subsets of Emacs' files
with -fomit-frame-pointer, and see which file is the cause of the
problem.  Then by splitting that file at various points you can find
out which function is the cause of the problem.  Once we know that,
maybe we can spot the problem by eyeballing it.