emacs-devel
[Top][All Lists]
Advanced

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

Re: other/9818 (Was: Re: Emacs - MinGW32 on W2K)


From: Wolfgang Bangerth
Subject: Re: other/9818 (Was: Re: Emacs - MinGW32 on W2K)
Date: Mon, 24 Feb 2003 10:03:30 -0600 (CST)

> A reproducible recepie for a crash:
> - Launch emacs
> - Load library "recentf" (load-library "recentf")
> - Set the recentf mode (recentf-mode t)
> - M-x recentf-open-files
> CRASH (Emacs Abort Dialog: A fatal error has occured!)

It is entirely possible that this is due to a bug in gcc. However, please 
note that it happens _somewhere_ inside Emacs, a package of several 
100,000 lines of code. Since gcc developers are not usually acquainted 
with large packages, it is standing gcc policy [1] to ask submitters to 
try to narrow down such problems. What we would need is what we call a 
self-contained testcase: one file with one or more functions where the 
miscompilation happens.

I understand that this is placing a burden on submitters of bug reports. 
However, it is virtually impossible to find a miscompilation if you don't 
know what a code does -- you just don't know what expected and wrong 
effects of functions are.

The usual procedure for reducing bugs like yours is to try compiling the 
files which you suspect might be miscompiled with a different compiler and 
see whether this helps. If you have found the one file that is 
miscompiled, try to figure out which function inside it might be the 
problem and try to successively reduce the size of the problem from
"emacs-size" to something that might fit into a few 100 lines of code at 
most.

Sorry for not being of more help,
  Wolfgang

[1] Please see http://gcc.gnu.org/bugs.html

-------------------------------------------------------------------------
Wolfgang Bangerth             email:            address@hidden
                              www: http://www.ticam.utexas.edu/~bangerth/






reply via email to

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