Re: SOLVED: Emacs (all versions) dumps core when built against XFree86 4.3

[David Kastrup]

Hin-Tak Leung <hintak_leung@yahoo.co.uk> writes:

> Symptom:  emacs dump cores at XtInitializeWidgetClass () with
>               'Fatal Error (11).Segmentational Fault'
> Solution: Adding '-z nocombreloc' to LDFLAGS
> Reason:   'combreloc' has become the default for recent GNU ld, which
>               breaks the unexec/undump of both emacs and xemacs
>               (all versions).
> 
> Many after-thoughts:
> 
> (1) The problem has been very recently reported against CVS
>      version of GNU emacs in March 2003 (i.e. two months ago),
>      and even caught Mr Stallman's attention briefly,
>      no solution is known to the GNU Emacs people; and
>      yet the xemacs people apparently had the fix integrated to
>      their ./configure for almost a year now...

Those are different projects with completely different ./configure
files.  It can hardly be expected that they synchronize arbitrary
problem reports unless somebody accidentally happens to work on both
projects.  But overlap of core developers is very small since there
is a _reason_ for both being separate projects.

> (2) The GNU emacs people/community has not been at all helpful...

And you have not exactly been an icon of amicability either, posting
long tirades about what you want and don't want to hear, how people
should behave and so on.

>      I guessed the origin of the problem myself, made my own
>      suggestion to do a gdb back trace, did my own search for any
>      problem associated with 'XtInitializeWidgetClass'...

Emacs comes with error reporting instructions.  You make it sound as
if it was an insolence that you had to properly report an error
before somebody could take a look at it.  But you started tirading
almost immediately before somebody even had a chance to tell you
anything, and every post contained long complaints about how bad
people were because they dared to mention that you were overstepping
basic politeness.

> (3) Sadly, judging from the reaction to the same problem in both
>      archives, I have to say the xemacs community probably are
>      more "free" (in the beer sense) than the GNU emacs community...

So what payment was exacted from you?

>      xemacs people gave sound technical advices and solutions, GNU
>      emacs people gave lectures on "employer/employee
>      relationships"...

I am sure that if somebody had noticed your particular problem
before, it would have got mentioned here, as well.  You were just
lucky that the problem was already known there and so people had not
to go through the notions of trying to get you to deliver any useful
information.  Perhaps take a look at the XEmacs archives and compare
how the people there reported the original problem in contrast to how
you did here.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum