🚀 go-pugleaf

I received a segmentation violation when attempting to read a file
which had a 2.1 Mbyte symbol table.  A backtrace revealed the culprit
to be the alloca() call near line 1500 of dbxread.c in
symbol_file_command().  I'm not sure whether this is a gdb problem or
a kernel bug.  What seems to be happening is the kernel is generating
a segmentation violation when the stack is extended in alloca.
Evidently it assumes a 2.1 Mbyte stack extension all at once really
means the program is out-to-lunch.  Another possibility is that there
is a kernel parameter that needs to be tuned for this type of
application.  At any rate, changing the alloca to malloc fixes my
problem but also requires a free at the end of the function.

--
 |  Glenn R. Engel
 |  Hewlett-Packard 
 |  (206) 335-2066
 |  glenne%hplsla@hplabs.hp.com