Mail Archives: djgpp/1998/04/17/09:17:09
Andris Pavenis (LU Astro.) (pavenis AT lanet DOT lv) wrote:
> On Fri, 17 Apr 1998, Paul Derbyshire wrote:
> > gcc -c foo.cc -o foo.o -O6 -mpentium -Wall -Werror -fno-implicit-templates
> > c:/djgpp/lang/cxx/stl_rope.h: In function `class
> > rope<__wchar_t,__default_alloc_template<false,0> > &)':
> > c:/djgpp/lang/cxx/stl_rope.h:1991: virtual memory exhausted
> > What the hell is wrong with the cc1plus.exe in pgcc 1.0.1?!?!?!
> > There is NO WAY IN HELL it is using 64 entire megabytes and then
> > trying to get more! What is REALLY wrong and why doesn't
> > cc1plus.exe issue a more accurate and informative error message???
I reported similar behaviour two weeks ago I think, but nobody
confirmed nor denied it. The combination -O -Wall is doing it, I
noticed. I still don't know if it's a bug or gcc *really* needs this
much memory (over 100Mb with my 3-line program).
> > I would like to have some feedback about this and an estimate for
> > when this extremely serious bug will be fixed, ASAP, since a
> > project I'm working on is completely stalled until I have a
> > cc1plus.exe that will faithfully compile my code under the sole
> > condition that the code be syntactically correct.
Try without -Wall, it helped for me.
> Perhaps You can complain Micro$oft that only 64MB are available for
> DOS session. Try quitting Windows and using CWSDPMI and have enough
> swap space on HD. Your example really eats much memory. I tested
> this with pgcc-1.0.2 under Linux. As I tested that the maximal
> amount was about 30Mb. Perhaps DJGPP version of gcc eats more...
I don't think this is the problem. CWSDPMI have me the same problem.
also enlarging the stack and/or heapsize didn't help. I understand it
needs memory, but *that* much?
--
Groeten, Michel. http://www.cs.vu.nl/~mdruiter
____________
\ /====\ / "You know, Beavis, you need things that suck,
\/ \/ to have things that are cool", Butt-Head.
- Raw text -