* Re: broken 970929 snapshot
@ 1997-10-02 21:56 Aaron Jackson
1997-10-02 23:01 ` Robert Lipe
0 siblings, 1 reply; 6+ messages in thread
From: Aaron Jackson @ 1997-10-02 21:56 UTC (permalink / raw)
To: egcs
Ok, the code for enquire.c is of course fine, a gcc -c enquire.c proves
this. looking through the logfile of the build it would seem that
enquire is the first bit of code compiled with the first stage
compiler, so that means that the first stage compiler wasn't built
properly (pretty obvious). I tried the offending command again, this
time with the verbose option (see below). Any help, hints, comments?
Aaron Jackson jackson@msrce.howard.edu
Reading specs from ./specs
gcc version egcs-2.90.11 970929 (gcc2-970802 experimental)
./cpp -lang-c -v -I./include -I. -isystem ./include -undef -D__GNUC__=2 -D__GNU
C_MINOR__=90 -D__ANSI_COMPAT -DMIPSEL -DR3000 -DSYSTYPE_BSD -D_SYSTYPE_BSD -Dbsd
4_2 -Dhost_mips -Dmips -Dultrix -Dunix -D__ANSI_COMPAT -D__MIPSEL__ -D__R3000__
-D__SYSTYPE_BSD__ -D_SYSTYPE_BSD -D__bsd4_2__ -D__host_mips__ -D__mips__ -D__ult
rix__ -D__unix__ -D__MIPSEL -D__R3000 -D__SYSTYPE_BSD -D__bsd4_2 -D__host_mips -
D__mips -D__ultrix -D__unix -Asystem(unix) -Asystem(bsd) -Acpu(mips) -Amachine(m
ips) -g -D__LANGUAGE_C -D_LANGUAGE_C -DLANGUAGE_C -D__SIZE_TYPE__=unsigned int -
D__PTRDIFF_TYPE__=int -DIN_GCC -DNO_MEM -DNO_LONG_DOUBLE_IO ./enquire.c /usr/tmp
/cca04446.i
GNU CPP version egcs-2.90.11 970929 (gcc2-970802 experimental) [AL 1.1, MM 40] D
ECstation running ultrix
#include "..." search starts here:
#include <...> search starts here:
include
.
include
/usr/local/include
/usr/local/mips-dec-ultrix4.4/include
/usr/include
End of search list.
./cc1 /usr/tmp/cca04446.i -quiet -dumpbase enquire.c -g -O0 -version -o /usr/tm
p/cca04446.s
GNU C version egcs-2.90.11 970929 (gcc2-970802 experimental) (mips-dec-ultrix4.4
) compiled by GNU C version 2.7.2.3.f.1.
as -g -nocpp -v -o enquire.o /usr/tmp/cca04446.s
/usr/lib/cmplrs/cc/as0 -v -G 8 -EL -g2 -O1 /usr/tmp/cca04446.s -o /tmp/ctmaa0444
9 -t /tmp/ctmsta04449
as0: overflow setmode farewell croak xmalloc maximum_int main eek_a_bug describe
i_define u_define f_define floor_log ceil_log exponent fake_f_rep f_rep bitpatt
ern endian missing fmissing false type_of ftype_of typerr ftyperr promotions che
ck_defines cprop basic sprop usprop fStore fSum fDiff fMul fDiv fSelf fCheck fpr
op efprop iprop uiprop dStore dSum dDiff dMul dDiv dSelf dCheck dprop edprop lpr
op ulprop ldStore ldSum ldDiff ldMul ldDiv ldSelf ldCheck ldprop eldprop
as0: Error: ./enquire.c, line 1: undefined symbol in expression
.word $LECIE1-$LSCIE1
as0: Error: ./enquire.c, line 1: undefined symbol in expression
.word $LEFDE1-$LSFDE1
as0: Error: ./enquire.c, line 1: Symbol must have absolute value: $LFB1
.word $LFE1-$LFB1
as0: Error: ./enquire.c, line 1: Symbol must have absolute value: $LFB1
.word $LCFI0-$LFB1
as0: Error: ./enquire.c, line 1: Symbol must have absolute value: $LCFI0
.word $LCFI1-$LCFI0
...700 lines or so...
as0: Error: ./enquire.c, line 1: Symbol must have absolute value: $LCFI246
.word $LCFI247-$LCFI246
as0: Error: ./enquire.c, line 1: Symbol must have absolute value: $LCFI247
.word $LCFI248-$LCFI247
4.8u 6.9s 0:12 96% 99+760k 0+29io 4pf+0w
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: broken 970929 snapshot
1997-10-02 21:56 broken 970929 snapshot Aaron Jackson
@ 1997-10-02 23:01 ` Robert Lipe
0 siblings, 0 replies; 6+ messages in thread
From: Robert Lipe @ 1997-10-02 23:01 UTC (permalink / raw)
To: Aaron Jackson; +Cc: egcs
Disclaimer: I don't think I've ever seen an Ultrix system in my life.
I have been known to tinker with MIPS parts, but I think I recognize
the symptoms in your case.
> as0: Error: ./enquire.c, line 1: undefined symbol in expression
> .word $LECIE1-$LSCIE1
I think you're getting nailed by the same thing that got the OpenServer
COFF port. You're getting DWARF2 exceptions (you'd have to look at the
generated assembly to be sure) and they're emitting constructs like
this, but the assembler is unable to handle the subtraction of the
forward references. You can roam through the archives for conversations
on this topic.
To prove or disprove this theory, try the following, in decreasing ease:
Manually compile with file with either -fno-exceptions which
turns exceptions off completely.
Compile the file with -fsjlj-exceptions to force the use of
the setjmp/longjmp exceptions.
Install GNU as.
If either of these work, this is very likely your problem. From here,
your options would include:
Using the GNU assembler for your target.
Experimenting with disabling the DWARF2 exceptions by adding a
#define DWARF2_UNWIND_INFO 0
in your tm.h file
(remember to make clean and rebuild) This is the
approximate equivalent of the -fsjlj-exceptions flag.
Getting the dwarf2 code to not emit forward refs.
--
Robert Lipe http://www.dgii.com/people/robertl robertl@dgii.com
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: broken 970929 snapshot
@ 1997-10-03 13:24 jackson
0 siblings, 0 replies; 6+ messages in thread
From: jackson @ 1997-10-03 13:24 UTC (permalink / raw)
To: robertl; +Cc: egcs
I 'made distclean', './configure --verbose >&logfile', added
#define DWARF2_UNWIND_INFO 0 to gcc/tm.h, 'make bootstrap-lean >>& logfile &'
and then went to sleep. Imagine my surprise when I get to work and 'tail -f
logfile' gave me the same errors as before. Is there any other way to disable
the DWARF2 exceptions?
Aaron Jackson jackson@msrce.howard.edu
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: broken 970929 snapshot
@ 1997-10-02 23:35 Aaron Jackson
1997-10-03 7:28 ` Robert Lipe
0 siblings, 1 reply; 6+ messages in thread
From: Aaron Jackson @ 1997-10-02 23:35 UTC (permalink / raw)
To: jackson, robertl; +Cc: egcs
>Disclaimer: I don't think I've ever seen an Ultrix system in my life.
You might not ever see one in the future since DEC killed them off...
>To prove or disprove this theory, try the following, in decreasing ease:
> Manually compile with file with either -fno-exceptions which
> turns exceptions off completely.
> Compile the file with -fsjlj-exceptions to force the use of
> the setjmp/longjmp exceptions.
Both flages fixed the problem.
> Install GNU as.
I'll wait a little before I try this option ;)
>You can roam through the archives for conversations on this topic.
I guess I should read more of the posts...but I tend to ignore
subjects that I know nothing about.
>Experimenting with disabling the DWARF2 exceptions by adding a
> #define DWARF2_UNWIND_INFO 0
> in your tm.h file
I'll give it a try. How about on the command line? I'll have
to try that tomrrow afternoon. Sleep now!
Aaron Jackson jackson@msrce.howard.edu
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: broken 970929 snapshot
1997-10-02 23:35 Aaron Jackson
@ 1997-10-03 7:28 ` Robert Lipe
0 siblings, 0 replies; 6+ messages in thread
From: Robert Lipe @ 1997-10-03 7:28 UTC (permalink / raw)
To: Aaron Jackson; +Cc: egcs
> >Disclaimer: I don't think I've ever seen an Ultrix system in my life.
> You might not ever see one in the future since DEC killed them off...
:-)
> >To prove or disprove this theory, try the following, in decreasing ease:
> > Manually compile with file with either -fno-exceptions which
> > turns exceptions off completely.
> > Compile the file with -fsjlj-exceptions to force the use of
> > the setjmp/longjmp exceptions.
>
> Both flages fixed the problem.
OK. This means the code your assembler is finding offensive is
related to exceptions - just like the situation I ran into recently.
> > Install GNU as.
> I'll wait a little before I try this option ;)
It really shouldn't be too painful. Ultrix is listed as a supported
target in current GAS.
> >You can roam through the archives for conversations on this topic.
>
> I guess I should read more of the posts...but I tend to ignore
> subjects that I know nothing about.
If we all did that, you'd be a little less close to getting your target
to work again, now wouldn't you? :-)
> >Experimenting with disabling the DWARF2 exceptions by adding a
> > #define DWARF2_UNWIND_INFO 0
> > in your tm.h file
>
> I'll give it a try. How about on the command line? I'll have
> to try that tomrrow afternoon. Sleep now!
Don't know of a way to do that. If you specify the flags described
above you'll force either no exceptions or non-dwarf2 exceptions.
As a kludge, you can change it from '1' to '0' at the end of gcc/defaults.h.
> Aaron Jackson jackson@msrce.howard.edu
>
>
>
--
Robert Lipe http://www.dgii.com/people/robertl robertl@dgii.com
^ permalink raw reply [flat|nested] 6+ messages in thread
* broken 970929 snapshot
@ 1997-10-02 16:51 Aaron Jackson
0 siblings, 0 replies; 6+ messages in thread
From: Aaron Jackson @ 1997-10-02 16:51 UTC (permalink / raw)
To: egcs
It would appear that some recent changes to enquire.c breaks egcs
for ultrix 4.4 since I can no longer get it to build even the first stage.
I used a 'configure; make bootstrap-lean' on a freashly unpaced 970929
distribution using gcc 2.7.2.3 and g77 0.5.21 and got the following
error messages. Since I am only an amateur hacker I have absolutly no
idea what these errors mean. Any comments, help or suggestions?
Aaron Jackson jackson@msrce.howard.edu
...chugging along until...
rm -f include/limits.h
cp xlimits.h include/limits.h
chmod a+r include/limits.h
rm -f include/README
cp ./README-fixinc include/README
chmod a+r include/README
touch stmp-int-hdrs
./xgcc -B./ -DIN_GCC -g -I./include -DNO_MEM -DNO_LONG_DOUBLE_IO -O0 -I.
-c ./enquire.c
as0: Error: ./enquire.c, line 1: undefined symbol in expression
.word $LECIE1-$LSCIE1
as0: Error: ./enquire.c, line 1: undefined symbol in expression
.word $LEFDE1-$LSFDE1
as0: Error: ./enquire.c, line 1: Symbol must have absolute value: $LFB1
.word $LFE1-$LFB1
as0: Error: ./enquire.c, line 1: Symbol must have absolute value: $LFB1
.word $LCFI0-$LFB1
as0: Error: ./enquire.c, line 1: Symbol must have absolute value: $LCFI0
.word $LCFI1-$LCFI0
as0: Error: ./enquire.c, line 1: Symbol must have absolute value: $LCFI1
.word $LCFI2-$LCFI1
as0: Error: ./enquire.c, line 1: Symbol must have absolute value: $LCFI2
...for 600 lines or so...
as0: Error: ./enquire.c, line 1: Symbol must have absolute value: $LFB63
.word $LCFI245-$LFB63
as0: Error: ./enquire.c, line 1: Symbol must have absolute value: $LCFI245
.word $LCFI246-$LCFI245
as0: Error: ./enquire.c, line 1: Symbol must have absolute value: $LCFI246
.word $LCFI247-$LCFI246
as0: Error: ./enquire.c, line 1: Symbol must have absolute value: $LCFI247
.word $LCFI248-$LCFI247
*** Error code 1
Stop.
*** Error code 1
Stop.
*** Error code 1
Stop.
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~1997-10-03 13:24 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1997-10-02 21:56 broken 970929 snapshot Aaron Jackson
1997-10-02 23:01 ` Robert Lipe
-- strict thread matches above, loose matches on Subject: below --
1997-10-03 13:24 jackson
1997-10-02 23:35 Aaron Jackson
1997-10-03 7:28 ` Robert Lipe
1997-10-02 16:51 Aaron Jackson
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).