Dear Richard,
I am having problems building CLN 1.1.4 with the newly released GCC 3.1. I attach below the signature the final part of the make log. Any idea what the problem may be? All the best
Roberto
Hi,
There are some problems with GCC 3.1. Bruno has made some fixes in CVS but that won't compile at all until I commit a small patch I prepared last week. I still have to ponder that patch, though. As soon as I have done this and run the check on my computer farm I will release the maintainance release CLN 1.1.5. Please be patient for a couple of hours or days. You'll hear from me then...
Regards -richy.
On Fri, 24 May 2002, Roberto Bagnara wrote: [...]
Any idea what the problem may be?
In hindsight, no, not really. Could you please try the current CVS version of CLN and tell me if it works? It should, now. If it does I would like to release it this week. Just go to GiNaC's download instructions and issue the usual CVS commands, just replacing GiNaC with cln.
Regards -richy.
On Fri, 24 May 2002, Roberto Bagnara wrote:
I am having problems building CLN 1.1.4 with the newly released GCC 3.1. I attach below the signature the final part of the make log.
Please give it a try with CLN 1.1.5. It should compile fine with any recend g++ from GCC. It's available at the usual places.
Regards -richy.
Richard B. Kreckel wrote:
On Fri, 24 May 2002, Roberto Bagnara wrote:
I am having problems building CLN 1.1.4 with the newly released GCC 3.1. I attach below the signature the final part of the make log.
Please give it a try with CLN 1.1.5. It should compile fine with any recend g++ from GCC. It's available at the usual places.
Hi Richard,
it seems we are still not there. Here is the final part of the make log for CLN 1.1.5 with GCC 3.1. All the best,
Roberto
Hi there,
I should have looked closer at your bugreport when you first submitted it. This is indeed a bug in the assembler 'gas', so in the binutils package, that was triggered by CLN and fixed by Bruno Haible a couple of weeks ago.
On Fri, 31 May 2002, Roberto Bagnara wrote: [...]
gcc -g -O2 -I../include -I../../cln-1.1.5/include -I../../cln-1.1.5/tests -I../../cln-1.1.5/src -I../../cln-1.1.5/src/base -I../../cln-1.1.5/src/float -I../../cln-1.1.5/src/float/lfloat -I../../cln-1.1.5/src/integer -I../../cln-1.1.5/src/base/random -I../../cln-1.1.5/src/base/digitseq -I../../cln-1.1.5/src/base/digit -I../src/base -I../../cln-1.1.5/src/base -c exam.s -o exam.o exam.s: Assembler messages: exam.s:216: Error: File number 53 already allocated make[1]: *** [exam.o] Error 1 make[1]: Leaving directory `/usr/local/distrib/buildcln/tests' make: *** [all] Error 2
It seems like binutils-2.12.1 does not fix the problem yet. Hence, you have three options: 1) Downgrade binutils to a version that has worked before 2) Wait for a new version of binutils. 3) Manually apply the following patch to gas/dwarf2dbg.c that was applied upstream by Nick Clifton nickc@cambridge.redhat.com on May 11 and build and install your own binutils: =================================================================== RCS file: /cvs/src/src/gas/dwarf2dbg.c,v retrieving revision 1.46 diff -c -3 -p -w -p -r1.46 dwarf2dbg.c *** gas/dwarf2dbg.c 3 May 2002 02:25:33 -0000 1.46 --- gas/dwarf2dbg.c 11 May 2002 10:26:52 -0000 *************** dwarf2_emit_insn (size) *** 290,296 **** return; loc_directive_seen = false;
! dwarf2_where (&loc); dwarf2_gen_line_info (frag_now_fix () - size, &loc); }
--- 290,301 ---- return; loc_directive_seen = false;
! /* Use the last location established by a .loc directive, not ! the value returned by dwarf2_where(). That calls as_where() ! which will return either the logical input file name (foo.c) ! or the physical input file name (foo.s) and not the file name ! specified in the most recent .loc directive (eg foo.h). */ ! loc = current; dwarf2_gen_line_info (frag_now_fix () - size, &loc); }
Regards -richy.
participants (2)
-
Richard B. Kreckel -
Roberto Bagnara