[Bug fortran/36342] [4.4 Regression] Missing file name in compilation diagnostics of preprocessed fortran source
--- Comment #17 from burnus at gcc dot gnu dot org 2008-07-19 18:27 --- Subject: Bug 36342 Author: burnus Date: Sat Jul 19 18:27:00 2008 New Revision: 137985 URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=137985 Log: 2008-07-19 Tobias Burnus [EMAIL PROTECTED] PR fortran/36342 * scanner.c (load_file): Add argument to destinguish between true filename and displayed filename. (include_line,gfc_new_file): Adapt accordingly. Modified: trunk/gcc/fortran/ChangeLog trunk/gcc/fortran/scanner.c -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36342
[Bug fortran/36342] [4.4 Regression] Missing file name in compilation diagnostics of preprocessed fortran source
--- Comment #18 from burnus at gcc dot gnu dot org 2008-07-19 18:30 --- FIXED on the trunk (4.4) [4.3 is not effected]. -- burnus at gcc dot gnu dot org changed: What|Removed |Added Status|REOPENED|RESOLVED Resolution||FIXED http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36342
[Bug fortran/36342] [4.4 Regression] Missing file name in compilation diagnostics of preprocessed fortran source
-- rguenth at gcc dot gnu dot org changed: What|Removed |Added Priority|P3 |P4 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36342
[Bug fortran/36342] [4.4 Regression] Missing file name in compilation diagnostics of preprocessed fortran source
--- Comment #16 from burnus at gcc dot gnu dot org 2008-07-17 13:30 ---
The problem seems to be this code in scanner.c's gfc_new_file:
if (gfc_cpp_enabled ())
{
result = gfc_cpp_preprocess (gfc_source_file);
if (!gfc_cpp_preprocess_only ())
result = load_file (gfc_cpp_temporary_file (), true);
}
else
result = load_file (gfc_source_file, true);
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36342
[Bug fortran/36342] [4.4 Regression] Missing file name in compilation diagnostics of preprocessed fortran source
--- Comment #15 from laurynas dot biveinis at gmail dot com 2008-06-20 14:00 --- Junk output and valgrind errors should be fixed by a patch for PR fortran/34908 and PR fortran/36276. The /tmp/... file name issue needs additional fixing though. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36342
[Bug fortran/36342] [4.4 Regression] Missing file name in compilation diagnostics of preprocessed fortran source
--- Comment #11 from pault at gcc dot gnu dot org 2008-06-18 07:17 --- (In reply to comment #10) Reopened. Hah! OK - thanks Paul *with tail between his legs* -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36342
[Bug fortran/36342] [4.4 Regression] Missing file name in compilation diagnostics of preprocessed fortran source
--- Comment #12 from P dot Schaffnit at access dot rwth-aachen dot de 2008-06-18 07:20 --- Hi! Stupid question: is it related to http://gcc.gnu.org/ml/fortran/2008-06/msg00203.html ? Philippe -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36342
[Bug fortran/36342] [4.4 Regression] Missing file name in compilation diagnostics of preprocessed fortran source
--- Comment #13 from burnus at gcc dot gnu dot org 2008-06-18 15:43 --- (In reply to comment #10) Reopened. I can still observe the problem with trunk from 20080615 (r136827). Can you post some more debug information? As written, I cannot reproduce it here anymore with 4.4.0 20080616 trunk revision 136838 on x86-64 - not even with valgrind. (I used to be able to reproduce it.) I also tried several locales. Question: I get the following: /tmp/ccpoBhQR.f90:8: warning: 'junk_2' is used uninitialized in this function What puzzles me is that one gets a /tmp/... filename and not the filename one passed to gfortran -cpp. Is this expected? I thought that when using libcpp directly, no temporary file is written anymore. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36342
[Bug fortran/36342] [4.4 Regression] Missing file name in compilation diagnostics of preprocessed fortran source
--- Comment #14 from dfranke at gcc dot gnu dot org 2008-06-18 16:48 --- Question: I get the following: /tmp/ccpoBhQR.f90:8: warning: 'junk_2' is used uninitialized in this function What puzzles me is that one gets a /tmp/... filename and not the filename one passed to gfortran -cpp. Is this expected? I thought that when using libcpp directly, no temporary file is written anymore. Heisenbug it seems. On x86_64 (20080617) I get: $ gfortran-svn -Wall -O pr36342.F90 pr36342.F90: In function 'test': pr36342.F90:4: warning: 'junk_2' is used uninitialized in this function Running f951 as output by gfortran [...] -v gives: pr36342.F90: In function 'test': pr36342.F90:4: warning: 'junk_2' is used uninitialized in this function BUT, prefixing valgrind (otherwise identical to the call above) results in: /tmp/ccIl84Yu.f90: In function 'test': /tmp/ccIl84Yu.f90:4: warning: 'junk_2' is used uninitialized in this function In addition, valgrind shows multiple invalid-reads as quoted in earlier comments. $ gfortran-svn -v Using built-in specs. Target: x86_64-unknown-linux-gnu Configured with: /data/home/daniel/svn/gcc-svn/configure --prefix=/data/home/daniel/x86_64-unknown-linux-gnu/gcc --program-suffix=-svn --with-gmp=/data/home/daniel/x86_64-unknown-linux-gnu/gmp-4.2.1 --with-mpfr=/data/home/daniel/x86_64-unknown-linux-gnu/mpfr-2.2.1 --with-system-zlib --enable-bootstrap --enable-threads=posix --enable-shared --enable-version-specific-runtime-libs --enable-languages=c,fortran --disable-nls --disable-multilib --disable-libmudflap Thread model: posix gcc version 4.4.0 20080617 (experimental) (GCC) -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36342
[Bug fortran/36342] [4.4 Regression] Missing file name in compilation diagnostics of preprocessed fortran source
--- Comment #8 from burnus at gcc dot gnu dot org 2008-06-16 17:56 --- Mark as waiting. Can any of you try whether the bug still exists for you? -- burnus at gcc dot gnu dot org changed: What|Removed |Added Status|NEW |WAITING http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36342
[Bug fortran/36342] [4.4 Regression] Missing file name in compilation diagnostics of preprocessed fortran source
--- Comment #9 from pault at gcc dot gnu dot org 2008-06-16 20:22 --- (In reply to comment #8) Mark as waiting. Can any of you try whether the bug still exists for you? This works for me - x86_ia64/Fedora8 - gcc version 4.4.0 20080615 Taking a risk, given Tobias' report - marking as fixed. Paul -- pault at gcc dot gnu dot org changed: What|Removed |Added Status|WAITING |RESOLVED Resolution||FIXED http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36342
[Bug fortran/36342] [4.4 Regression] Missing file name in compilation diagnostics of preprocessed fortran source
--- Comment #10 from dfranke at gcc dot gnu dot org 2008-06-16 21:49 --- Reopened. I can still observe the problem with trunk from 20080615 (r136827). $ gfortran-svn -O -Wall ~/pr/pr36342.F90 #382;: In function 'test': hqë·#353;:4: warning: 'junk_2' is used uninitialized in this function In addition, the invalid reads are still shown by valgrind. FX, could you have a look which part of your patch (r134992, one of the wide-char patches I think) might have introduced the problem? -- dfranke at gcc dot gnu dot org changed: What|Removed |Added Status|RESOLVED|REOPENED Resolution|FIXED | http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36342
[Bug fortran/36342] [4.4 Regression] Missing file name in compilation diagnostics of preprocessed fortran source
--- Comment #7 from burnus at gcc dot gnu dot org 2008-06-13 21:19 --- Works for me. Possibly fixed by PR preprocessor/36479. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36342
[Bug fortran/36342] [4.4 Regression] Missing file name in compilation diagnostics of preprocessed fortran source
-- jsm28 at gcc dot gnu dot org changed: What|Removed |Added Target Milestone|--- |4.4.0 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36342
