bug or feature?: egcs dies too late on syntax error

Markus Werle markus@lufmech.rwth-aachen.de
Wed Nov 22 02:36:00 GMT 2000


Hi!

I wonder why egcs delays an exit when syntax errors occur.
If a syntax error occurs, no other message is of any interest to me.

In the case below with several lines of output the error message is
hard to find (see ---) and I first wondered why a warning about
multi-line-comments killed g++.

Btw.: How do I stop that annoying multi-line warning?


g++ -o /work/markus/OBJECTS/NewtonSolver.o -c -I.
-ftemplate-depth-1000  -Wall -Wno-comment     -mpa-risc-2-0  \
-DPETSc -DHAVE_MPI_COMM_F2C -DUSE_PETSC_DEBUG -DUSE_PETSC_LOG  \
-I. -I/opt/petsc-2.0.29.2/include -I/opt/mpi/include NewtonSolver.C
In file included from Differentiations.h:16,
                 from NewtonSpecialClasses.h:7,
                 from NewtonSolver.h:18,
                 from NewtonSolver.C:2:
LinearSolver.h:31:2: warning: #warning "I use PETSc now"
--------------error message appears at this point -----
Lumping.h:255: syntax error before "affects"
---------------------------------------------
In file included from NewtonSolver.h:22,
                 from NewtonSolver.C:2:
GalerkinDiscretizedPDE.h:364:1: warning: multi-line comment
In file included from NewtonSolver.C:3:
NewtonSolver.h:149:1: warning: multi-line comment
gmake: *** [NewtonSolver.o] Error 1


Markus

P.S.: Thank You for egcs on hpux-11.0 (seems to work fine since last
week):
Reading specs from
/work/markus/opt/egcs-001117/lib/gcc-lib/hppa2.0w-hp-hpux11.00/2.97/specs

Configured with: ../egcs/configure --prefix=/work/markus/opt/egcs-001117
--enable-languages=c++ --with-gnu-as --without-gnu-ld
--disable-libstdcxx-v3
gcc version 2.97 20001113 (experimental)





More information about the Gcc-bugs mailing list