This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: begin_function debug hook.


With the patch above, the h8300 port does not build any more.

gcc -c -DCROSS_COMPILE -DIN_GCC    -O2 -fno-strength-reduce -ggdb -save-temps -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wtraditional -pedantic -Wno-long-long  -DHAVE_CONFIG_H    -I. -I. -I/home/kazu/gnu/gcc/gcc-head/gcc/gcc -I/home/kazu/gnu/gcc/gcc-head/gcc/gcc/. -I/home/kazu/gnu/gcc/gcc-head/gcc/gcc/config -I/home/kazu/gnu/gcc/gcc-head/gcc/gcc/../include /home/kazu/gnu/gcc/gcc-head/gcc/gcc/dbxout.c -o dbxout.o
/home/kazu/gnu/gcc/gcc-head/gcc/gcc/dbxout.c: In function `dbxout_source_line':
/home/kazu/gnu/gcc/gcc-head/gcc/gcc/dbxout.c:575: `sdb_begin_function_line' undeclared (first use in this function)
/home/kazu/gnu/gcc/gcc-head/gcc/gcc/dbxout.c:575: (Each undeclared identifier is reported only once
/home/kazu/gnu/gcc/gcc-head/gcc/gcc/dbxout.c:575: for each function it appears in.)

I am wondering if this has anything to do with making
sdb_begin_function_line static.  A quick grep shows that
sdb_begin_function_line is referred to from many targets in the
definitions of of ASM_OUTPUT_SOURCE_LINE.  In turn,
ASM_OUTPUT_SOURCE_LINE is used in dbxout.c, which has no access to

Could somebody take a look?


Kazu Hirata

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]