This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: [RFA:] Skip symbol after warning symbol in gas/write.c:write_object_file
- From: Ian Lance Taylor <ian at airs dot com>
- To: Hans-Peter Nilsson <hans-peter dot nilsson at axis dot com>
- Cc: binutils at sources dot redhat dot com
- Date: 06 Feb 2005 13:31:07 -0500
- Subject: Re: [RFA:] Skip symbol after warning symbol in gas/write.c:write_object_file
- References: <200502061503.j16F30FH024456@ignucius.se.axis.com>
Hans-Peter Nilsson <hans-peter.nilsson@axis.com> writes:
> I'm not sure whether the symbol following the warning stab
> should be used other than picking the name, but I guess it
> actually should, so there's an xfailed test-case for that (or
> two if you count ELF which is xfailed for not supporting the
> construct in its stabs support).
I don't think we should use the symbol following the a.out warning
stab for anything. The warning stab should be followed by an
undefined symbol, which is what we should issue the warning about.
That undefined symbol should not be otherwise used for anything. Note
that in typical uses, the symbol will be defined somewhere else in the
object file in any case.
Let me know if I am missing something.
> gas:
> * write.c (write_object_file): Recognize warning-symbol construct
> and skip object- and target- handling for the second symbol.
>
> ld/testsuite:
> * ld-cris/globsymw2.s, ld-cris/globsymw3.s: New files.
> * ld-cris/warn3.d, ld-cris/warn4.d, ld-cris/warn5.d,
> ld-cris/warn6.d: New tests.
This is OK, if you adjust the testsuite according to the above.
Ian