This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH 3/3] Add a test case for the jit-reader interface.
- From: Jan Kratochvil <jan dot kratochvil at redhat dot com>
- To: Sanjoy Das <sanjoy at playingwithpointers dot com>
- Cc: gdb-patches at sourceware dot org
- Date: Wed, 9 Jan 2013 13:37:06 +0100
- Subject: Re: [PATCH 3/3] Add a test case for the jit-reader interface.
- References: <1352048631-25042-1-git-send-email-sanjoy@playingwithpointers.com> <1352048631-25042-4-git-send-email-sanjoy@playingwithpointers.com> <20121201204751.GB22812@host2.jankratochvil.net> <CAMiUf7f5EXkxi7hD4Rm1bU=_tUUtB3vgLbvzB4S-qbWYgh4_fQ@mail.gmail.com> <20121210163800.GA24342@host2.jankratochvil.net> <CAMiUf7dcxE-nt7OYd2RrKTeMGBBb0-1Buod+KKsYrdu1JYJFTg@mail.gmail.com>
On Tue, 08 Jan 2013 21:40:02 +0100, Sanjoy Das wrote:
> > OK, that makes sense. But in such case please make there a clear separate
> > jit.h copy (with a different name and some added #ifdef/#error or something
> > like that protection so that if one mistakenly includes gdb/jit.h it gets
> > error-reported, with an explanation you made to me now above).
>
> Fixed.
> +#ifdef JIT_H
> +#error "We don't include jit.h directly since we'd like the jit-reader unit \
> + tests to break if we make ABI incompatible changes to the structures \
> + re-declared here."
> +#endif
OK, thanks for the explanation, it looks clear to me now.
> + gdb_test_no_output "jit-reader-load ${jit_reader_bin}"
Do not use full binary path in test name, those differ across testsuite runs.
That is use for example:
gdb_test_no_output "jit-reader-load ${jit_reader_bin}" "jit-reader-load"
Therefore these three parts are OK for check-in.
There was a never-replied mail but that suggestion is not required for this
check-in.
Re: [PATCH 1/3] Fix segfault when unwinding JIT frames using a custom reader.
http://sourceware.org/ml/gdb-patches/2012-12/msg00007.html
Message-ID: <20121201202522.GA22812@host2.jankratochvil.net>
Thanks,
Jan