This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH 0/7] first batch of test suite updates
- From: Pedro Alves <palves at redhat dot com>
- To: Tom Tromey <tromey at redhat dot com>
- Cc: gdb-patches at sourceware dot org
- Date: Thu, 20 Jun 2013 17:06:17 +0100
- Subject: Re: [PATCH 0/7] first batch of test suite updates
- References: <1371675821-9959-1-git-send-email-tromey at redhat dot com>
On 06/19/2013 10:03 PM, Tom Tromey wrote:
> I've been working on a long-term project to fully parallelize the gdb
> test suite. The end goal is that it should be possible to run each
> .exp file in parallel.
Looking forward.
> This project has many parts; and the full patch is not quite finished
> (and is also very large). I wanted to start sending out some of the
> more straightforward changes, mostly so that I wouldn't have to
> constantly rebase them. I think these changes are reasonable in their
> own right.
Agreed.
>
> This series holds all the "easy" conversions from gdb.base. This is
> just straightforward changes to use standard_testfile and
> standard_output_file, plus also prepare_for_testing or clean_restart
> as well.
>
> This series fixes a few spots where .exp files did not interact well.
> For example, cases where the tests used the same executable name.
>
> This regression tests cleanly; however since some file names changed
> there are a few minor output changes:
>
> Missing tests:
> gdb.base/corefile.exp: args: -core=coremaker.core: PASS
> gdb.base/corefile.exp: args: execfile -core=coremaker.core: PASS
> gdb.base/info-proc.exp: core break.gcore: PASS
>
> New tests:
> gdb.base/corefile.exp: args: -core=corefile.core: PASS
> gdb.base/corefile.exp: args: execfile -core=corefile.core: PASS
> gdb.base/default.exp: set the history filename: PASS
> gdb.base/info-proc.exp: core info-proc.gcore: PASS
>
> I don't consider this to be a problem.
Agreed.
>
> I split the patch up into 7 roughly equal-sized pieces based on file
> name.
>
> After this series, gdb.base is still not parallel-clean. There are
> some lurking "trickier" cases I omitted from this series, as they
> depend on new testsuite/lib infrastructure. I plan to work through
> all the easy/obvious bits first, before starting to upstream that
> stuff.
I think I saw places in the context of the patches that could be
made to use build_executable instead of gdb_compile, and didn't look
tricky, but, this is all good forward progress as is!
Thanks a lot for doing this.
I tried skimming the series, but got extremely bored after
patch #1. :-)
I'm all for this.
--
Pedro Alves