This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH 05/13] fix up gdb.xml
- From: Tom Tromey <tromey at redhat dot com>
- To: Pedro Alves <palves at redhat dot com>
- Cc: gdb-patches at sourceware dot org
- Date: Thu, 17 Oct 2013 12:42:43 -0600
- Subject: Re: [PATCH 05/13] fix up gdb.xml
- Authentication-results: sourceware.org; auth=none
- References: <1377290210-483-1-git-send-email-tromey at redhat dot com> <1377290210-483-6-git-send-email-tromey at redhat dot com> <521C97AE dot 4040707 at redhat dot com>
>>>>> "Pedro" == Pedro Alves <palves@redhat.com> writes:
Pedro> Should be:
Pedro> set msg "set tdesc filename tdesc-arch.xml ($which architecture)"
Thanks, fixed here.
>> set cmd_regex [string_to_regexp $cmd]
>> gdb_test_multiple $cmd $msg {
>> -re "^$cmd_regex\r\n$gdb_prompt $" {
>> @@ -86,31 +89,32 @@ proc set_arch { arch which } {
>> "The target architecture is set automatically \\(currently $arch\\)" \
>> "$cmd ($which architecture)"
>>
>> - file delete "${subdir}/tdesc-arch.xml"
>> - remote_file host delete "tdesc-arch.xml"
>> + remote_file host delete $filename
Pedro> I noticed this seems to leave the build machine's copy
Pedro> behind. Was that on purpose?
Yeah. It seemed simpler to just leave it around. I think it's
generally a mistake to delete intermediate files on the build machine --
it makes debugging the test suite harder, and with the parallel mode it
is generally trivial to clean up anyway.
>> # Similarly, we need to copy files under test into the objdir.
>> -proc load_description { file errmsg } {
>> +proc load_description { file errmsg {should_cd 1} } {
Pedro> Could you add a comment on the should_cd parameter please?
Pedro> It's not obvious to me why a test should cd, while the other
Pedro> shouldn't.
Done.
Tom