This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH] Run catch-load.exp on Windows
- From: Pedro Alves <palves at redhat dot com>
- To: Yao Qi <yao at codesourcery dot com>
- Cc: Pierre Muller <pierre dot muller at ics-cnrs dot unistra dot fr>, "'Eli Zaretskii'" <eliz at gnu dot org>, gdb-patches at sourceware dot org
- Date: Mon, 16 Sep 2013 15:26:24 +0100
- Subject: Re: [PATCH] Run catch-load.exp on Windows
- Authentication-results: sourceware.org; auth=none
- References: <1378947563-26245-1-git-send-email-yao at codesourcery dot com> <83li32iwrs dot fsf at gnu dot org> <5231A766 dot 8060108 at codesourcery dot com> <000301ceafcd$d3bbcda0$7b3368e0$ at muller@ics-cnrs.unistra.fr> <5237058E dot 50409 at codesourcery dot com>
On 09/16/2013 02:20 PM, Yao Qi wrote:
> On 09/12/2013 11:36 PM, Pierre Muller wrote:
>> I don't get, I always thought that
>> when using remote debugger, the debuggee is cross-compiled
>> on the machine where GDB is run, not on the remote machine where gdbserver is run.
>
> This is remote-target, not remote-host. Dejagnu is running on build,
> test case is compiled on host, and executable is running on target. In
> the case you described, host == build, so the test case is not copied.
> However, in remote-host, host != build, Dejagnu has to copy source file
> to the remote host first, and launch compiler on the remote host to
> compile the test case. The executable can be run on the same machine or
> a different one.
We do have several other tests that have .h files though.
$ find . -name "*.h" | wc -l
33
Sounds like we should come up with a way to make dejagnu handle
them in this scenario then. (tbc, not saying you should do that
yourself).
>
>> Thus, I always believed that there is no need to copy
>> the source files to the remote machine, as GDB has the sources available.
>>
>> Are these assumptions wrong?
>
> Usually, we don't, but in remote-host testing, we have to.
--
Pedro Alves