Re: [RFC] (windows) GDB/MI crash when using "-list-thread-groups --available"

[Pedro Alves <palves@redhat.com>]

On 05/10/2018 10:02 PM, Simon Marchi wrote:

> 
> So when an inferior is started, "-list-thread-groups --available" works correctly on Windows?  What is the target beneath then, which provides the list of available processes on Windows?

After the inferior is started, the Windows target is pushed in the target
stack, so there will be a target beneath, either the exec target, or the
dummy target directly.  Either of those returns TARGET_XFER_E_IO for
this target object.  

The issue here is that before the inferior is started, the
Windows target is not pushed on the target stack.
See target_get_osdata.

> which provides the list of available processes on Windows?

I don't think the feature works at all on Windows.  It's probably
returning an empty list of processes.

> 
> Looking at how it works on Linux, it's the process stratum, inf_ptrace_target, that answers this request.  On Windows, shouldn't windows_nat_target answer this request?  After all, it's the responsibility of windows_nat_target to communicate with the Windows OS to debug processes natively on it.

Yeah, it's just that the Windows port doesn't really implement the
feature at all, AFAICT.  Ideally it'd be implemented.  Otherwise,
I guess handling the case of not having a target beneath
here is reasonable.

> 
>> Also, The testcase I am proposing fails on the -list-thread-groups test
>> when run on GNU/Linux because, on that platform, the command returns
>> more output than the expect buffer can handle, resulting in an UNRESOLVED
>> status. How does one usually handle this? The only why I can think of
>> is a loop of gdb_test_multiple... Other ideas?
> 

> What's the difference between the new test case and gdb.mi/list-thread-groups-available.exp?  In that one too, -list-thread-groups --available is executed with no inferior started.  It also uses mi_gdb_test though, so it probably hits the same limitation.

I've actually saw that testcase fail before because of this issue.  :-)
I probably saw it when I was running many test jobs in parallel (thus many
processes) or something like that.

> 
> As a quick and dirty hack, is it possible to just increase temporarily the size of the buffer to something that will surely be large enough?  Otherwise, using gdb_test_multiple or maybe gdb_expect to consume the output little by little sounds good.

Yeah, the best way to address this is to consume
output in chunks, with exp_continue.  That fixes it for good.
See for example:

commit 11859c310cd6b6fd892337a5ee1d36921e6d08d8
Author:     Andrew Burgess <andrew.burgess@embecosm.com>
AuthorDate: Mon Apr 9 00:18:34 2018 +0100

    gdb/testsuite: Handle targets with lots of registers


I'd prefer that over increasing buffer sizes.

Thanks,
Pedro Alves