[Lldb-commits] REGRESSIONS: [lldb] r181091 - After recent OperatingsSystem plug-in changes, the lldb_private::Process and lldb_private::Thread subclasses were changed and the API was not respected properly.

[Malea, Daniel]

Hmm, I applied the patch locally and ran the tests; it seems to introduce
more regressions.

With linux.patch applied, I see 94 failing suites vs 81 in trunk right
now...

Andy, since you're looking at thread support, is there any chance you can
take a look at Greg's patch too? You might know more about stop reasons
than I do.


Thanks,
Dan


On 2013-05-06 12:49 PM, "Greg Clayton" <gclayton at apple.com> wrote:

>
>On May 6, 2013, at 8:34 AM, "Thirumurthi, Ashok"
><ashok.thirumurthi at intel.com> wrote:
>
>> 
>> 
>> -----Original Message-----
>> From: Thirumurthi, Ashok
>> Sent: Monday, May 06, 2013 11:13 AM
>> To: 'Greg Clayton'
>> Subject: REGRESSIONS: [Lldb-commits] [lldb] r181091 - After recent
>>OperatingsSystem plug-in changes, the lldb_private::Process and
>>lldb_private::Thread subclasses were changed and the API was not
>>respected properly.
>> 
>> Hi Greg,
>> 
>> This commit introduced 89 test failures on the Linux buildbots.  Here
>>are some snippets from the log that suggest that breakpoints are broken.
>> Are analogous changes required in ProcessPOSIX to remain in step with
>>the new code around thread lists?
>
>The test suite was passing on MacOSX. We will need to look at these on
>linux and figure out what is going on. For linux, there really shouldn't
>be any changes, though I would have to look at how the POSIX plug-in was
>reporting the stop reasons. I just took a look at the POSIXThread and
>noticed that it doesn't use:
>
>lldb_private::Thread::m_actual_stop_info_sp, it uses its own
>POSIXThread::m_stop_info. Can someone try and remove
>"POSIXThread::m_stop_info" and use the "SetStopInfo" accessor:
>
>void
>Thread::SetStopInfo (const lldb::StopInfoSP &stop_info_sp);
>
>Here is a patch that fixes the issue:
>
>