On 18-04-19 18:17, Pedro Alves wrote:
> On 4/18/19 4:48 PM, Tom de Vries wrote:
> 
>> I've tested this both with native and native-gdbserver and found no
>> regressions.
>>
>> I've also verified that both blocks mentioned above are triggered by the
>> new test-cases.
>>
>> LGTM.
>>
> 
> Great, I've pushed it in, ...
> 
>>> From 575fecd185d07cd0d2f9d9aed5325e7b09b675e0 Mon Sep 17 00:00:00 2001
>>> From: Pedro Alves <palves@redhat.com>
>>> Date: Thu, 18 Apr 2019 09:57:45 +0100
>>> Subject: [PATCH] [gdb] Handle vfork in thread with follow-fork-mode child
> 
> ... after fixing the authorship back to you.  Looks like squashing in
> my fixes made stgit/git reset the author and I hadn't noticed.
> 

Is this assertion fix ok for 8.3.1?

The patch doesn't apply cleanly, but the merge conflict is trivial
(showing here the relevant diff between the patch on master and the
backport one):
...
$ diff -u master.patch bp.patch
-@@ -964,7 +968,7 @@ handle_vfork_child_exec_or_exit (int exec)
+@@ -963,7 +967,7 @@ handle_vfork_child_exec_or_exit (int exec)
          if (print_inferior_events)
            {
-             std::string pidstr
+             const char *pidstr
 -              = target_pid_to_str (ptid_t (inf->vfork_parent->pid));
 +              = target_pid_to_str (ptid_t (vfork_parent->pid));

@@ -164,7 +162,7 @@

        if (non_stop && resume_parent != -1)
...

Build on x86_64-linux, and ran test-cases belonging to the patch.

Thanks,
- Tom