On 18/05/2025 07:11, Takashi Yano wrote:
On Sat, 17 May 2025 15:00:53 +0100
Jon Turney wrote:
This fixes constantly replaying the exception if we have a segfault
while a debugger is already attached, e.g. stracing a segv, see:
https://cygwin.com/pipermail/cygwin/2025-May/258144.html
(I'm tempted to remove the 'debugging' static from exception::handle()
and everything associated with it, since replaying the exception the
next half a million times it's hit seems really weird)
I adjusted the wording of this.
(This would seem to make try_to_debug() less useful, as it then does
nothing if you're just run under gdb, but it's what the code used to
do...)
This is me struggling to articulate that the first chunk is overkill.
When call stack looks like:
api_fatal
api_fatal_debug
try_to_debug
We probably want to break_here() when we were started under a debugger
(e.g. gdb)
(This is something I added in an attempt to be helpful in aa822482).
I don't understand what the sentences in ()s mean.
Clarified, hopefully.
Fixes: 91457377d6c9 ("Cygwin: Make 'ulimit -c' control writing a coredump")
Please add "Signed-off-by:". I also think it is better to add
"Reported-by:".
Done.
[...]
Otherwise, LGTM. Please push.
Thanks.
I dropped the first chunk for the reasons above.
I also updated 3.6.2 release note.
