Race condition hangs on multiple mintty/tcsh? Brad Wetmore

[Thomas Wolff]

Am 11.08.2020 um 19:37 schrieb Thomas Wolff:
>
> On 11.08.2020 12:30, Corinna Vinschen wrote:
>> On Aug 11 19:02, Takashi Yano via Cygwin wrote:
>>> On Tue, 11 Aug 2020 11:37:58 +0200
>>> Corinna Vinschen wrote:
>>>> On Aug 11 13:59, Takashi Yano via Cygwin wrote:
>>>>> Hi Thomas,
>>>>>
>>>>> On Thu, 6 Aug 2020 15:31:24 +0200
>>>>> Thomas Wolff wrote:
>>>>>> Am 06.08.2020 um 13:46 schrieb Thomas Wolff:
>>>>>>> Am 06.08.2020 um 01:23 schrieb Kevin Schnitzius via Cygwin:
>>>>>>>> On Wednesday, August 5, 2020, 06:56:48 PM EDT, Thomas Wolff
>>>>>>>> <towo@towo.net> wrote:
>>>>>>>>> Am 04.08.2020 um 12:02 schrieb Thomas Wolff:
>>>>>>>>>> Am 04.08.2020 um 00:13 schrieb Brad Wetmore via Cygwin:
>>>>>>>>>>> Hi,
>>>>>>>>>>>
>>>>>>>>>>> I generally kick off multiple (10) mintty sessions, and 
>>>>>>>>>>> place them
>>>>>>>>>>> around the screen.
>>>>>>>>>>> [...]
>>>>>>> I could reproduce one case of one of three terminals being
>>>>>>> unresponsive also with xterm.
>>>>>> And it also happens if I drop select() from mintty (and use just
>>>>>> non-blocking read()).
>>>>>>> Does anybody familiar with pty/select or recent changes have any 
>>>>>>> idea?
>>>>> I looked into this problem. After much struggle, I think
>>>>> I have found a workaround for this issue.
>>>>>
>>>>> I am not sure why this solves the issue at all, however,
>>>>> this works for me.
>>>>>
>>>>> Could you please test a patch attached?
>>>> Shall I commit the patch and create a developer snapshot for testing?
>>> Thanks so much. That would be more convenient for potential testers.
>> Done.  Please try the latest snapshot from https://cygwin.com/snapshots/
> Thanks. I could not reproduce the issue on one system in ~27 cases. 
> I'd like to retest on another system where I could previously produce 
> the issue more frequently, but that's currently broken...
I gave it another 22 test instances on my other machine (the more 
sensitive one), no issue reproduced with the patch.