Testsuite failure in 3.11

2021-01-11 Thread Jordi Mallach
Hey, I have uploaded 3.11 and 3.11.1 to the Debian build servers and most of the builds are failing with a stuck testsuite. I cannot reproduce the bug locally. Automated builds are being killed after 150 minutes of no activity (ie no terminal output). I don't see a pattern in the architectures th

Re: Testsuite failure in 3.11

2021-01-11 Thread Sergey Poznyakoff
Hi Jordi, > 54: strout > E: Build killed with signal TERM after 150 minutes of inactivity > > Any ideas, Sergey? None at all. This is the simplest tastcase one can imagine: read characters from the standard input and reproduce it on the standard output.

Re: Testsuite failure in 3.11

2021-01-11 Thread Jordi Mallach
Hi! El dl. 11 de 01 de 2021 a les 14:05 +0200, en/na Sergey Poznyakoff va escriure: > Hi Jordi, > > >  54: strout > > E: Build killed with signal TERM after 150 minutes of inactivity > > > > Any ideas, Sergey? > > None at all.  This is the simplest tastc

Re: Testsuite failure in 3.11

2021-01-11 Thread Sergey Poznyakoff
Jordi Mallach ha escrit: > It seems the test is writing junk for 30 seconds until it terminates. How interesting! I need to find a way to reproduce it. Any suggestions? > While I'm here, is RUNTESTFLAGS="-v -v -v -debug" the correct way of > enabling debug for testsuites? Nope. This is for

Re: Testsuite failure in 3.11

2021-01-11 Thread Sergey Poznyakoff
Sergey Poznyakoff ha escrit: > Jordi Mallach ha escrit: > > > It seems the test is writing junk for 30 seconds until it terminates. > > How interesting! I need to find a way to reproduce it. Any > suggestions? Forget it. Please, try commit dd8ed22c2dfcecf419e2c7a35cd344698448e2ad. Let me k

Re: Testsuite failure in 3.11

2021-01-11 Thread Sergey Poznyakoff
Additionally, the following patch fixes test failures on 32-bit architectures: 77cf5f0fd756ccef6ae133fd62fe13073f42373f Regards, Sergey

Re: Testsuite failure in 3.11

2021-01-11 Thread Sergey
On Monday 11 January 2021, Jordi Mallach wrote: > I don't see a pattern in the architectures that fail, though. I've seen this error with 3.10 when building package for ALT Linux. x86 is built only in my case (i586 and x86_64 both). -- Regards, Sergey