Re: [BUG?] applypatch-msg hook no-longer thinks stdin is a tty

2015-10-03 Thread Johannes Sixt
Am 03.10.2015 um 09:37 schrieb Chris Packham: On Sat, Oct 3, 2015 at 6:43 AM, Junio C Hamano wrote: If you want to go interactive from the hook, you'd have to open and interact with /dev/tty yourself in your hook anyway. That may be what I have to do, although I have

Re: [BUG?] applypatch-msg hook no-longer thinks stdin is a tty

2015-10-03 Thread Chris Packham
On Sat, Oct 3, 2015 at 6:43 AM, Junio C Hamano wrote: > Junio C Hamano writes: > >> Chris Packham writes: >> >>> As of git 2.6 this has stopped working and stdin always fails the tty >>> check. >> >> We now run that hook thru

Re: [BUG?] applypatch-msg hook no-longer thinks stdin is a tty

2015-10-02 Thread Junio C Hamano
Chris Packham writes: > As of git 2.6 this has stopped working and stdin always fails the tty > check. We now run that hook thru run_hook_ve(), which closes the standard input (as the hook is not reading anything). Perhaps you can check if your output is connected to

Re: [BUG?] applypatch-msg hook no-longer thinks stdin is a tty

2015-10-02 Thread Junio C Hamano
Junio C Hamano writes: > Chris Packham writes: > >> As of git 2.6 this has stopped working and stdin always fails the tty >> check. > > We now run that hook thru run_hook_ve(), which closes the standard > input (as the hook is not reading anything).

[BUG?] applypatch-msg hook no-longer thinks stdin is a tty

2015-09-30 Thread Chris Packham
Hi, I have a applypatch-msg hook that implements some policy for acceptable commit messages and reject non-conformant patches. It also is able to prompt me to override it's rejection. The prompting only happens when stdin is a tty (as determined by pythons sys.stdin.isatty()) For example this