11 Commits

Author	SHA1	Message	Date
Foo	2d877c9240	Move shell tests inputs and outputs to separate directories	2017-05-01 01:58:48 +03:00
Foo	5a3e19847e	Refactor shells test code ... Tests passing locally: bash (no)daemon C busybox (no)daemon C fish (no)daemon C mksh (no)daemon C rc (no)daemon C tcsh (no)daemon C zsh (no)daemon C zsh zpython ipython pdf module pdf subclass dash has problems with job control (bgscript.sh not finished by kill)	2017-05-01 00:28:25 +03:00
ZyX	74e81bf788	Restore the timeout ... There may be a reason for a test to take 3 minutes long, but there is no reason for the test to take even half a minute long *without any output*.	2015-02-21 19:55:39 +03:00
ZyX	678faed4c2	Force destroying child after everything was done ... Normally this line should not do anything useful.	2015-02-21 19:55:39 +03:00
ZyX	b7c7d29aa0	Use “non-blocking” read and stop on TIMEOUT and EOF	2015-02-21 19:55:39 +03:00
ZyX	6a6991bcb4	Use 3 minutes for timeout	2015-02-21 19:55:38 +03:00
ZyX	7d98218b64	Do not wait, just .read() ... .read() without arguments should read until EOF is received.	2015-02-21 19:55:38 +03:00
ZyX	0a87b558d9	Do not use child.wait(), it hangs	2015-02-21 19:55:38 +03:00
ZyX	0710f2760f	Open log as binary file for writing ... Child output is binary.	2015-02-21 16:11:04 +03:00
ZyX	22162a9656	Replace screen-based tests with pexpect-based tests ... Currently works for all tests, but dash ones. Not much of a problem since dash tests were disabled for being unstable.	2015-02-21 15:59:10 +03:00
ZyX	8524ee35e7	Ignore stderr from printf and env ... socat may close pipe before they succeed to write something which results in env: write error: Broken pipe (according to my experience leading printf’s always succeeds to write before socat closes pipe).	2015-02-21 13:56:58 +03:00