On 7/16/25 11:59 PM, andy pugh wrote:
Both tests 1a and 2a have the entire lifetime of mb2hal in the
'expected' output file. The '-W' flag only waits for the component to
become ready and does not wait until it is finished.
True, but the failure was a file truncation (consistently) at:
[snip]
(In fact it was spookily consistent, to the character, My initial
thought was a problem in the sed filtering in the checkresult script)
That is not uncommon in file-system operations. There is buffering at
several levels and these problems can look eerily consistent.
I have put a "sleep 1" in the checkresults script, but that doesn't
feel like a good solution.
And it isn't the right way to solve the problem because it will just
hide the race until it fails again for someone else.
From the halcmd man-page under loadusr:
* -W to wait for the component to become ready.
* -w to wait for the program to exit
You need to change the '-W' into '-w' (lower case 'w'). That should
ensure that loadusr will not finish until the program (mb2hal) has
exited. That then ensures that halrun will not exit prematurely to run
checkresults on unfinished data.
--
Greetings Bertho
(disclaimers are disclaimed)
_______________________________________________
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
