Increase cppcheck test timeout.

[nuclearsandwich]

test_rclcpp is still seeing cppcheck timeouts on Windows.
This adds another 30 seconds to the timeout in order to provide more headroom.

Example builds that failed this week:

• https://ci.ros2.org/view/nightly/job/nightly_win_deb/1712/testReport/
• https://ci.ros2.org/view/nightly/job/nightly_win_rel/1660/testReport/

Builds running the cppcheck tests with retest-until-fail to increase the chance of getting a timeout:

• Windows debug
• Windows release

[dirk-thomas]

For how long do these tests run when they don't reach the timeout? And what is the variability in their runtime? E.g. the first successful build would have passed without this change since it was 2 and a half minutes.

[chapulina]

For how long do these tests run when they don't reach the timeout? And what is the variability in their runtime?

It looks like it's usually less than a couple seconds for both win_rel and win_deb:

• https://ci.ros2.org/view/nightly/job/nightly_win_deb/1712/testReport/(root)/projectroot/cppcheck/history/
• https://ci.ros2.org/view/nightly/job/nightly_win_rel/1660/testReport/(root)/projectroot/cppcheck/history/

This suggests that increasing the timeout may not solve the issue. The test is probably hanging indefinitely.

[nuclearsandwich]

It looks like it's usually less than a couple seconds for both win_rel and win_deb:

I don't know what test those reported numbers are for but I don't think it's test_rclcpp's cppcheck run. Looking at the log lines from my two sets of builds (I had retest until fail set to see multiple runs with or without timeouts) I see:

00:53:11.337     Test #77: cppcheck .........................   Passed  195.14 sec
00:56:27.559     Test #77: cppcheck .........................   Passed  196.22 sec
00:59:42.725     Test #77: cppcheck .........................   Passed  195.16 sec
01:02:58.566     Test #77: cppcheck .........................   Passed  195.84 sec
01:06:16.630     Test #77: cppcheck .........................   Passed  198.06 sec
01:09:29.517     Test #77: cppcheck .........................   Passed  192.88 sec
01:12:44.611     Test #77: cppcheck .........................   Passed  195.09 sec
01:16:00.152     Test #77: cppcheck .........................   Passed  195.54 sec
01:19:17.493     Test #77: cppcheck .........................   Passed  197.34 sec
01:22:36.435     Test #77: cppcheck .........................   Passed  198.94 sec
01:25:53.278 1/1 Test #77: cppcheck .........................   Passed  196.84 sec

00:44:18.416     Test #77: cppcheck .........................   Passed  147.34 sec
00:46:43.931     Test #77: cppcheck .........................   Passed  145.51 sec
00:49:09.923     Test #77: cppcheck .........................   Passed  145.99 sec
00:51:36.023     Test #77: cppcheck .........................   Passed  146.10 sec
00:54:02.496     Test #77: cppcheck .........................   Passed  146.47 sec
00:56:28.408 1/1 Test #77: cppcheck .........................   Passed  145.91 sec

Last night's Debug gave

76/80 Test #77: cppcheck ........................................................   Passed  156.32 sec

and release timed out after 180s.

For how long do these tests run when they don't reach the timeout? And what is the variability in their runtime? E.g. the first successful build would have passed without this change since it was 2 and a half minutes.

So based on the above it seems highly variable. I'm not sure whether cppcheck in these cases is IO or CPU bound but either metric is prone to temporary exhaustion based on the instance CPU burst token availability and IOPS availability.