RESOLVED FIXED 116858
Bad value in tests counter at new-run-webkit-tests in --debug-rwt-logging mode
https://bugs.webkit.org/show_bug.cgi?id=116858
Summary Bad value in tests counter at new-run-webkit-tests in --debug-rwt-logging mode
Dariusz Frankiewicz
Reported 2013-05-28 03:06:17 PDT
Created attachment 203033 [details] log When running new-run-webkit-tests with --debug-rwt-logging parameter, logger shows wrong number of actual running test. For example when we've got 9 tests it starts from 0 and ends at 8. So in last line we could think that script ran one less test, until we've look at the begining. [0/9] editing/text-iterator/backward-textiterator-first-letter-crash.html "ps -eo comm,command" took 0.02s worker/0 editing/text-iterator/backward-textiterator-first-letter-crash.html passed [1/9] editing/text-iterator/find-after-mutation.html worker/0 editing/text-iterator/find-after-mutation.html passed [2/9] editing/text-iterator/findString.html worker/0 editing/text-iterator/findString.html passed [3/9] editing/text-iterator/first-letter-rtl-crash.html worker/0 editing/text-iterator/first-letter-rtl-crash.html passed [4/9] editing/text-iterator/first-letter-word-boundary.html worker/0 editing/text-iterator/first-letter-word-boundary.html passed [5/9] editing/text-iterator/range-to-from-location-and-length.html worker/0 editing/text-iterator/range-to-from-location-and-length.html passed [6/9] editing/text-iterator/rtl-first-letter-text-iterator-crash.html worker/0 editing/text-iterator/rtl-first-letter-text-iterator-crash.html passed [7/9] editing/text-iterator/rtl-selection-crash.html worker/0 editing/text-iterator/rtl-selection-crash.html passed [8/9] editing/text-iterator/thai-cursor-movement.html In my opinion it should indicate value one higher so at the end we've got [9/9] To reproduce this you can run: ./new-run-webkit-tests --efl -2 -v --debug --debug-rwt-logging LayoutTests/editing/text-iterator
Attachments
log (6.06 KB, text/plain)
2013-05-28 03:06 PDT, Dariusz Frankiewicz
no flags
Patch (1.47 KB, patch)
2013-05-29 03:14 PDT, Dariusz Frankiewicz
dpranke: review-
Proposed patch with improvements (7.27 KB, patch)
2013-06-04 03:27 PDT, Dariusz Frankiewicz
no flags
Dirk Pranke
Comment 1 2013-05-28 09:40:34 PDT
The status line is actually printing the number of tests that have completed, not the number of tests that have been started, which is why it starts at 0. The status line is also designed to be overwritten in place, so when you actually complete the last (9th) test, we immediately erase it and move on, so you never actually see [9/9]. This was by design, but arguably is not the best design, and I'm open to changing it.
Dariusz Frankiewicz
Comment 2 2013-05-29 03:14:52 PDT
Dirk Pranke
Comment 3 2013-05-29 12:19:47 PDT
Comment on attachment 203146 [details] Patch if you incremented 'num_completed' before the test has completed, that's wrong, isn't it? :) At the very least, you should change the name to 'num_started', and make sure you update the other places it is used.
Dariusz Frankiewicz
Comment 4 2013-06-04 03:27:14 PDT
Created attachment 203673 [details] Proposed patch with improvements
WebKit Commit Bot
Comment 5 2013-06-05 01:08:34 PDT
Comment on attachment 203673 [details] Proposed patch with improvements Clearing flags on attachment: 203673 Committed r151201: <http://trac.webkit.org/changeset/151201>
WebKit Commit Bot
Comment 6 2013-06-05 01:08:36 PDT
All reviewed patches have been landed. Closing bug.
Note You need to log in before you can comment on or make changes to this bug.