|
From: Nicholas N. <nj...@cs...> - 2005-07-21 00:16:33
|
Hi, I've rewritten the nightly test script to test both the current Valgrind and the one from 24 hours ago. It's attached. If there is no difference between the 24-hour-old results and the current results, it just shows the current results, same as it does now. If there is a difference, it shows the diff and both sets of output. An example of this is attached. If there are no objections I'll check it in, and Tom and Julian can roll it out onto their test machines. Thanks. N |
|
From: Julian S. <js...@ac...> - 2005-07-21 08:57:07
|
Great stuff! I like it. My only reservation is that the output is a bit verbose. At a minimum, could we have the current results first, and then the 24-h results, and then the diff? So I don't have to scroll down in kmail to see the most recent results. Ideally I was hoping for the following, but perhaps it's a bit much as I realise this would involve extracting the relevant parts of the short logfiles, which isn't obvious how to do. (note: the implication is I don't really care about the 24-h ago build results, which is true). J Nightly build on nemesis ( SuSE 9.0 ) started at 2005-07-20 16:08:00 CDT ### these are the current build results Checking out vex source tree ... done Building vex ... done Checking out valgrind source tree ... done Configuring valgrind ... done Building valgrind ... done Running regression tests ... failed ------------ Regression test results ------------ == 180 tests, 3 stderr failures, 0 stdout failures ================= memcheck/tests/stack_changes (stderr) none/tests/faultstatus (stderr) none/tests/x86/int (stderr) ------------ Yesterday's regression test results ------------ == 180 tests, 4 stderr failures, 2 stdout failures ================= memcheck/tests/stack_changes (stderr) memcheck/tests/xml1 (stderr) none/tests/cmdline1 (stdout) none/tests/cmdline2 (stdout) none/tests/faultstatus (stderr) none/tests/x86/int (stderr) ------------ Difference ------------ ! == 180 tests, 4 stderr failures, 2 stdout failures ================= memcheck/tests/stack_changes (stderr) - memcheck/tests/xml1 (stderr) - none/tests/cmdline1 (stdout) - none/tests/cmdline2 (stdout) none/tests/faultstatus (stderr) --- 10,13 ---- ! == 180 tests, 3 stderr failures, 0 stdout failures ================= memcheck/tests/stack_changes (stderr) none/tests/faultstatus (stderr) |
|
From: Nicholas N. <nj...@cs...> - 2005-07-21 13:00:10
|
On Thu, 21 Jul 2005, Julian Seward wrote: > My only reservation is that the output is a bit verbose. At > a minimum, could we have the current results first, and then the > 24-h results, and then the diff? So I don't have to scroll down > in kmail to see the most recent results. Ok, I'll change it. > Ideally I was hoping for the following, but perhaps it's a bit > much as I realise this would involve extracting the relevant > parts of the short logfiles, which isn't obvious how to do. > (note: the implication is I don't really care about the 24-h > ago build results, which is true). It also assumes that we got as far as the regtests. Sometimes building (or even checking out) fails. Remember that most of the time the results will be the same as the previous day's, so the more verbose output will be the exception. N |