On Fri Aug 15 07:00:38 2008, coke wrote: > #not ok 1 - Line length ok > # Failed test 'Line length ok' > # at t/codingstd/linelength.t line 80. > # Lines longer than coding standard limit (100 columns) in 1 files: > # /home/smoke/parrot/compilers/pirc/new/pirsymbol.c:256: 104 cols > # Looks like you failed 1 test of 1. > > This causes -all- smolder reports to be marked as failures. >
Coke et. al., May I offer a dissenting opinion? It ain't broke, so we shouldn't fix it. Just now, I took a look at http://smolder.plusthree.com/app/public_projects/smoke_reports/8 and requested the 30 most recent reports. I saw that some reports from earlier today were passing 99.99% of the tests and that the failing test, as you reported, was this coding standards test. Later in the day, the test was fixed, so most reports (on well functioning OSes) resumed 100% passing. Is that a bad thing? Isn't that exactly what we want out of our Smolder tests (and our other smoke tests as well)? So what if a coding standard test is less important than a core test; do we want to find out about it quickly or not? And I, for one, find myself going to the Smolder site much more often than our 'official' site these days -- precisely because I can spot new test failures more quickly there and jump in with a quick fix. You don't have to plow thru year-old tests on OSes that no one is actively developing on, as you do at smoke.parrotcode.org. Now, we might want to permit *individual* smolder testers have the option of submitting, say, only 'make coretest'. We can do that now at smoke.parrotcode.org. But I would like to see the *default* setting for Smolder remain 'make test'. Thank you very much. kid51
