On Sep 7, 2011, at 4:58 PM, Ojan Vafai wrote: > On Wed, Sep 7, 2011 at 4:48 PM, Alexey Proskuryakov <[email protected]> wrote: > > 16.08.2011, в 17:45, Alexey Proskuryakov написал(а): > > >> I think that a style bot rule complaining about new files in script-tests > >> directories (outside fast/js) would be great. If there are no new script > >> tests, then there are no incorrectly prepared ones. > > > > I'm proposing a change for make-script-test-wrappers to only look inside > > fast/js in <https://bugs.webkit.org/show_bug.cgi?id=66357>. In next steps, > > I'd like to remove TEMPLATE.html files, and then merge most or all existing > > split tests into their .html files. > > Maciej has expressed concern that making people create monolithic script > tests manually is undesirable. You need to find an existing script test, copy > boilerplate, and fix the number of ".." path components, which involves > several annoying steps. > > I'm proposing a make-new-script-test script for that in > <https://bugs.webkit.org/show_bug.cgi?id=67747>. > > IMO, it's only undesirable because of the current amount of boilerplate. We > could change script test JS code so that the only boilerplate you need is: > <!DOCTYPE HTML> > <script src="path/to/test-harness.js"></script> > > For example, there's no need for a JS file to run after the body of the test > because the initial script can just listen to the window load event. This is > how dump-as-markup.js works and I think it's preferable to needing a script > to generate the test and to having the test be split up between an HTML and a > JS file.
Do you have an idea for how to avoid needing to adjust the number of "../" path segments in the path to the harness? That's the most annoying step of the current copy/paste solution, and it does not seem like simplifying the boilerplate as you suggest resolves that issue. Regards, Maciej
_______________________________________________ webkit-dev mailing list [email protected] http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
