Since Lex and I are push/pull'ing you on the flag name, I'd leave it the way you have it unless a consensus otherwise emerges. But the when-to-enable flag is wrong, and I think you've moved from a too-expansive to a too-restricted spec!
http://gwt-code-reviews.appspot.com/47802/diff/2003/2009 File user/src/com/google/gwt/junit/JUnitShell.java (right): http://gwt-code-reviews.appspot.com/47802/diff/2003/2009#newcode779 Line 779: if (!argList.contains("-XdisableAssertions") && argList.contains("-web")) { I think my reaction on this one got lost in an off-line thread... I actually think "assertions are on for tests, unles s you turn them off" is simpler/better than "assertions are on for tests, but only in web mode,...", which I think makes the JUnitShell constructor change superior. There are also ways to run "web mode" that do not use -web (namely, -remoteweb, -selenium, and -manual), so this contains test is flawed anyway. My suggestion for doing it here was actually to always assert -ea (in position 0, and regardless of anything in argList), and let that be overridden by a later -XdisableAssertions flag if need be. So, regardless of intended spec, the test here is wrong. I claim that "correct" is to be mode-insensitive (i.e. JUnitShell defaults to assertions enabled, but hosted mode and compiler default to disabled). Copy GWTC for dissent, speak now or hold your peace. http://gwt-code-reviews.appspot.com/47802 --~--~---------~--~----~------------~-------~--~----~ http://groups.google.com/group/Google-Web-Toolkit-Contributors -~----------~----~----~----~------~----~------~--~---
