Hi again Marius, On Jan 25, 2010, at 9:25 PM, Marius Dumitru Florea wrote:
Hi Vincent,
Vincent Massol wrote:
Hi Marius,
On Jan 25, 2010, at 8:09 PM, Marius Dumitru Florea wrote:
Hi devs,
I'd like to upgrade platform/web to GWT 2.0 before the 2.2M2 release. Even if the WYSIWYG users won't notice any difference (maybe a small decrease in loading time and a small increase in responsiveness dues to compiler optimizations) this upgrade will help ease the development. You can find the detailed release notes at http://code.google.com/webtoolkit/doc/latest/ReleaseNotes.html . I'm most interested in:
* ability to run the unit tests on the build server (a 32-bit JVM isn't needed anymore, finally, since the tests are run by default with HtmlUnit) * ability to run and debug the Java code directly in the browser (new development mode using a FF plugin) * ability to split the code and load it on demand
In order to upgrade I need http://google-web-toolkit-incubator.googlecode.com/files/gwt-incubator-20091... in our maven repo:
mvn install:install-file -Dfile=gwt-incubator-20091216-r1739.jar -DgroupId=com.google.gwt -DartifactId=gwt-incubator -Dversion=20091216-r1739 -Dpackaging=jar
Who can help me with that?
All unit and integration tests pass locally so here's my +1.
My take is that it's just too dangerous to do now, especially since we should already have released XE 2.2M2 and we're late. I don't think there's any value in taking unnecessary risks at this point.
I don't understand why you think this upgrade is so dangerous.
Any upgrade done at the last moment is dangerous. And even more for a major release upgrade. If it were about upgrading from Groovy 1.7 to Groovy 2.0 I'd say the same.
Selenium tests cover a significant number of use cases and if GWT 2.0 leads to a regression we can still fix it in RC1. We can even revert the upgrade if it's really bad. I don't have bad experiences with GWT upgrades. The last one (to GWT 1.7) produced only one regression on the menu bar, which was caught by the selenium tests. GWT 2.0 is supposed to be better and fixes some bugs over GWT 1.7 which means that the widgets used by the editor are better.
I'm definitely not against upgrading. For me it's just a timing issue. Since everyone else seems to think it's a good idea I'm moving my vote from -1 to -0. Please just make sure that the final release is not delayed because of this (in case something goes wrong). Thanks -Vincent
Thanks, Marius
Personally I'd much rather wait for XE 2.3M1 (i.e a few days since we'll create a branch at the end of this week - I also need that branch to continue my work on references which I have also stopped last Thursday in order to not cause more unnecessary risks).
So I'd be -1 right now since I don't see a compelling reason to do that now. If everyone else agrees, I'll reconsider my judgement and move that to a -0.
Thanks -Vincent
PS: If you really want to start committing it now, we can easily do it on a branch and merge in a few days. The merge would probably not be hard.