On Mon, Sep 1, 2014 at 3:55 AM, Seamus McMorrow <[email protected]> wrote:
> Hi guys, > > I have just tried out the 2.7-SNAPSHOT. I have a fairly large multi-module > maven project. > For development, I am using STS 3.6, windows7, SSD, i7 3720qm 2.6Ghz, > > Starting SuperDevmode: > Unification traversed 147422 fields and methods and 11825 types. 11773 are > considered part of the current module and 11773 had all of their fields and > methods traversed. > Compilation succeeded -- 113.359s > Linking succeeded -- 2.768s > 120.982s total -- Compile completed > The first compile in -XcompilePerFile is slower than a regular compile since it has to traverse a larger set of types and traverse them more fully (to ensure later cache accesses are correct). I'm guessing that normally you get about an 80s first compile? Eventually we should be able to make even this first compile faster by persisting the MinimalRebuildCache to disk and reusing it between different SDM process launches. Also (as Roberto pointed out recently) if we switche from whole-type linking to per-function linking then we could probably get away with doing the old-style partial-class traversal, which would mean faster first compiles and a little bit smaller output. > Changing One File in one module project/jar: > Unification traversed 5282 fields and methods and 2440 types. 238 are > considered part of the current module and 238 had all of their fields and > methods traversed. > Compilation succeeded -- 9.081s > Linking succeeded -- 1.566s > 13.436s total -- Compile completed > > Some other things to note: > > When using GWT 2.6.1, a recompile used to take about 45 secs or so. > Sourcemap debugging in Chrome now works perfectly! I can now debug the > javascript via java in chrome and all the line numbers match. (This was not > the case with 2.6.1) > On the third compile I tried changing 3 files, however it gave me errors > about not finding some class files. > Can you post these error messages and whether they are related to a Generator? Are you weren't seeing the new warning about resources being loaded via ClassLoader instead of being registered in the .gwt.xml file and loaded via ResourceOracle? > I will try to keep using this, and report any other findings. > > Keep up the great work guys, GWT development is going to get so much > cooler. :) > Looking forward to the official release > Great! Any bugs/repro cases you can report will help us get this to release-worthy, sooner. > > > -- > You received this message because you are subscribed to the Google Groups > "GWT Contributors" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/d/msgid/google-web-toolkit-contributors/f300a560-19e8-44bd-ba39-b2e48abb8522%40googlegroups.com > <https://groups.google.com/d/msgid/google-web-toolkit-contributors/f300a560-19e8-44bd-ba39-b2e48abb8522%40googlegroups.com?utm_medium=email&utm_source=footer> > . > > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "GWT Contributors" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/google-web-toolkit-contributors/CAFw3gJ_%2BMPO3eWemLf6OBqO5W8KMoPeMrKGktgm2UMiwZa7cAQ%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
