incubator-wave-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Thomas Wrobel <darkfl...@gmail.com>
Subject Re: Setting up Wave for client development (my experience setting up on WindowsXP)
Date Wed, 04 Dec 2013 20:50:56 GMT
Might be worth at some point looking into alternatives for multi-language
support?
Should be possible without multiplying compile times.
I'll have to look into it once I am up and running.

~~~
Thomas & Bertines online review show:
http://randomreviewshow.com/index.html
Try it! You might even feel ambivalent about it :)


On 4 December 2013 21:33, Yuri Z <vega113@gmail.com> wrote:

> Permutation is not only per browser, it is also per language and I think
> WIAB supports 4 languages.
> Anyway, it is very strange it took 4 hours, probably 2 GB is too little,
> you ll need about 4 GB.
>
>
> On Wed, Dec 4, 2013 at 10:25 PM, Thomas Wrobel <darkflame@gmail.com>
> wrote:
>
> > 32??? There isnt 32 different browser engines is there :?
> > My own GWT projects, (using 2.5.1) use 7 at most. (and even that should
> go
> > down in newer versions as Opera phase's out Presto.)
> >
> > It wasn't GWT permutations taking the bulk of the time  anyway, mind you
> -
> > it seemed to mostly be the testing and (strangely) expanding JAR files.
> > That was just my perception though. Certainly before it got to
> > "compile-gwt:" took at least 4 hours.
> > Would the log help here?
> >
> > My machine is a 4200 dual core Amd. Not much ram (2GB), was running
> chrome
> > at the same time, but not doing anything intensive. I wouldn't be
> surprised
> > if my machine is to blame here, but I cant think why it would be this
> > different.
> > ---
> > Anyway, dinner now, then back to poking at things.
> >
> >
> > ~~~
> > Thomas & Bertines online review show:
> > http://randomreviewshow.com/index.html
> > Try it! You might even feel ambivalent about it :)
> >
> >
> > On 4 December 2013 20:59, Yuri Z <vega113@gmail.com> wrote:
> >
> > > We have 32 GWT permutations the moment, we used to have only 4... Some
> > last
> > > changes caused this increase... We need to be more cautious about
> > updating
> > > GWT client code.
> > > I tried
> > > ant clean dist-server compile-gwt test
> > > It took me about 16 minutes. If you tried the default target which also
> > > includes running tests then it could take about 6 minutes more.
> > > So max 21 minutes on 2-core laptop. This is for the full prod build, if
> > you
> > > run the server from compiled source with dev GWT setting(only 2
> > > permutations) then it takes only a few minutes, or even less.
> > >
> > > Basically running wave is simple like:
> > >
> > > git clone git://git.apache.org/wave.git wave
> > > cd wave
> > > cp server.config.example server.config
> > > ant dist-server compile-gwt run-server
> > > Open the browser at http://localhost:9898
> > >
> > >
> > > On Wed, Dec 4, 2013 at 9:40 PM, Ali Lown <ali@lown.me.uk> wrote:
> > >
> > > > > "BUILD SUCCESSFUL
> > > > > Total time: 312 minutes 41 seconds"
> > > >
> > > > Err.. it takes ~5 minutes on my dev machine! Is this a single core
> vm,
> > > > doing lots of swapping, and with shared io?
> > > >
> > > > > Suggestion;
> > > > > Would it be possible to have a virtual machine with everything set
> up
> > > > > already? or is there technical/license reasons for that to be
> > > unsuitable?
> > > >
> > > > I suspect this would be difficult. (And you don't really want to be
> in
> > a
> > > > VM).
> > > >
> > > > > Query:
> > > > > Can Wave be updated to JDK7? is there big issues holding it back
?
> Or
> > > is
> > > > > there more open alternatives we can use instead - one that doesn't
> > > > require
> > > > > handing over personal details to a company?
> > > >
> > > > (OpenJDK 1.6 works fine, so...)
> > > >
> > > > This is quite difficult to do for the codebase. You would also need
> to
> > > > upgrade all the third-party components.
> > > >
> > > > Please continue to provide feedback.
> > > >
> > > > Thanks.
> > > > Ali
> > > >
> > >
> >
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message