And thanks for the suggestion.
Something I will definitely try out.
However, I am not sure how practical this approach actually is once you depend on third party gwt libraries that do not keep a release pace matchin your internal nexus GWT 2.8 releases...
Some people seem to be keeping pace with GWT trunk, Vaadin being one such example.
I am not so sure that is a gold standard though.
In any case, I believe your approach is acceptable if you own most of the code compiled by gwt.
By the way, out of curiosity, why the main build system for GWT still ant?
This reminds of Mojarra, also still built by ant.
Kind regards.
On Monday, May 30, 2016 at 10:06:52 PM UTC+2, Ignacio Baca Moreno-Torres wrote:
I said this many times but... I strongly recommend that you create your own releases. This is what we have been doing for years...1. create a release using something like... (tools and trunk are github clones, build-deploy.sh is the script to deploy to your company repo, you should create it yourself)
system in /Users/you/Code/gwt/
› ll
total 16
-rw-r--r-- 1 you staff 247 jul 17 2015 README.md
-rwxr--r-- 1 you staff 297 ago 14 2015 build-deploy.sh
-rwxr--r-- 1 you staff 191 abr 12 2015 build-local.sh
-rwxr--r-- 1 you staff 159 mar 12 2015 push-to-m2.sh
drwxr-xr-x 10 you staff 340 abr 7 10:56 tools
drwxr-xr-x 30 you staff 1020 mar 20 09:20 trunk
drwxr-xr-x 4 you staff 136 sep 6 2014 workspace
system in /Users/you/Code/gwt/
› cat build-deploy.sh
#!/bin/sh
export GWT_VERSION=2.8.0-i`date +%Y%m%d`
export GWT_MAVEN_REPO_URL=http://maven.yourcompany.com/ext-release- local
export GWT_MAVEN_REPO_ID=maven.yourcompany .com
echo building version $GWT_VERSION...
cd trunk
ant clean buildonly elemental dist-dev
maven/push-gwt.sh
git tag -f $GWT_VERSION2. update your project gwt version... done. :)
Yep quite easy. We usually make 1 release a week, some time more some time less, GWT is extremely stable, usually if something is wrong is really wrong, which is a good thing because problems are detected quite early. You can spend more time, but I don't think that you are going to find any surprise if you develop during one or two week with a new version. So we usually change the GWT version just after a release, and we just continue developing using the new version, if no-one detects anything before the next release we just release. Looks like google compile most of it's projects using the gwt head, so you find frequently that if some commit makes something inestable they revert it in a few day (frequently the next day). So if you test your version during more than a week and you review the last commits since your release tag, you can see even more info if something is going to fail.
This has obvious advantages, you get used to compile gwt so if you find some bug which affect your project you can always hack it (which is far easier that find a perfect solution) and fix the problem for your clients, obviously, you get the cool advantage of being always using the last features of gwt, and finally your are going to start feeling like GWT is much live than you thought it be only seeing its releases.
Please, atleast try, make a release in your company maven report. Compile your project, and see what happen. Only good things can happens XD
On Monday, May 30, 2016 at 8:50:02 PM UTC+2, NGdeM wrote:One can understand that the amount of resources behind Angular are far superior to those behind GWT.In any case, the project's last official hear-beat dates from December 3, 2015, and this was a beta release.For anybody watching, it is scary to see a UI framework having releases going out at such a slow pace.It most likely far from being trivial to bring full support to the Java 8 api.But in that case, could one say that the scope for the deliverable is not properly planned?Could the scope of the GWT releases have not been sliced into multiple final incremental stable releases?None the less, even if the scope of the 2.8 release is to support the full Java 8 API, then several beta releases like the beta 1 could have already come out. Such as every 4 months there should be a goal to have something stable coming out.One could enumerate the set of of open work tasks and APIs not yet supported...At this point in time, would looking at the release notes in the GWT documentation just thinks the project is terminated.Would a beta-2 be possible any time soon?Many libraries such as Vaadin GWT widgets no longer support 2.8 beta1.Many thanks and kind regards.
On Sunday, May 8, 2016 at 5:26:22 PM UTC+2, steve Zara wrote:GWT 2.8 is now well over a year behind what seemed to be the original schedule. I'm having to deal with colleagues who say they have lost confidence in the GWT project, which is a problem as I have GWT projects to support and further develop, as part of what I hope will be a globally used and long-lived system. Confidence in software requires the presence of established releases, NOT betas, which put off investors.Is there anything that can be done to assist with progress? Is there a problem with lack of interest in GWT from, say, Google? Does GWT 2.8 involve too many features when compared to 2.7? Is there a lack of developers working on GWT? Are more testers needed?
Regards
Steve Zara
You received this message because you are subscribed to the Google Groups "GWT Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscribe@googlegroups.com.
To post to this group, send email to google-web-toolkit@googlegroups.com.
Visit this group at https://groups.google.com/group/google-web-toolkit.
For more options, visit https://groups.google.com/d/optout.
No comments:
Post a Comment