Re: What's the future of the GWT Designer?

[Thomas Broyer]

[+gwt-steering]

Thanks Eric. That was my impression but now we have an authoritative answer.

Do you think you'd have the time to at least make GWT Designer compatible with GWT 2.6 and GWT.next? How if someone provides a patch? (probably not me, and probably untested if the open-source repo doesn't build)

In the mean time, we have stopped "advertizing" GWT Designer on gwtproject.org: https://gwt-review.googlesource.com/6590, and we have a fix for GWT 2.6.1: https://gwt-review.googlesource.com/6342 (which we'll release soon; we'll discuss this Wednesday during our Steering Committee meeting).

I'll do a call for maintainers in the GWT community on Google+ later today (at a time where I'll hopefully reach both EMEA and the Americas).

On Sun, Mar 9, 2014 at 9:02 PM, Eric Clayberg <clay...@google.com> wrote:

The folks who built GWT Designer are no longer involved in GWT work, but would be happy to help hand it over to someone who wants to maintain it. I'm sure we could help getting it to build, for example.

On Fri, Mar 7, 2014 at 5:52 AM, Thomas Broyer <t.br...@gmail.com> wrote:

Hi all,

What's the future of the GWT Designer?

We're currently discussing whether to update or remove a link to the GWT Designer in the gwtproject.org doc: https://gwt-review.googlesource.com/6590

It seems to me like the GWT Designer is not actively maintained anymore. We broke the compatibility with GWT 2.6.0 and, while we have a fix in 2.6.1 (not ETA yet), GWT Designer will have to be updated for the GWT.next version.

I don't think any GWT maintainer uses the GWT Designer so we have no real incentive making sure it doesn't break (it uses some "semi private / internal" APIs from GWT, and we have no real knowledge of how it works and thus what could break it – which is exactly the reason it's now broken with GWT 2.6.0). Apparently, nobody within Google uses it either, as they'd have noticed the breakage way earlier than the 2.6.0 release; so I'd understand that the GPE team only does the strict minimum updating it for each new Eclipse release and possibly fixing a few bugs here and there.

It also looks like the GPE team is more concerned about App Engine than GWT (and that's fine, provided they at least maintain the basic GWT integration that many people use).

So let me ask again: what's the future of the GWT Designer? (both standalone, and as part of the GPE)

--
Thomas Broyer
/tɔ.ma.bʁwa.je/

[Eric Clayberg]

Making GWTD compatible with GWT 2.6 and beyond is certainly possible. I suppose that depends on what exactly is wrong with it. If it is just a matter of patching the problem described in Issue 8558, and rebuilding it as-is, we should be able to make that happen (although it sounds like you are restoring the old API that GWTD needed, so is that really needed?).

As far as on-going maintenance is concerned, we could certainly add new committers to the project and turn them loose. The existing committers from Google are all working on other projects and have very limited time for anything GWTD-related. There are a couple of non-Google committers (also working on other projects) who might be able to do some work on it, but I can't speak for them. If anyone wants to be added as a committer to GWTD, send me a note.

The biggest challenges with GWTD are actually unrelated to GWT and any changes there. Keeping GWTD updated and working with new GWT versions was always fairly straightforward. The hardest problem was dealing with OS changes (almost always on the Linux side) that broke it. Every time a window manager or distro changed, it would seem to break GWTD. Incompatibilities between Linux distros, versions, and window managers was a very special kind of Hell I never want to revisit. Working on GWTD made me (and the other committers) really hate Linux for a long time (and I still have a bad taste in my mouth from it).

The other major issue is the build environment. The current build is unfortunately tied into internal Google infrastructure. GWTD really needs a completely new build based on the latest standard Eclipse build tools. Creating that will probably take someone a several days to a week of work. I don't see any help coming from our end on doing that. Even sanitizing the current build system enough to dump it out there unsupported (and definitely non working) would take some time from folks I can't speak for.

So... if someone wants to take it over, we can help make that happen (and try to answer questions as they come up).

[Thomas Broyer]

On Wed, Mar 12, 2014 at 5:55 PM, Eric Clayberg <clay...@google.com> wrote:

Making GWTD compatible with GWT 2.6 and beyond is certainly possible. I suppose that depends on what exactly is wrong with it. If it is just a matter of patching the problem described in Issue 8558, and rebuilding it as-is, we should be able to make that happen (although it sounds like you are restoring the old API that GWTD needed, so is that really needed?).

We've only restored it in the release/2.6 branch, not in master (so not in GWT.next).

 

As far as on-going maintenance is concerned, we could certainly add new committers to the project and turn them loose. The existing committers from Google are all working on other projects and have very limited time for anything GWTD-related. There are a couple of non-Google committers (also working on other projects) who might be able to do some work on it, but I can't speak for them. If anyone wants to be added as a committer to GWTD, send me a note.

The biggest challenges with GWTD are actually unrelated to GWT and any changes there. Keeping GWTD updated and working with new GWT versions was always fairly straightforward. The hardest problem was dealing with OS changes (almost always on the Linux side) that broke it. Every time a window manager or distro changed, it would seem to break GWTD. Incompatibilities between Linux distros, versions, and window managers was a very special kind of Hell I never want to revisit. Working on GWTD made me (and the other committers) really hate Linux for a long time (and I still have a bad taste in my mouth from it).

The other major issue is the build environment. The current build is unfortunately tied into internal Google infrastructure. GWTD really needs a completely new build based on the latest standard Eclipse build tools. Creating that will probably take someone a several days to a week of work. I don't see any help coming from our end on doing that. Even sanitizing the current build system enough to dump it out there unsupported (and definitely non working) would take some time from folks I can't speak for.

So... if someone wants to take it over, we can help make that happen (and try to answer questions as they come up).

FYI, I made a "call for volunteers" on Google+ https://plus.google.com/113945685385052458154/posts/TfEAqNwkZGV that links to this thread.

Also, we've just had a GWT Steering Committee meeting and we were thinking maybe we (you actually) should stop bundling it with the Google Plugin for Eclipse and only have it available as a standalone plugin (that would eventually no longer be endorsed by Google); but maybe wait a bit for new volunteers to maintain the project before doing that.

Also, please wait for the meeting minutes, I might have misunderstood something as English is not my native language ;-)