Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    Stephen Haberman
    @stephenh
    and moving is going to kind of suck.
    what makes it harder is that there is not a blessed/obvious framework to move to.
    that is kind of good, in terms of there is competition about which one is best, and best fits your setup, but it means you're back in the traditional Java approach of "piece together your own custom stack" instead of just using something with everything already wired together.
    for cell tables, the question is nuanced. :-)
    a) yes, tessell "supports" them, b) it actually has some docs http://www.tessell.org/celltable.html, but c) i don't currently use any cell widgets in our app, so YMMV
    i think the typical/non-tessell approaches to cell widgets kind of suck.
    and that tessell could go a long way to making them easier to use/easier to test.
    e.g. tessell's approach is typically "okay, we'll stub out the browser, and just pretend things still work"
    but the implementation of cell widgets, even with things that you would think don't need direct access to the DOM (like sorting/paging/etc.) had random DOM-specific code in them, like scheduled deferreds/etc.
    so, the hairy implementation + my general preference to just not use them, means their support status is "...basically..."
    Anton Johansson
    @anton-johansson
    Allright, I'll investigate some more about dispatching commands instead. The only problem I see is that the commands would require dependencies from the OSGi container in my case, but there's maybe a good solution for that. :)
    Okay, yeah that makes sense. But what would you use instead of a CellTable?
    I guess you could use FlexTable, which I haven't looked at very much. But doesn't have the same issues? Also it being harder to automatically bind to a model?
    What I would like is really some widget that simply represents a regular <table> that is completely dumb. Doesn't even know about the column layout. Instead you'd set a TableModel to it, which has all the information you need, which columns it has, how to render them etc, and also the possibility to bind the data to a model.
    Anton Johansson
    @anton-johansson
    Hey Stephen!
    How's it going?
    Stephen Haberman
    @stephenh
    hi anton; i've been doing fine.
    funny it was mid-july that we were chatting; time flies.
    you'd been asking about cell tables last time; kind of looks like i didn't mention it, but http://tessell.org/rowtable.html is how all of the tables in our app work, and while it doesn't have celltable's flyweight/scalable architecture, it's more dumb & simple, and works really well with tessell's generated stubs (as both the header and row view get stubs generated).
    Anton Johansson
    @anton-johansson
    yeah, time flies...
    Unfortunately, our GWT-project got put on hold due to more prioritized things. :(
    Allright! Dumb & simple sounds better to me, when working with GUI's. :)
    Stephen Haberman
    @stephenh
    ah sure, that is how things go. :-)
    Anton Johansson
    @anton-johansson
    Maybe we should move tessell-maven-plugin into your original repository, and have it build it together with the rest of tessell?
    Anton Johansson
    @anton-johansson
    I added an issue for adding a flag to disable AppViews-generation, like we discussed a while back.
    I'll see if I can get a PR up for it, when I have the time. :)
    Anton Johansson
    @anton-johansson
    By the way, are you only developing for Java 8 and GWT 2.8?
    Stephen Haberman
    @stephenh
    hi anton; yeah, we're only developing on java 8 and gwt trunk these days. lambdas were too tempting. :-)
    Anton Johansson
    @anton-johansson
    Haha I completely understand.
    We haven't moved over to Java 8 yet (for some reason) and we're going mad not being able to utilize lambdas.
    Anton Johansson
    @anton-johansson
    What do you think about merging the maven plugin with the original project, to keep them versioned together?
    Stephen Haberman
    @stephenh
    hi anton. i think for now i'm fine with them being separate projects.
    especially since i don't use the maven plugin personally, i would not be surprised that even if they lived in the same project, there is not a strong guarantee i'd always keep it fixed.
    unless it was part of the travis build, i suppose, then i would have to :-)
    in theory the class names/parameters of tessell's code generators are pretty stable/haven't changed in awhile either, so i don't think there are a lot of situation where they would move in lock step.
    Anton Johansson
    @anton-johansson
    Okay, that's fine. It's pretty easy to change which version of tessell the plugin should use anyway. :)
    I just need to get commit stephenh/tessell@988222e into the master branch:
    Anton Johansson
    @anton-johansson
    Speaking of which, I added a new PR to add this commit into the master-branch. I noticed I previously had a PR up, which looked rather strange, so I closed that one.
    Anton Johansson
    @anton-johansson
    Also, is it much work for you to release a new V2 release? It would be nice to have a V2 that works with the tessell-maven-plugin (currently, I've used the V2 snapshot, built locally)
    Stephen Haberman
    @stephenh
    ah yeah, i can do a v2 release...will try and do that over the next day or so...
    Anton Johansson
    @anton-johansson
    Okay, nice. No rush though. :)
    Stephen Haberman
    @stephenh
    when elemental-2 shows up, the html elements will be interfaces:
    which is great, given i'd wanted old-school gwt widgets to get interfaces, but they were never added/accepted
    so, i think spiking a "widget-less" version of tessell, that uses uibinder + jstypes only, will be interesting in the next few months.
    the html element interfaces will come for free/in elemental-2, and then i'll start adding stub implementations of all the elements.
    Anton Johansson
    @anton-johansson
    Cool. I've seen lots of comments recently about avoiding widgets.
    Stephen Haberman
    @stephenh
    yeah. we've not actually had a problem with widgets, and if anything their attach/detach events are nice. but right, they are actively discouraged, so eventually will have to use something else...
    Jim Northrup
    @jnorthrup
    Hi Stephen how far along is this?
    Stephen Haberman
    @stephenh
    @jnorthrup sorry, missed your message, but i haven't made any progress on a non-widget approach.