[Buildbot-devel] About the front end part of buildbot
tardyp at gmail.com
Wed Mar 11 14:10:09 UTC 2015
Le mer. 11 mars 2015 à 10:31, chase <i at chasezhang.me> a écrit :
> Hi Pierre,
> I am Chase, whose GitHub name is *shanzi*. I am studying on the front end
> part of buildbot and here are some questions I’d like to discuss with you.
Hi Chase, please use the buildbot mailing list for this kind of discussion.
> I noticed that you wrote Guanlecoja
> <https://github.com/buildbot/guanlecoja> and applied it in buildbot. It
> seems that Guanlecoja is very similar to yeoman <http://yeoman.io/>.
It is explained in guanlecoja documentation. guanlecoja is a ready to use
build system for coffeescript, jade, bower, etc. It actually has a plugin
> I have some experiences with yeoman so I’m wondering if the two are the
> same. So could you provide more instructions about working with it? For
> 1. Are we supposed to modify guanlecoja/config.coffee ourselves?
> yes, config.coffee is the file where you configure guanlecoja for
overriding default config. This is especially where you describe your bower
dependancies. There is no bower.json with guanlecoja projects, it is
actually generated from the onfig.coffee.
> 1. You included bootstrap in Guanlecoja, but not
> <http://sdorra.github.io/angular-dashboard-framework/>. Should we add
> them to Guanlecoja or directly in buildbot if I’d like to finish #3147
> #3147 is not started, so yes, you would need to include adf inside the
> 1. Angular.ui <https://angular-ui.github.io>’s ui-bootstrap
> <http://angular-ui.github.io/bootstrap> component has ported many
> components of bootstrap to Angular. What about add it to Guanlecoja? It
> seems that Angular.ui is quite similar to Guanlecoja-ui
> ui-bootstrap is already part of guanlecoja-ui.
Guanlecoja-ui regroups reusable ui components needed for buildbot and other
dashboard application that. It provides some integration with ui-router,
and ui-bootstrap, as well as a side menu system, notifiation system, etc.
You can see it as a higher level reusable components that bootstrap-ui, and
> The questions above are about keep developing on the base of current
> structures. And I have yet another idea.
> That is, we don’t use bootstrap any more and migrate totally to Angular
> Material <https://material.angualrjs.org>. The reasons to do so:
> 1. Officially provided by Angular Community, Angular Material works
> very well with Angular.js.
> 2. It contains most things that provided by Guanlecoja-ui and many
> more components that we will need in the future.
> That is not true, as guanlecoja-ui is much more high level as md. You will
see that if you dig more into the code.
> 1. The UI style is more elegant (in my opinion) than bootstrap. And as
> bootstrap has been abused so much, material design might be a good
> 1. The dashboard sortable effect of Angular-dashboard-frameworks seems
> can be easily implemented with ng-sortable
> <https://github.com/a5hik/ng-sortable>. We can use the containers
> provided by Angular Material along with ng-sortable to build a beautiful
> dashboard-like UI.
> This is only one aspect of the problem. It is not only about creating
beautiful UI, it is also about having configurable and reusable widget
components. This is what for me is interesting in adf.
> 1. An icon font is used as the icon set on the page. But I don’t think
> it is the best choice. I am very familiar to Web Icon Font (I even write a icon
> font builder <https://github.com/shanzi/myicons> myself) but according
> to this article <https://css-tricks.com/icon-fonts-vs-svg/> from css
> tricks, SVG is obviously a better choice. Angular Material provide a
> directive md-icon to load SVG icons which is the most promising way to
> display icons.
> 2. The color styles of Angular Material is highly customized. I have
> had a good experience with it in my last project.
> So what’s your opinion? Is it practical for us to migrate to Angular
I follow with attention the progress of angular Material. Last time I
checked, this was very young beta project. Now with the ng-conf finished
last week, it appears to have much more stability.
The problem I can see is that we would have to rewrite the whole UI.
Designing the current already took quite long, and if we recreate it each
time there is a new web design technology, we never will release a working
So I would be more encline to publish a bootstrap version, and wait a
little bit so that there is a strong community around md, a community as
strong as bootstrap.
That said, I am very happy that you are making such kind of proposal. I
encourage you to make me change my mind.
If you can make a proof of concept quickly with the build
<http://nine.buildbot.net/#/builders/1/builds/118> page converted to
material design, I think that would be a great GSOC project.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the devel