Recently I have been considering bringing back the managed source editor we have in MD. It needs a lot of work, especially in the areas of performance, mouse handling and scrolling, but I think with some hacking, it can be made to work really well.
There are a couple reasons for this thinking. First of all, the one piece of MonoDevelop that defines it is the source editor. I mean, all the bells and whistles are just cruft if your editor doesnt do what you need it to. Now, gtksourceview does what we need, and we have been able to extend it to do a lot of the other stuff we want it to do. However we have no control over it, and having no control over the primary piece of your application is a thorny situation. The gtksourceview guys are great, and the library is wonderful, but the ability to fix bugs and then release MD with the bugfixes is pretty damn important.
I have yet to really make a decision one way or the other, but it is absolutely on my mind.
More gettextification... introduced a fatal bug into compilation/running doing so (thanks for catching it lluis!). I have basically finished the base assembly and am now doing the source editor assembly. And I found that we can easily add translation to our addin xml files by renaming them to whatever.addin.xml and changing the parser to accept _label instead of label. Which is really really good news. Havnt tested if it will work yet, but my hopes are high, gettext seems to do a pretty damn good job.
Posted by tberman at April 4, 2004 12:17 AMI'm certainly not qualified to judge whether gtksourceview is a good long-term choice for MD... but let me get on a soapbox anyway. :)
Is this another example of development splitting into different directions to the detriment of both choices? I'm probably making the same oversimplification about labor that you see on /. all the time ("Why don't 'you' all just concentrate on one thing?")
But I'd think gtksourceview benefits from having MD as a "customer" and MD benefits somewhat from not having to implement the same functionality again.
Another issue seems to be whether it's long-term good sense for MD to use a non-managed widget. This is probably a trump card on the decision. But I'm picturing (unrealistically?) a scenario where MD benefits from improvements to gtksourceview, and gtksourceview (and other consumers of it) benefit from the features that MD demands of it.
Or maybe I haven't a clue as to what I'm talking about. :)
Posted by: Brian at April 4, 2004 05:25 PMYou could also distribute the gtksourceview source as a part of the MonoDevelop source. Then you can fix and release whenever you need to and still contribute the fixes back (though that might be a bit more complicated)
Posted by: Kenneth Christiansen at April 4, 2004 05:56 PM