Update: I found some time to finally implement the word breaking algorithm (excluding dictionary lookups). It's something I've been postponing for a while now, and your email reminded me how needed it is.
I'll hopefully get everything to a working state soon. Will update you when I do. -- Tom. On 07/12/11 11:04, Tom Hacohen wrote: > Dear MJ, > > Thanks a lot for the patch, this is a very needed feature, however I > have a couple of comments on the implementation. > > The reason I was postponing implementing word selection for entry, is > that it's not that trivial. > > There's a complete standard on word boundaries that has to be > implemented (http://unicode.org/reports/tr29/#Word_Boundaries) for it to > work correctly for all languages. > > I'm not against implementing an incomplete implementation (i.e one that > just filters according to some chars) until we get around to doing the > complete implementation, but in that case, > edje_object_part_text_word_separators_set must be removed, as it won't > be useful in the future. We need to make the incomplete implementation > look like a complete implementation and just behave differently in the > details. Personally, I prefer having an implementation inside textblock > as it has more information, something like: "cursor_to_word_end" and > "cursor_to_word_start". > > Also, I can see you changed the way _curs_back behaves, but judging from > the textblock code, your change is not needed as it's already like that > there... Also, I don't see how it relates to your change because a > paragraph separator is a word separator anyway. > > The comments above are especially relevant for languages that don't use > spaces that often, or that break words without hyphens. > > I believe that using the line breaking algorithm (that we already have) > will give better results than working with delimiters (although it won't > be perfect itself). > > > In conclusion, I wouldn't mind applying a patch that just automatically > handles this (without exposing any API to the user) as first step > towards a complete implementation, but I do believe implementing it > inside textblock as described above is a better solution. > Therefore, this patch can't be applied in it's current form, but needs > to be adjusted as Vincent suggested in his mails, and as I suggested above. > > > Waiting to hear your thoughts on the matter. > > -- > Tom. ------------------------------------------------------------------------------ Learn Windows Azure Live! Tuesday, Dec 13, 2011 Microsoft is holding a special Learn Windows Azure training event for developers. It will provide a great way to learn Windows Azure and what it provides. You can attend the event by watching it streamed LIVE online. Learn more at http://p.sf.net/sfu/ms-windowsazure _______________________________________________ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
