09 Sep 2005

I cannot agree more with what tml said.

In fact to properly do caseless string search/comaprison in utf8 you also need normalization beside casefolding.
GtkSourceView supports caseless search in a TextBuffer, but not without headaches and some corner cases are still not ok. For instance bug 168247 and bug 303239.

This reminds me of Project Ridley: caseless search is something that definately belongs in GtkTextView itself, however it’s not something as simple as cut&pasting a bunch of code from GtkSourceView to GtkTextView: the current implementation of caseless searching works “well enough” for us, but it operates on the whole buffer which is not acceptable for something as widely used as Gtk. Thus the Right Way ™ to do things is to implement incremental utf8 normalization support in glib. If someone tackles that, he’ll be my hero :)

Leave a Reply