Details
-
Type: Bug
-
Status: Resolved (View Workflow)
-
Priority: Major
-
Resolution: Fixed
-
Affects Version/s: 1.5.11
-
Fix Version/s: None
-
Component/s: filters
-
Labels:None
Description
I suspect this to be a filter issue of some sort (as in a problem in filters themselves) but it may instead be a bug in diatheke. In either case, some extensive testing is necessary to determine where the problem(s) lie.
From IRC:
[13:49] <refdoc> Osk - another thing - I am playing with writing filters
[13:49] <refdoc> one problem is that the diatheke seems not to do much with them despite the -o switch
[13:49] <refdoc> e.g arabic shaping
[13:49] <refdoc> does zilch
[13:50] <Osk|log> yeah, I've heard that there are issues
[13:50] <Osk|log> with Diatheke and filters
[13:50] <Osk|log> are we storing shaped Arabic though? if it's already shaped, the filter shouldn't do anything
[13:50] <refdoc> no
[13:51] <refdoc> we store unshaped
[13:51] <refdoc> so it should shape
[13:51] <Osk|log> and Diatheke is compiled with ICU support?
[13:51] <refdoc> yes
[13:52] <refdoc> and the arabic vowel strip filter (copied off you Hebrew points) is doing not much in the diatheke
[13:52] <refdoc> despite me integrating it in exact teh same fashion as the hebrew points
[13:53] <refdoc> part of teh problem might be that I am not filtering enough
[13:53] <refdoc> (the abundance of Arabic script diacritics
[13:53] <refdoc> is confusing)
[13:54] <refdoc> but part is right now that diatheke is not doing much with the filters in the first place
[13:54] <refdoc> even the existing ones (not just my new one)
Format filters (e.g. to HTML/HTMLHREF) also need some checking.