For purposes of text extraction from PDF files, glyphs should be named in accordance with Adobe Glyph Naming convention (http://www.adobe.com/devnet/opentype/archives/glyph.html). Most glyphs in Lohit fonts follow this convention properly, but some do not. I did not do an extensive review, but I noticed several occurances of glyph names like: u0919_u094D.half_u0915_u094D.half.half Which is wrong, since step 1 of the mapping algorithm in the above link will drop the part of glyph name after the first occurence of a peroid, so only u0919_u094D will remain which is assume is not what is wanted here. A proper name would then be: u0919_u094D_u0915_u094D.half.half.half or something like that (the part after the peroid is completely ignored, so it can be anything).
Thanks for raising this issues. This is really important. I will try to fix this in coming few months. Gone through the adobe AGL. So name should be "string_text_with_underscore"."end_here"
This bug appears to have been reported against 'rawhide' during the Fedora 20 development cycle. Changing version to '20'. More information and reason for this action is here: https://fedoraproject.org/wiki/BugZappers/HouseKeeping/Fedora20
With the Lohit2 project we are redoing all stuff. Just done release of Lohit-Devanagari 2.91.0, it fixes this issue. https://www.redhat.com/archives/lohit-devel-list/2013-October/msg00001.html
This bug has been resolved. with latest Lohit releases.