On Sun, Mar 4, 2012 at 10:34 AM, Nasir Khan nasir8891@gmail.com wrote:
Hi,
From the very beginning, there is a problem in the PDF Download tool and that is it can not properly render the Bengali and some of the other Indic languages (might be all Indic languages). A few bug was reported earlier and the developers tried to solve the issues but unfortunate it was not completely fixed.
The PDF library used with the Collection extension of Mediawiki is reportlab. Reportlab uses its own internal rendering engine and not capable of handling complex scripts like Indic. Such a rendering engine will not scale to meet the requirements of hundreds of scripts/languages that we support.
Here i have a question that at this moment is there any developer of developer group is working on this issue? if yes then i want to join with him.
I had discussed this with Pediapress developers in the past. I started a project many months back to develop a general purpose PDF rendering library for complex scripts and I had announced it here: http://lists.wikimedia.org/pipermail/wikimediaindia-l/2011-February/002198.h... And I have an online version based on the code in development http://silpa.org.in/Render. It can create PDFs from Bengali wiki pages and possibly most of the scripts that we support. The text rendering engine used is Pango. This is my petproject and 2-3 people joined me, but recently it is inactive in development since we are not getting free time. The library require lot of work to reach a reasonably usable state. Then it require good amount of effort to integrate with collection extension.
The project is hosted at https://savannah.nongnu.org/projects/pypdflib and available in Debian.
Thanks Santhosh