thank you for the links. But since PDF indexing is a feature of Jahia, I hope that the Jahia team is still working on resolving issue in their software (pdf support). It's not about performance, but failures we talk. If pdfbox is just crap, it would be a good idea to at least provide adapters for other engines that support productive environments. For me it's like having HSQL DB as a free and instarun database for Jahia (which is good) and ALSO have additional support from scratch for MySQL and Oracle (which is absolutely required).
Warning, warning, dangerous ground ;-) :
Jahia is released under an open and community based philosophy (collaborative sourcing = nearly open sourcing). So the paradigm is to try to offer the best prices and to commoditize the high-end/mid-range CMS and Portal market. The cons is that we need to rely on other "open source" layers which we integrate to avoid certain R&D or OEM costs. Same is true for RedHat/Suse or others: They can try to help the community improve certain layers with their own dedicated staff but they are not Microsoft, they do not have control on the full Linux layers they package. Then they also have to rely/wait after the work done by the community and follow their decisions/priorties (or mandate someone to do a custom important improvement but then indirectly bill it to the end-customers through some "expensive" support programs).
So regarding bug fixes in the Jahia core kernel, I think our history (cf: changelog Jahia 4: http://www.jahia.org/download/jahia4/4_0/pr/history.txt) speaks for itself and I doubt lots of editors offer better and faster bugs resolutions.
Regarding supporting third party layers and/or adding support for other commercial products, this is more complex. What if there is a bug in Tomcat. Because we package it by default, this would mean we also need to support it? So ideally, we should have some neutral abstraction layers everywhere in the code in order for the customer to use any commercial or open source layer he wants to replace. Practically this is not so easy to do especially when the external contributor did not think about it while developing his extension. Moreover IMHO there are other more strategical "abstraction layer" we need to implement before the PDF plug-in one (e.g. the persistance layer with Hibernate or Apache OJB, the search engine layer with other tools such as Verity, the caching layer with tools such as Tangosol, etc..).
So then it is a question of priority... But here again, Jahia is released under a collaborative source license. If THIS is a priority for you, please do not hesitate to develop it or sponsorize the integration of another PDF parsing library in exchange of a license discount... In opposite to other proprietary license, you have this choice and this right.
Peace
St�phane
