|
From: Bruce M. <br...@mc...> - 2003-05-23 04:04:07
|
This what I have been doing in the MAIN branch: 1. Added the com.babeldoc.core.pipeline.feeder package. This package contains a number of classes that making adding documents to pipelines easier. The fundamental objects are SynchronousFeeder and AsynchronousFeeder. The Sync. feeder feeds in the documents to the pipeline and then waits until that processing is done before accepting another document. The Async. feeder submits all the documents to a queue asynchronously. Then there is a threaded implementation for sucking them off the queue and running them through the pipelines. Please be aware that the multithreaded nature of babeldoc is not guaranteed although Erik and Dejans fine work has remedied most of the problems. 2. Split off the xslfo and svg module from core. This can make realworld babeldoc deployments smaller if they do not require these functions. regards, Bruce McDonald. PS. Erik - I will revert the nextStage issue in the 1.0 branch. I have not decided which approach is better for the long run. If we can remedy this using a generalized error handling design. I also put in a word for you re. the opportunity we spoke about. The job is in MD / Wash DC - waiting for the boss to speak. Possible telecommuting. Keep in touch. PPS. Dejan - I have finished your dependency so that you may go full steam with your scanner changes - I am excited to see what you got in your bag of tricks. PPPS. As always YIM me during the day. Please note the badly misconfigured / broken proxy firewall sometimes messes with YIM chats. So if I dont answer - repeat the message. |