Red Hat Bugzilla – Bug 612819
Clean_ids should retain the chapter name in the id
Last modified: 2010-11-23 23:18:17 EST
Description of problem: When publican Clean_ids is run it truncates section ids with the result id=Book_Name-section_Name. For some book, this causes a clash (duplication) of ids resulting in an inability to build the book. Version-Release number of selected component (if applicable): version=2.1 How reproducible: When section names are identical across book chapters Steps to Reproduce: 1. create a book with the same section name in two separate chapters: Introduction Other relevant resources on the topic Using TestNG project Other relevant resources on the topic 2. run publican Clean_ids Actual results: The build fails Expected results: The build succeeds Additional info: Retaining the chapter name in the truncated section ids will minimise the possibility of id clashes (duplication).
There's only so much clean_ids can do; I've added the following warning to the Publican User Guide: ================ Important — ID conflicts can occur The clean_ids command is intended to facilitate building a DocBook structure around documents ported from other formats such as HTML. However, clean_ids is file-based and and only has access to information in the XML file that it is currently processing and to the document name. Therefore, nodes of the same type that have the same title receive the same IDs. These duplicate IDs will prevent the document from building. Use the clean_ids command to assist you in laying out your document, but expect that some manual adjustment to IDs might be necessary. We recommend that you do not run clean_ids on an already well established document. ================
Since this is outside the scope of how clean_ids should work, closing as NOTABUG