Larry,
You're probably not understanding the problem correctly because I'm not
articulating it correctly. ; )
I guess I'm having some trouble envisiging possible scenarios because
I'm new to CVS:
A. (what I think you have suggested)
1. Ant calls svn to check out the *entire* project to temp dir. This
would include master file, component files, XSL files etc.
2. Ant calls db -> XML mapping tool to obtain latest extract
3. Assemble master file from components with 'faux' XInclude (XSL
stylesheets using xsl:copy, xsl:copy-of) or XInclude-aware XSL
processor
4. Process master file with XSL-T and XSL-FO processors to produce
HTML and PDF
5. Get coffee (my favourite part)
OR some of the component files might be in different svn projects /
directories, I might have to parse each XInclude individually:
B.
1. Ant opens master document (perhaps from svn)
2. Ant converts each individual XInclude element to a svn call
3. Files / projects checked out of svn into a temp directory
4. Ant calls SQL -> XML mapping tool, stores resulting XML extract in
the same temp directory
5. With all component files in place, Ant begins to process master
document with 'faux' XInclude (XSL stylesheets using xsl:copy,
xsl:copy-of) or XInclude-aware XSL processor
6. Ant stores compiled master document in same temp directory
7. Ant processes compiled master document with XSL-T and XSL-FO
processors to produce HTML and PDF
8. get coffee (my favourite part)
Does this make sense?
TIA,
Chris
Chris Johnson
Web Developer
Capilano College
North Vancouver, Canada
604.986.1911 ext. 3455
cjohnson@capcollege.bc.ca
Larry Garfield <larry@garfieldtech.com> 04/26/2005 6:00:35 PM >>>
Perhaps I'm not understanding the problem correctly, but why pull the
flat files out of svn as they're being processed? Assuming that they
won't change during the compilation, wouldn't an ant-run routine that
does the following be sufficient:
- svn update the current version of the files
- extract SQL->XML files (optional if they're as rare as you say)
- start processing the master DocBook file
- Go get coffee while it compiles
XInclude or not, I believe the entire document needs to be parsed at
once before processing begins, so you'll need all the files in place
anyway before the XSLT scripts can do their thing.
Chris Johnson wrote:
Hi,
I'm looking into using SubVersion to manage some of my docbook
files.
It's a modular document made up of ~120 XIncluded files and a couple
of
database extract (SQL -> XML) files. The flat files are edited
periodically (could be at any time), while the SQL -> XML extracts
occur
when I get the go-ahead (~4-6 times a year).
I would like to manage the flat files with SubVersion, and use XML
catalogs to abstract the calls to SubVersion and the database SQL ->
XML
process.
I envisage something like an Ant process:
- grab the the master XML file , which consists primarily of
XIncludes.
- start processing the XIncludes using XML catalog
- the XML catalog would abstract the XIncludes to both SubVersion
for
the latest commits and to SQL -> XML extract process
- assemble document
- process document to produce HTML and PDF.
Does this sound feasible? Or is there an easier way to achieve this?
Any links to docs / faq's appreciated.
Chris
Chris Johnson