What I have in mind is a repository on opensolaris.org, and a child workspace on whatever system you generate the PDFs on. Think of the workspace as a semi-permanent staging directory for the PDFs.
If your current process only updates the PDFs where there was a source change, then you would copy those new PDFs to the directory. (If you're currently regenerating all the PDFs, we can surely figure out a way so that only the revised documents actually get a new PDF.) Then, instead of tarring up the directory, do "hg commit -m'snapshot of YYYYMMDD'; hg push" or "svn commit -m'snapshot of YYYYMMDD'". At that point you're basically done--the SCM software will make a note of the changed files and ship them to the server. The only thing manual step I can think of is when you add a document--you'll need to tell the SCM software to track the new PDF. But that's a one-time operation. Rainer> My script runs on everything in the (local to me) directory the Rainer> XML sources go into. Sounds like a job for make(1), then. mike _______________________________________________ website-discuss mailing list [email protected]
