[tei-council] Guidelines Building

Sebastian Rahtz sebastian.rahtz at it.ox.ac.uk
Thu Dec 19 09:59:11 EST 2013


On 19 Dec 2013, at 13:51, Martin Holmes <mholmes at uvic.ca> wrote:
> 
> 
> I think the problem (correct me if I'm wrong) is the coupling of the 
> stylesheet build with the P5 build. Jenkins always looks to the latest 
> successful build of the Stylesheets project for its XSL, but if you 
> build locally, by default you're using the Stylesheets in the installed 
> TEI Debian packages. Stylesheet development goes hand-in-hand with P5 
> development, so unless you check out the latest Stylesheets code from 
> github and build that, then point your make at the results, there's a 
> strong possibility things will fail.

i think thats a bit of an exaggeration. “small” rather than “strong”.
As it happens, Syd had to make a change, which has this unfortunate side
effect, but that should be very rare these days (I try hard to be good about it).

> Another possibility is to have the local build process point at the 
> lastSuccessfulBuild of the Stylesheets on one of the Jinks machines. 
> That would presumably work for anyone working only on the trunk in SVN 

true, you could make that the default in the P5 Makefile, not
a bad idea, so that it mimics what Jenkins does.

however, just tested it with

 make XSL=http://bits.nsms.ox.ac.uk:8080/jenkins/job/Stylesheets/lastSuccessful/artifact/dist/xml/tei/stylesheet/

and it doesn’t work:

/Volumes/Repo/TEI/Sourceforge/tei-code/P5/antbuilder.xml:58: stylesheet /Volumes/Repo/TEI/Sourceforge/tei-code/P5/http:/bits.nsms.ox.ac.uk:8080/jenkins/job/Stylesheets/lastSuccessfulBuild/artifact/dist/xml/tei/stylesheet/odds/odd2json.xsl doesn't exist.

anyone care to diagnose that?

Sebastian



More information about the tei-council mailing list