Thorsten Scherler wrote: > David Crossley escribi??: > > [EMAIL PROTECTED] wrote: > > > Automated build for forrest-sample-2 FAILED > > > ... > > > ... > > > [java] X [0] /icon-b.png BROKEN: > > > /export/home/config/forrestbot-trunk/conf/work/forrest-sample-2/src/documentation/content/xdocs/icon-b.png > > > (No such file or directory) > > > ... > > > > This same example works using skins. > > > > <img src="/icon-b.png" ... > > > > The resources.xmap and locationmap resolve to find > > source file in src/documentation/resources/images/ directory. > > > > However with Dispatcher it seems to fail the locationmap > > and tries to find it at xdocs/icon-b.png > > > > It is a silly example and perhaps it should be removed > > but why the difference between skins and Dispatcher? > > The main reason is that png can be generate from svg. This is controlled > by skins/dispatcher. Maybe the best would be to extract the svg > generation to a plugin of its own (it is a couple of dispatcher > contracts and structurer definitions).
But the abovementioned link works fine as-is in 'forrest run' mode. > The problem with the resources.xmap is that it is a mix between skin > specific matches (like round corner) and general core matches. This xmap > I personally consider as one of the highest complex code in our whole > code base that is the reason why I did not yet found the time to slim it > down to a resources-core.xmap and resources-skin.xmap. > > Will try to have a look tomorrow. Is that wise when we are trying to crank out the release? It works fine in html production now. BTW, i see that the dispatcher copy of resources.xmap does not use the locationmap at all, whereas the main one does. -David