GH> The file-entities.php script already distinguishes between the cygwin
GH> and the windows compiled PHP, so if the cygwin PHP is used, cygwin path GH> values are generated into the output (so that implies a cygwin installed GH> openjade). It does not need to be relative. Even if a windows PHP is GH> used, we can pass on a parameter to generate cygwin path values. Why is GH> that a problem later on?
The problem is to autodetect which openjade version we use - cygwined or not.
GH> Why? In configure, you can do quite some iteresting checks :) If the GH> ../phpdoc-tools/jade/jade.exe is used, it is not cygwin, otherwise it is.
It could work. Just don't forget about ../phpdoc-tools/openjade/openjade.exe =)
The new configure.in I committed yesterday now checks if [open]jade was found in the ../phpdoc-tools/ folder. The executable name or the subfolder further down there is irrelevant, just that it is in ../phpdoc-tools.
Goba
