Cache of pre-computed information for the
CAnnotationProcessor
. The cache is helpful to
cut substantially the time to build of an svm image, when many CAnnotation are being used. This
is the case, namely for Walnut, when the compilation time to generate the information for the C
processor is by far the dominant cost.
For simplicity, the cache cannot be updated: it can be re-created from scratch, or used. The two
options are mutually exclusive, and the re-creation of the cache erase all previously cached
information. responsibility for the accuracy of the cached content is up to the end user.
A CAP cache is just a directory with a file for each
NativeCodeInfo
, where the file is
pretty much the output of the "query" program generated by the CAnnotationProcessor.