Hi Gábor, > If we later decide that compression is desired we can still go with a text > file with references.
This wouldn’t always be feasible, because it would require *manual* reference tracking. The packager would need to know what references need to be retained and then ensure that they are part of that text file. > Anyway, do you think that extending the reference scanner would be a good > idea? We had the same problem with gzipped R files and zipped Python things. For the R files a more advanced reference scanner would not work, because the files just include a gzipped section; it’s not the whole file that’s simply gzipped. Detecting sections is too hard. But for Python and Jars it might just be enough. [PS: please don’t top-post.] -- Ricardo GPG: BCA6 89B6 3655 3801 C3C6 2150 197A 5888 235F ACAC https://elephly.net
