Hello, I decided to rewrite the script in Perl, as adding more features become cumbersome in a shell script.
It's an ugly mess, but it works nicely... http://files.housegordon.org/tmp1/gnu_savannah_eval.pl Usage is: gnu_savannah_eval.pl --project "NAME" DIRECTORY | pandoc --from markdown --to html > 1.html Some of the added features: 1. better copyright detection 2. detect "canonical" copyright ("Copyright (C) YEAR NAME") but also non-canonical forms. gives a warning about non-canonical forms. 3. better license detection 4. tries to detect exact type and version (GPL/LGPL/AGPL 2/2.1/3 + or later) or BSD/MIT/zlib/perl/gnu-permissive/etc. 5. failing that, tries to detect some common phrases that hint about a non-standard license wording. 6. detects linux vs gnu/linux, and open-source vs free-software 7. Generates a table of all files, their detected type, license and copyright. Testing on GNU packages, there are still false-positives, but also many valid warnings: there are many files (mostly tests) which aren't copyrighted. http://files.housegordon.org/tmp1/coreutils_eval.html http://files.housegordon.org/tmp1/gawk_eval.html http://files.housegordon.org/tmp1/grep_eval.html http://files.housegordon.org/tmp1/sed_eval.html http://files.housegordon.org/tmp1/tar_eval.html And these are evaluation of currently pending projects: http://files.housegordon.org/tmp1/jnotebook_eval.html http://files.housegordon.org/tmp1/rufascube_eval.html To be continued, - Assaf.
