By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement . We’ll occasionally send you account related emails.

Description:
Will CPD's "--exclude" option support regular expression some day? Using regular expression to skip paths is so much more effective than listing all the folders I want to exclude.

For instance, a large JavaScript project can include tons of *.min.js files in different directories. We won't want these files to be scanned. Using regular expression for "--exclude" argument can easily solve this problem. Otherwise, we have to find all the *.min.js files first, then combine them into a long argument for the cpd command.

I don't think regular expressions would be a good fit here, but maybe using Ant paths would be better… it tells very explicitly how to glob and what to leave out, it's flexible, but not as flexible / obscure as regexps, and it's going to require a lot less escaping.

What do you think?

Another problem I encountered here is: the "--exclude " option does not support directories with special characters, such as Chinese.
I tried --exclude /data/iccen/test_exclude/core-js/test1- 副本.js , and the file is still scanned :(

This would be really helpful for languages where nominally excluded files, like tests, are intermixed with code (Go, for example). Unfortunately --exclude *_test.go (or variations) simply results in [main] WARN net.sourceforge.pmd.cli - No such file *_test.go .

The best option to dynamically find code files is something like:

find . -type f -name "*.go" ! -name "*_test.go" > go_files.txt
pmd cpd --file-list="go_files.txt" --language="go"

It was actually pretty easy to exclude both individual files and full paths with a specific string in the from the report:

find . -type f -name "*.php" ! -name "*settings.php" ! -name "*settings*php" ! -name "_ping.php" ! -name "*Test.php" -not -path "*tests*" > non-test-php-files.txt
pmd cpd --minimum-tokens 100 -f csv -l php -z . --file-list non-test-php-files.txt  > code-all-duplicates.csv