>> If you pull up $WRKSRC/cocci.ml, then you get to the possible root of the >> problem, see around line 485. >> >> egrep fails because it relies on \b which is not supported in BSD grep? > > Good catch. > > Below is a diff against stuarts port. I also tried a different one > using gegrep (from sysutils/ggrep) instead of egrep, and got the > same output from make regress in both cases.
Cool, I integrated your diff and found another case of gdiff which was blocking the simple.cocci test. Now everything works fine, I checked on simple.cocci, proc_info.cocci (which is biggest size cocci file). Have a go at it. I will try to feed your diff upstream and hope it can get integrated into RC2. I will cc ports@ on that. I changed the DESC file to copy the paragraph from their website (inserted INRIA, their lab where it wasn't clear who 'us' refers to)... Attaching the complete port for testing. thanks, amit
cocci.tgz
Description: GNU Zip compressed data
