Hi all, Bogdan asked me to send this to devel list in case some more people will be interested this time :)
I had some problems working with core opensips / modules since there is no good way to test the code without restarting it and throwing actual messages at it. This makes conditions like out of memory, or specific transaction states very hard to reproduce (especially timing-based actions). At some point I created a script to convert opensips into an autotools build environment (can be found at https://bitbucket.org/viraptor/opensips_autocheck/changeset/489f863e3240). It also adds the libcheck framework to the build and includes one proof of concept test in check_tests/check_textops.c It should apply to one of late 1.5.x or early 1.6.x versions... (can't remember now which one) I think something like this could be a good idea for future development, especially since 2.0 is taking some shape right now. I'm not saying either autotools or libcheck is "the" choice here, but the general idea seems working and useful. I'm happy to convert the rest of modules to work with autotools if the development team wants to try this way. I think this could improve opensips in the future (give ability to create regression tests, create ways to test modules without actual traffic, benchmark module code in tight loop without the network processing overhead, add some implicit "behaviour documentation" even if it's the test code only). So... comments? ideas? how do people feel about it? (maybe I'm the only one finding sipp silly for debugging ;) ) Regards, Stan _______________________________________________ Devel mailing list [email protected] http://lists.opensips.org/cgi-bin/mailman/listinfo/devel
