> Has anybody in the gcov developer community pondered or done such a > thing?
I have pondered it, yes. Although it might not be a particularly helpful answer, have you considered using other tools to get what you want rather than modify gcov and the runtime code? You could run the instrumented binary with (say) one set of options, run gcov, gather the coverage output, and then merge it into an "accumulated" report that includes what you want. Repeat. The main downside is that you have to run gcov over ever source file after each binary invocation and then process the output. Cheers, Ben -- Ben Elliston <[EMAIL PROTECTED]> Australia Development Lab, IBM
