Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Understanding GCOV outputs

512 views
Skip to first unread message

spectre

unread,
Jan 23, 2001, 2:32:21 PM1/23/01
to
I'm having a hard time understanding what the verbage in the output from
gcov means: "x%/5 branches executed in function foo" verses "x%/5 branches
taken at least once in function foo". Exactly what does "executed" mean
versus "taken at least once"?

Additionally, in the source that gets outputted, there doesnt seem to be a
clear definition of whether a decision went true or false and how many
times it went what way. Is there a way to determine that from the source
output? Or am I better off buying a comercial tool like C-Coverage from
Bullseye software?

spectre

unread,
Jan 24, 2001, 1:43:20 PM1/24/01
to
Well, can someone at least tell me how to ask the maintainer of gcov? Or
is this just an example of where a commercial product is superior to the
GNU one and forget about trying to use gcov for decision coverage testing?

Joe Durusau

unread,
Jan 24, 2001, 6:32:37 PM1/24/01
to
I don't know who maintains gcov, but in more-or-less standard
syntax, 'executed' on a branch instruction means that the program
got to the branch. 'taken' means that it was a conditional branch,
which at the hardware level generally amounts to 'go straight on
to the next instruction or go elsewhere', and that the elsewhere
path was followed.

Speaking only for myself,

Joe Durusau

PS for more detail, see the gcov source. It should clarify the matter.

0 new messages