Code review comment for lp://staging/~jameinel/subunit/omit-times-623654

On Sat, Oct 23, 2010 at 3:11 PM, John Arbash Meinel
<email address hidden> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 10/23/2010 5:29 AM, Jonathan Lange wrote:
>> On Fri, Oct 22, 2010 at 7:02 PM, Robert Collins
>> <email address hidden> wrote:
>>> Right, thats clearly bong.
>>> I'd like to see, I think:
>>> time:
>>> time:
>>> test: event
>>> time:
>>> error: event
>>> time:
>>> time:
>>
>> I don't see at all what having two "time:" statements here would mean.
>> In subunit protocol terms it means exactly the same as having only the
>> second time statement.
>>
>> jml
>>
>
> Not entirely. I believe there are test runners that tell you how *long*
> an event took. (bzr selftest -v). In which case doesn't it subtract the
> known time from the next time it sees?
>

From the README:
"The time directive acts as a clock event - it sets the time for all future
events. The value should be a valid ISO8601 time."

time: A
test: X
time: B
success: X
time: C
time: D

Thus, the net effect of two consecutive time events is to set the time
for all future events to whatever value was provided with the second
time event. The first is ignored.

jml