Add "filters/count" filter.

Add "filters/count" filter.

This filter allows you to track the number of times a give API has been
invoked, as well as the number of errors that the API has returned. Will mostly
be used for testing.

R=dnj@chromium.org, estaab@chromium.org, martiniss@chromium.org, vadimsh@chromium.org
BUG=

Committed: https://chromium.googlesource.com/infra/infra/+/d457a6aaf151534810385ddec52b97a15aa5ccb1

[iannucci, 2015-07-16 23:47:38]

PTAL, this one is pretty boring too (mostly copy-pasta).

[Vadim Sh., 2015-07-16 23:56:12]

lgtm sans unsynchronized ++

https://codereview.chromium.org/1239103003/diff/1/go/src/infra/gae/libs/gae/f...
File go/src/infra/gae/libs/gae/filters/count/count.go (right):

https://codereview.chromium.org/1239103003/diff/1/go/src/infra/gae/libs/gae/f...
go/src/infra/gae/libs/gae/filters/count/count.go:51: fullName :=
runtime.FuncForPC(pc).Name()
I wonder how fast call stack introspection is in Golang?

With a little bit of copy-pasta it can be eliminated here (just remove the magic
and pass function name explicitly as a string).

https://codereview.chromium.org/1239103003/diff/1/go/src/infra/gae/libs/gae/f...
go/src/infra/gae/libs/gae/filters/count/count.go:68: ce.Successes++
this should be under the lock or updated using atomic instruction. Consider two
goroutines calling the same datastore function.

[iannucci, 2015-07-16 23:58:18]

I think the stack inspection is OK (it's just for testing after all), but I
think I'm going to replace the map with a straight-up list of embedded
structures (then I can do sync atomic ++ on them), and would make fewer strings
to mess with.

[iannucci, 2015-07-17 00:22:52]

PTAnL, it's more differenter now. No stack walking, no maps, no mutex.

[iannucci, 2015-07-17 00:23:10]

On 2015/07/17 00:22:52, iannucci wrote:
> PTAnL, it's more differenter now. No stack walking, no maps, no mutex.

(and best of all, more copypasta!)

[Vadim Sh., 2015-07-17 00:41:27]

lgtm

imho it's much cleaner

[iannucci, 2015-07-17 00:44:39]

On 2015/07/17 00:41:27, Vadim Sh. wrote:
> lgtm
> 
> imho it's much cleaner

Agreed... though it kinda makes me want to generate the code from the
interfaces... probably better that I don't.

[iannucci, 2015-07-17 00:44:43]

The CQ bit was checked by iannucci@chromium.org

[commit-bot: I haz the power, 2015-07-17 00:44:55]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1239103003/20001

[commit-bot: I haz the power, 2015-07-17 00:48:11]

Committed patchset #2 (id:20001) as
https://chromium.googlesource.com/infra/infra/+/d457a6aaf151534810385ddec52b9...

[nodir, 2015-09-19 05:45:49]

nodir@chromium.org changed reviewers:
+ nodir@chromium.org

[nodir, 2015-09-19 05:45:50]

https://codereview.chromium.org/1239103003/diff/20001/go/src/infra/gae/libs/g...
File go/src/infra/gae/libs/gae/filters/count/count.go (right):

https://codereview.chromium.org/1239103003/diff/20001/go/src/infra/gae/libs/g...
go/src/infra/gae/libs/gae/filters/count/count.go:26: func (e *Entry) up(errs
...error) error {
why do you accept a slice but you use only first element if present?