Add User Testing interface to impl/memory.

Add User Testing interface to impl/memory.

R=dnj@chromium.org, estaab@chromium.org, vadimsh@chromium.org
BUG=https://github.com/luci/gae/issues/24

Committed: https://github.com/luci/gae/commit/f45f7f7faae7017ece70eb35666ddfcbb33696c3

[iannucci, 2015-12-13 19:54:17]

On 2015/12/13 at 19:51:35, iannucci wrote:
> 

No tests yet, put PTAL for general idea. There are some TODO's in the memory
impl... I'm not sure if what I have there is sufficient for testing, especially
around the oauth stuff. I think I need one more interface for doing something
with oauth scopes or something, but I don't know how that stuff works, so I'd
really appreciate some suggestions :)

If it looks good then I'll add some tests n' stuff to make sure it works ok.

[Vadim Sh., 2015-12-13 21:20:48]

https://codereview.chromium.org/1522783002/diff/20001/impl/memory/user.go
File impl/memory/user.go (right):

https://codereview.chromium.org/1522783002/diff/20001/impl/memory/user.go#new...
impl/memory/user.go:45: return u.data.user
return nil if user.ClientID != "" (to mimic behaviour of real API)

https://codereview.chromium.org/1522783002/diff/20001/impl/memory/user.go#new...
impl/memory/user.go:52: return u.data.user, nil
return nil if user.ClientID == "" (to mimic behaviour of real API)

https://codereview.chromium.org/1522783002/diff/20001/impl/memory/user.go#new...
impl/memory/user.go:75: func (u *userImpl) OAuthConsumerKey() (string, error) {
I think this thing is also deprecated (as the rest of OAuth 1.0 stuff). Not
sure. At least I've never used it.

https://codereview.chromium.org/1522783002/diff/20001/impl/memory/user.go#new...
impl/memory/user.go:107: ID:       hex.EncodeToString(id[:]),
in real life it is usually int64 (as a string)

https://codereview.chromium.org/1522783002/diff/20001/impl/memory/user.go#new...
impl/memory/user.go:108: ClientID: hex.EncodeToString(clientID[:]),
in real life this is set only for return value of CurrentOAuth. In fact, some
checks in luci-go/server/auth rely on this (IIRC).

Also its value is sort of important (using in various access checks, just like
email). Probably better to make it an argument of Login?

[iannucci, 2015-12-14 08:49:01]

better?

https://codereview.chromium.org/1522783002/diff/20001/impl/memory/user.go
File impl/memory/user.go (right):

https://codereview.chromium.org/1522783002/diff/20001/impl/memory/user.go#new...
impl/memory/user.go:45: return u.data.user
On 2015/12/13 at 21:20:48, Vadim Sh. wrote:
> return nil if user.ClientID != "" (to mimic behaviour of real API)

done

https://codereview.chromium.org/1522783002/diff/20001/impl/memory/user.go#new...
impl/memory/user.go:52: return u.data.user, nil
On 2015/12/13 at 21:20:48, Vadim Sh. wrote:
> return nil if user.ClientID == "" (to mimic behaviour of real API)

done

https://codereview.chromium.org/1522783002/diff/20001/impl/memory/user.go#new...
impl/memory/user.go:75: func (u *userImpl) OAuthConsumerKey() (string, error) {
On 2015/12/13 at 21:20:48, Vadim Sh. wrote:
> I think this thing is also deprecated (as the rest of OAuth 1.0 stuff). Not
sure. At least I've never used it.

made it return an error, since we don't know how to mock it correctly.

https://codereview.chromium.org/1522783002/diff/20001/impl/memory/user.go#new...
impl/memory/user.go:107: ID:       hex.EncodeToString(id[:]),
On 2015/12/13 at 21:20:48, Vadim Sh. wrote:
> in real life it is usually int64 (as a string)

done

[Vadim Sh., 2015-12-14 18:54:08]

https://codereview.chromium.org/1522783002/diff/40001/impl/memory/user.go
File impl/memory/user.go (right):

https://codereview.chromium.org/1522783002/diff/40001/impl/memory/user.go#new...
impl/memory/user.go:104: clientID := hex.EncodeToString(clientIDRaw[:])
but now you can't mock non-OAuth case with Login(). Just make:

func (....) Login(email string, admin bool, clientID string) string

[iannucci, 2015-12-14 19:40:48]

On 2015/12/14 at 18:54:08, vadimsh wrote:
> https://codereview.chromium.org/1522783002/diff/40001/impl/memory/user.go
> File impl/memory/user.go (right):
> 
>
https://codereview.chromium.org/1522783002/diff/40001/impl/memory/user.go#new...
> impl/memory/user.go:104: clientID := hex.EncodeToString(clientIDRaw[:])
> but now you can't mock non-OAuth case with Login(). Just make:
> 
> func (....) Login(email string, admin bool, clientID string) string

how about this? if behavior/ui lg, I'll add tests.

[iannucci, 2015-12-14 21:00:33]

Added tests, PTAL

[Vadim Sh., 2015-12-14 21:01:38]

lgtm

[iannucci, 2015-12-14 21:02:06]

The CQ bit was checked by iannucci@chromium.org

[commit-bot: I haz the power, 2015-12-14 21:02:27]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-12-14 21:02:28]

This CL has an open dependency (Issue 1521823003 Patch 1). Please resolve the
dependency and try again.

[iannucci, 2015-12-14 21:02:46]

The CQ bit was checked by iannucci@chromium.org

[iannucci, 2015-12-14 21:02:47]

The patchset sent to the CQ was uploaded after l-g-t-m from vadimsh@chromium.org
Link to the patchset:
https://chromiumcodereview.appspot.com/1522783002/#ps120001 (title: "rebase")

[commit-bot: I haz the power, 2015-12-14 21:02:50]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1522783002/120001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1522783002/120001

[commit-bot: I haz the power, 2015-12-14 21:07:36]

Description was changed from

==========
Add User Testing interface to impl/memory.

R=dnj@chromium.org, estaab@chromium.org, vadimsh@chromium.org
BUG=https://github.com/luci/gae/issues/24
==========

to

==========
Add User Testing interface to impl/memory.

R=dnj@chromium.org, estaab@chromium.org, vadimsh@chromium.org
BUG=https://github.com/luci/gae/issues/24

Committed:
https://github.com/luci/gae/commit/f45f7f7faae7017ece70eb35666ddfcbb33696c3
==========

[commit-bot: I haz the power, 2015-12-14 21:07:36]

Committed patchset #7 (id:120001) as
https://github.com/luci/gae/commit/f45f7f7faae7017ece70eb35666ddfcbb33696c3