1 2012-10-30T00:02:23 <sunu> ThomasWaldmann: if we go directly to that url then we'll miss the login_submit flag. Hence handle_login won't be fired
2 2012-10-30T00:03:10 <sunu> or we can set the flag may be. let me try.
3 2012-10-30T00:05:54 <ThomasWaldmann> you just give some appropriate url to the gae login as return url
4 2012-10-30T00:06:12 <sunu> yes. i get it.
5 2012-10-30T00:08:04 <ThomasWaldmann> btw, put a comment there, what you are doing is slightly non-obvious :)
6 2012-10-30T00:08:27 <sunu> ThomasWaldmann: see now.
7 2012-10-30T00:08:45 <sunu> Oops missed the comment.
8 2012-10-30T00:10:06 <ThomasWaldmann> and iirc .show_root is enough
9 2012-10-30T00:10:54 <sunu> No, we need the login_submit to make sure handle_login is executed.
10 2012-10-30T00:11:38 <sunu> otherwise the login won't be called at root and the user won't be created.
11 2012-10-30T00:12:28 <ThomasWaldmann> sure. i mean using .show_root instead of frontend.show_root
12 2012-10-30T00:12:34 <sunu> Oh sorry you were talking about frontend.show_root :(
13 2012-10-30T00:13:08 <sunu> ThomasWaldmann: yes , I got that a bit late :)
14 2012-10-30T00:13:40 <sunu> let me update with a comment.
15 2012-10-30T00:13:54 <ThomasWaldmann> also look at superfluous imports
16 2012-10-30T00:18:41 <sunu> ThomasWaldmann: look now
17 2012-10-30T00:21:36 *** dwcramer
18 2012-10-30T00:23:33 <ThomasWaldmann> sunu: commented
19 2012-10-30T00:23:49 <ThomasWaldmann> btw, can you deploy next try to gae?
20 2012-10-30T00:24:58 <sunu> yeah already doin that.
21 2012-10-30T00:25:10 <sunu> doing*
22 2012-10-30T00:25:57 <ThomasWaldmann> i guess we can take the code soon :)
23 2012-10-30T00:26:51 <sunu> ThomasWaldmann: working fine. Have a look if you want http://moin-test.appspot.com/Home
24 2012-10-30T00:29:24 <ThomasWaldmann> not quite
25 2012-10-30T00:29:47 <ThomasWaldmann> it doesn't show the root item when simulating the login at the root url
26 2012-10-30T00:30:59 <sunu> yeah didn't notice that.
27 2012-10-30T00:31:27 <ThomasWaldmann> ehrm, you redirect to +login
28 2012-10-30T00:31:37 <ThomasWaldmann> (as return_to)
29 2012-10-30T00:33:18 <sunu> The redirection is not done because the gae_user is already there
30 2012-10-30T00:34:58 <ThomasWaldmann> i click on logout and i am logged out
31 2012-10-30T00:35:38 <ThomasWaldmann> then clicking login and i am immediately logged in (i think because i said that that site might use my account a few days ago)
32 2012-10-30T00:35:47 <ThomasWaldmann> http://moin-test.appspot.com/%2Blogin?login=1&login_submit=1 and i am at that url
33 2012-10-30T00:36:54 <sunu> ThomasWaldmann: the immediate login is because we don't clear gae_user at log out
34 2012-10-30T00:37:23 <ThomasWaldmann> clear what exactly?
35 2012-10-30T00:38:02 <sunu> I mean the users.get_current_user()
36 2012-10-30T00:38:49 <sunu> it's not cleared if we don't redirect the user to users.create_logout_url
37 2012-10-30T00:39:24 <ThomasWaldmann> ehrm, didn't you have that code already for logout?
38 2012-10-30T00:41:49 <ThomasWaldmann> ah, i remember
39 2012-10-30T00:41:54 <ThomasWaldmann> def logout(): flash(_("You are now logged out."), "info") flaskg.user.logout_session() return redirect(url_for('.show_root'))
40 2012-10-30T00:42:01 <ThomasWaldmann> that is the code from logout view
41 2012-10-30T00:42:02 *** sunu1
42 2012-10-30T00:42:31 <ThomasWaldmann> is there an issue with your ssh/screen session?
43 2012-10-30T00:43:33 <sunu> oh I just lost my connection for aa bit
44 2012-10-30T00:43:42 <sunu> so went onto the phone
45 2012-10-30T00:44:11 <ThomasWaldmann> ah
46 2012-10-30T00:44:20 <sunu> ThomasWaldmann: but it's better this way, no ? the way it is currently working ?
47 2012-10-30T00:44:53 <ThomasWaldmann> no, it doesn't really log out
48 2012-10-30T00:45:15 <ThomasWaldmann> it just kills the session, but the gae login is still somehow there, just not used
49 2012-10-30T00:45:17 <sunu> then should I have the logout method back ?
50 2012-10-30T00:45:37 <ThomasWaldmann> yes, guess you need to copy those 2 calls as you already had it
51 2012-10-30T00:46:26 <sunu> ok. btw have a look at CR's logout. It's the same as we have now.
52 2012-10-30T00:48:32 <sunu> and bringing back the old logout code will fix the current login issue too :)
53 2012-10-30T00:53:22 *** sunu1
54 2012-10-30T01:03:55 <sunu> ThomasWaldmann: Check now. http://moin-test.appspot.com/Home
55 2012-10-30T01:07:21 <ThomasWaldmann> hmm, is it intended that the logout is GLOBAL to all google apps?
56 2012-10-30T01:09:04 <sunu> no. It's not.
57 2012-10-30T01:10:26 <sunu> i mean loging out of moin won't log you out of your google account.
58 2012-10-30T01:10:46 <sunu> or any other app. it'll only log you out of moin.
59 2012-10-30T01:10:52 <ThomasWaldmann> it did
60 2012-10-30T01:12:28 <sunu> according to gae doc it shouldn't.
61 2012-10-30T01:12:35 <sunu> let me check.
62 2012-10-30T01:15:41 <sunu> ThomasWaldmann: so what should we do ?
63 2012-10-30T01:16:02 <ThomasWaldmann> logging out from moin also logs out gmail
64 2012-10-30T01:16:13 <ThomasWaldmann> logging out from gmail does not log out moin
65 2012-10-30T01:19:45 <sunu> yes. so should we rollback ?
66 2012-10-30T01:53:07 <sunu> ThomasWaldmann: there's nothing about it in the docs. So I guess it's the default behavior of create_logout_url
67 2012-10-30T02:04:14 <ThomasWaldmann> put a comment into the src, can be solved later
68 2012-10-30T02:04:16 <ThomasWaldmann> gn
69 2012-10-30T02:06:19 <sunu> good night ThomasWaldmann :) (though it's morning here now :D)
70 2012-10-30T03:58:38 *** dwcramer
71 2012-10-30T04:56:00 *** dwcramer
72 2012-10-30T05:30:00 *** dwcramer
73 2012-10-30T06:42:56 <dreimark> moin
74 2012-10-30T06:44:28 * dreimark on halloween you can get a crossover license for free
75 2012-10-30T10:03:35 *** greg_f
76 2012-10-30T11:32:09 <gwork> dreimark : huh? do you have any more info on that?
77 2012-10-30T12:27:29 *** jek
78 2012-10-30T12:28:27 *** jek
79 2012-10-30T13:53:39 *** dave_largo
80 2012-10-30T14:11:09 <sunu> moin
81 2012-10-30T14:39:18 <sunu> ThomasWaldmann: updated the CR with comment about the global logout http://codereview.appspot.com/6692048
82 2012-10-30T15:03:49 <ThomasWaldmann> replied
83 2012-10-30T15:12:58 *** dwcramer
84 2012-10-30T15:19:59 <sunu> ThomasWaldmann: couldn't understand which comment you are pointing to ?
85 2012-10-30T15:22:14 <sunu> no, got it now.
86 2012-10-30T15:27:04 <sunu> ThomasWaldmann: Done.
87 2012-10-30T15:42:36 <sunu> ThomasWaldmann: I think we could SIMULATE a workaround for the global logout issue using request.values. Let me upload the changes.
88 2012-10-30T15:57:28 *** dave_largo
89 2012-10-30T16:04:28 *** RogerHaase
90 2012-10-30T16:09:20 <sunu> no, it doesn't work as I expected.
91 2012-10-30T16:25:55 *** RogerHaase
92 2012-10-30T17:57:08 *** xorAxAx
93 2012-10-30T17:59:58 *** xorAxAx
94 2012-10-30T17:59:58 *** xorAxAx
95 2012-10-30T18:46:46 *** greg_f
96 2012-10-30T19:50:41 *** spy
97 2012-10-30T20:12:57 *** RogerHaase
98 2012-10-30T20:52:55 *** MattMaker
99 2012-10-30T20:53:13 *** MattMaker
100 2012-10-30T21:00:36 *** MattMaker
101 2012-10-30T21:08:20 *** MattMaker
102 2012-10-30T21:55:44 *** xjjk
103 2012-10-30T21:59:11 *** spy
104 2012-10-30T22:13:10 <RogerHaase> looks like pytest 2.3.2 breaks tests on windows/7: http://rn0.ru/show/QdMwuCeDlHezWwFFrOTQ/
105 2012-10-30T22:23:25 <ThomasWaldmann> hi roger
106 2012-10-30T22:27:46 <ThomasWaldmann> RogerHaase: from the error msg, it somehow looks like it is accessing one of flask's threadlocals (like request, session, ...), but I don't see why or which
107 2012-10-30T22:32:18 <RogerHaase> pytest works after I did "pip install -I pytest==2.2.4"
108 2012-10-30T22:32:50 <ThomasWaldmann> so, 2.3.x breaks it?
109 2012-10-30T22:33:01 <RogerHaase> yes
110 2012-10-30T22:33:33 <RogerHaase> or, at least 2.3.2
111 2012-10-30T22:33:46 <ThomasWaldmann> are you just running py.test from the toplevel dir or what exactly is the commandline?
112 2012-10-30T22:34:10 <ThomasWaldmann> it would be interesting to know whether 2.3.0 or 2.3.1 also breaks it
113 2012-10-30T22:34:33 <ThomasWaldmann> btw, we can also discuss on #pylib
114 2012-10-30T22:34:44 <RogerHaase> from top level, I have small .bat file:
115 2012-10-30T22:34:51 <RogerHaase> call env\Scripts\activate.bat
116 2012-10-30T22:34:51 <RogerHaase> py.test.exe > pytest.txt
117 2012-10-30T22:34:51 <RogerHaase> call env\Scripts\deactivate.bat
118 2012-10-30T22:37:18 <RogerHaase> will try 2.3.0 next
119 2012-10-30T22:41:05 <RogerHaase> pytest 2.3.0 has same error as 2.3.2: http://rn0.ru/show/oVByQ1L381lieIXzHiq3/
120 2012-10-30T22:44:30 <ThomasWaldmann> RogerHaase: i'll update the pypi mirror and try to reproduce on linux
121 2012-10-30T22:48:50 <ThomasWaldmann> RogerHaase: also happens on linux
122 2012-10-30T22:49:28 <RogerHaase> hooray, not a windows problem :-D -- one in a row
123 2012-10-30T22:50:22 *** dwcramer
124 2012-10-30T22:50:55 * ThomasWaldmann reads the changelog / docs
125 2012-10-30T23:17:35 <RogerHaase> gtg
126 2012-10-30T23:17:39 *** RogerHaase
127 2012-10-30T23:21:17 <dreimark> moin
128 2012-10-30T23:51:22 <moinBot> http://hg.moinmo.in/moin/2.0/rev/7a07ddddb256 2012-10-30 Thomas Waldmann <tw AT waldmann-edv DOT de> remove unused import that make trouble with py.test 2.3.x
129 2012-10-30T23:51:23 <moinBot> http://hg.moinmo.in/moin/2.0/rev/b02d552029c8 2012-10-30 Thomas Waldmann <tw AT waldmann-edv DOT de> require pytest < 2.3, because pytest 2.3.x doesn't work with our test code, likely due to the fixture changes done after 2.2