Just some more cleanup. There is no reason all user_save() calls have to endure some distributed auth code. I checked OpenID and it is unaffected by this patch.
| Comment | File | Size | Author |
|---|---|---|---|
| #11 | mw.patch | 3.65 KB | moshe weitzman |
| mw_88.patch | 1.76 KB | moshe weitzman |
Comments
Comment #1
moshe weitzman commentedin order to test this, just do a registration and a login via distributed auth (i.e. enable drupal.module and login with @drupal.org) and make sure all is well.
Comment #2
pwolanin commentedsubscribing
Comment #3
gábor hojtsySo where does this auth information saved instead?
Comment #4
moshe weitzman commenteddrupa.module calls user_set_authmaps() after this patch. openid calls into the authmap table directly.
Comment #5
gábor hojtsyAgain, this patch does not touch drupal.module, so drupal.module might register with user_external_login_register() then?!
Comment #6
moshe weitzman commented@Gabor - you are right. The relevant change is in user_external_login_register().
My statement in the other issue is incorrect. drupal.module is unchanged by this patch and can be booted at any time.
Comment #7
catchThis still applies cleanly. drupal.module has gone but I assume this is still a valid issue?
Comment #8
gábor hojtsyYes, in principle this looks good, but another set of eyes would never hurt.
Comment #9
moshe weitzman commentedsigh.
Comment #10
moshe weitzman commentedrefreshed. i did need to change two lines in openid module so that it creates its own authmap entry instead of relying on user_save().
Comment #11
moshe weitzman commentednow with patch
Comment #12
dries commentedI think this patch might need a re-roll because of the schema API patch that got committed?
Comment #13
moshe weitzman commentedI rerolled after that one landed. Should be good to go.
Comment #14
dries commentedLooks good to me. Makes for cleaner code. Committed this patch to CVS HEAD.
Comment #15
Anonymous (not verified) commentedAutomatically closed -- issue fixed for two weeks with no activity.