Closed
Bug 455182
Opened 16 years ago
Closed 16 years ago
Problems refreshing multiple Digest-authed CalDAV calendars
Categories
(Calendar :: Provider: CalDAV, defect)
Calendar
Provider: CalDAV
Tracking
(Not tracked)
RESOLVED
FIXED
0.9
People
(Reporter: browning, Assigned: browning)
Details
Attachments
(1 file)
6.55 KB,
patch
|
dbo
:
review+
|
Details | Diff | Splinter Review |
As noted at http://mxr.mozilla.org/mozilla1.8/source/calendar/providers/caldav/calDavCalendar.js#934 the way we have been queuing HTTP queries from refresh()es of calendars sharing an authrealm won't work with the new changelog code. This causes refresh() problems for anyone who uses a CalDAV server with Digest auth (which is the default Apple configuration) and multiple calendars with at least one of them cached. IMO this should be fixed for 0.9. This patch does fix it; I'm not particularly happy with the method, though, and interested in hearing of other approaches.
Flags: blocking-calendar0.9?
Attachment #338477 -
Flags: review?(daniel.boelzle)
Updated•16 years ago
|
Attachment #338477 -
Flags: review?(daniel.boelzle) → review+
Comment 1•16 years ago
|
||
Comment on attachment 338477 [details] [diff] [review] Do auth renegotiation in a synchronous HEAD query > refresh: function caldav_refresh() { > if (!this.mCheckedServerInfo) { > // If we haven't refreshed yet, then we should check the resource > // type first. This will call refresh() again afterwards > this.checkDavResourceType(null); > return; obsolete return >- } >- >- if (this.mAuthScheme != "Digest") { >- // Basic HTTP Auth will not have timed out, we can just refresh >- // Same for Cosmo ticket-based authentication >- this.safeRefresh(null); > } else { >- // Digest auth may have timed out, and we need to make sure that >- // several calendars in this realm do not attempt re-auth simultaneously >- if (this.firstInRealm()) { >- this.safeRefresh(null); >- } >+ this.safeRefresh(); > } > }, looks good, r=dbo, though hardly tested (only Bedework).
Comment 2•16 years ago
|
||
Checked in on HEAD, MOZILLA_1_8_BRANCH and SUNBIRD_0_9_BRANCH.
Updated•16 years ago
|
Flags: blocking-calendar0.9? → blocking-calendar0.9+
Updated•16 years ago
|
Status: NEW → ASSIGNED
Target Milestone: --- → 0.9
Comment 3•16 years ago
|
||
Are we keeping this bug open to find alternatives, or can it be closed? I think using HEAD is ok, but we don't really want to do this asynchronously. I know that we basically need to to make sure that other queries to the server are not sent, but maybe we could do this by queuing all operations instead and then doing the HEAD query asynchronously. We really want to identify the respective necko bug or file it if it doesn't exist. It may already have been mentioned, but if someone could either tell me about it or cc me on it, I would be grateful.
Comment 4•16 years ago
|
||
Setting to FIXED per IRC chat with Bruno.
Status: ASSIGNED → RESOLVED
Closed: 16 years ago
Resolution: --- → FIXED
You need to log in
before you can comment on or make changes to this bug.
Description
•