Closed Bug 199217 Opened 22 years ago Closed 21 years ago

confusion of sort by thread when switching from desc. sorted flat view

Categories

(SeaMonkey :: MailNews: Message Display, defect)

defect
Not set
minor

Tracking

(Not tracked)

VERIFIED FIXED

People

(Reporter: ch.ey, Assigned: sspitzer)

References

Details

Attachments

(1 file)

User-Agent: Mozilla/5.0 (X11; U; Linux i686; de-DE; rv:1.4a) Gecko/20030323 Build Identifier: Mozilla/5.0 (X11; U; Linux i686; de-DE; rv:1.4a) Gecko/20030323 A little hard to describe. Do a non-thread sort (like subject, sender, date a.s.o) in desending order. Click on thread-col heading. Mail list will be sorted by thread but m_sortOrder won't be set to the new sortOrder. It will be set to sortOrder not until reverse sorted by thread. This results in the asc-icon for thread-col heading will show up if sorted if the thread is sorted descending and vice versa. Normaly this malconduct can't be seen since the thread-col has the same icon for ascending and descending sorting order. But I noticed it while working on bug 195832. Reproducible: Always Steps to Reproduce:
Blocks: 195832
Attached patch proposed patchSplinter Review
Set m_sortOrder to new sortOrder after applying sort.
Your description is a little hard to follow, but I think I've figured it out. I think having Bug 199477 as a separate bug makes this even harder to follow; they seem to be very closely related. Here is my description of the problem I see going from Flat to Threaded sorting: When a new sort criterion is chosen, it should force to "Ascending" order. Flat-sort criteria do this; however, when Sort by|Thread is selected, the order used is the same as when Sort by Thread was *last* used. (This is Bug 199477.) If this new sort order is different from the sort order used for the flat criterion, then the UI displays the sort order in reverse: by Thread (Ascending) => by Date (Descending) => by Thread yields: messages sorted ascending by thread, but View|Sort by|Descending is checked by Thread (Descending) => by Sender (Ascending) => by Thread yields: messages sorted descending by thread, but View|Sort by|Ascending is checked Once the UI selection is reversed, it continues to be reversed: change the menu selection from Ascending to Descending, and the descending threads resort in ascending order. I can confirm this behavior in 1.4b: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4b) Gecko/20030408 Since my report now contains the essence of this bug and 199477, I suggest you mark that as a duplicate of this one, OK?
Status: UNCONFIRMED → NEW
Ever confirmed: true
Hm, both problems result in confustion about thread sorting, yes. But since there are two different points to fix, I'm not really sure if duping 199477 is the best. Point 1 is that the indicators show what they should (ascending when changing from one sort type to another) (this bug). Point 2 is to fix the behaviour of thread sorting itself (bug 199477).
*** Bug 199477 has been marked as a duplicate of this bug. ***
Christian, you may have noticed that the sorting behavior has changed a lot, due to the fix for bug 72493. I think that the problem described in this bug no longer occurs, altho it's a little hard to be sure because "Sort by |Thread" behaves somewhat differently than it used to; see also bug 219620.
I didn't inspect the new behaviour before. Now I did, and I'm still confused. > I think that the problem described in this bug no longer occurs I don't think so. The problem can't be visualized by the patch from bug 195832. This seems because of another change from patch for bug 72493. But m_sortOrder is still not set to sortOrder. To this the summary from bug 199477 is still true too. "Thread sorting isn't always ascending when coming from flat sort". Comming from flat sort to thread sort by clicking on the ThreadCol head sorts by thread in the order the previous flat sort had. So ascending sender sort -> ascending thread sort descending sender sort -> descending thread sort but ascending sender sort -> ascending subject (or more exact !thread) sort descending sender sort -> ascending subject (or more exact !thread) sort
Christian, I am seeing slightly different behavior now than described in comment 2. Following are some test results using Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.6a) Gecko/20031003 This version has the fix for bug 72493, plus followup fixes for bug 217031 and bug 218656. It is necessary now to track actions of View|SortBy|Thread (VST) vs. View|Messages|Threaded (VMT), as these two currently behave differently. (VST is equivalent to clicking on the Thread column header.) Also, it is helpful to display the Order Received column. Repeating the tests in comment 2: (1) VST (sort by Thread, Ascending) Menu state: Order Received, Ascending Order Received column: NO SORT ICON shown. => VMT (return to flat) Menu state: Order Received, Ascending Order Received column: Ascending sort arrow icon. => Sort by Date (Descending) => VST: Messages are threaded, sorted by Order Received, Descending. Sort menu state is Order Received, Descending. Order Received column: NO SORT ICON shown. (2) VST (by Thread, Ascending) => VMT (return to flat) => Sort by Date (Descending) => VMT: messages are threaded, sorted by Date, Descending. Sort menu state is Date, Descending. Date column has ASCENDING sort icon (arrow pointing down) (3) VST => View Sort Descending Menu state: Order Received, Descending Order Received column: Descending sort icon. => VMT (return to flat) Menu state: Order Received, Descending Order Received column: ASCENDING sort icon. => Sort by Sender (Ascending) => VST yields: Messages are threaded, sorted by Order Received, Descending. Sort menu state is Order Received, Descending. Order Received column: NO SORT ICON shown. (4) VST => View Sort Descending => VMT (return to flat) => Sort by Sender (Ascending) => VMT yields: messages are threaded, sorted by Sender, Ascending. Sort menu state is Sender, Ascending. Sender column: Ascending sort icon In these cases, the *menu* state always agrees with the actual sorting (which wasn't true before, a problem I described in comment 2). But the *sort icon* state can have various problems depending on the sequence of operations; and I think that's what you originally intended the bug for. Specifically: One symptom is that whenever View|Sort By|Thread is selected, the threads are sorted by Order Received, Ascending; but the Order Received column does not show an Ascending sort-arrow icon, until/unless the View|Sort By|Order Received, or Ascending, or Descending menu is selected, or until the Order Received column or Thread column is clicked. (I first discovered this while testing for bug 219787.) The other symptom is that if View|Messages|Threaded is selected, and the threaded sort is Descending, the actual sort order is not affected, but the column sort icon is always shown Ascending. This is true whether VMT is used to turn threading on or to turn it off. In addition, you are correct about the symptom from bug 199477: View|Sort By|Thread (still) does not (always) force the sort order to Ascending: if VST is selected from an unthreaded, Descending view, then the sort order is maintained as Descending. If VST is selected from a threaded, Descending view, the order is reverted to Ascending. Can you update your patch to address all three of these symptoms?
No Mike, I can't update my patch - I simply don't have the knowledge. The first "patch" was quite simple and I stumpled over it more by chance than by engineering.
Christian, the patch for bug 219787 is in place, for 1.6b nightlies; and it fixes the problems I described in comment 7 as well. Please comment as to whether this bug should be resolved or not. However, I believe that this fix complicates your goal of having different icons for ascending vs. descending in the thread column (bug 195832) -- the 'byThread' sort ordering is now never used, in favor of the 'byId' one. (Note that ID == Order Received)
Christian Eyrich, when you get a chance to test Mozilla 1.6, please comment here on what you wish to do with this bug.
I clicked around a lot now and couldn't (re-)produce any sort direction error. Though I didn't look at it in debugger or so, I think it has been fixed while working on some of the other sort-bugs.
Status: NEW → RESOLVED
Closed: 21 years ago
Resolution: --- → FIXED
v fix -- see bug 219787.
Status: RESOLVED → VERIFIED
Product: Browser → Seamonkey
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: