I've been looking at this the last few days and can't find a clean fix on top of the existing code. So I'm inclined to go back to the original code as it was after Bug 163964 landed. Here's a summary of the relevant changes in the order they landed: Bug 163964 caused initial discoverallboxes and select URL's to occur in parallel in two connection rather than serially in one connection to speed up startup for users with lots of folders. This change didn't intentionally queue any URLs to run later. Bug 1768121 was caused by the user putting in the wrong password and getting their account locked out. The fix for this was to do the discoverallboxes and queue the select until authentication completes, then run the select. This is where the queuing of URLs began and is the root of the problem. I think if this issue had been a "wontfix" the current problems we are seeing would not be present. Bug 1793599 added more queuing and other changes to handle the case where the password is not stored and has to be prompted for. Also handled the case where the first URL has no associated window so you can't prompt for password. Bug 1764770 then required some more fixes for the queuing due to the need to show an override prompt for TLS errors. Bug 1798161 allows imap PREAUTH not require a password. I think this is OK as-is, but will look closer. I'll look into getting rid of all these changes except for Bug 163964. I think the only downside is that some users may more easily get locked out of their accounts if they use the wrong password as occurred at Bug 1768121.
Bug 1854586 Comment 9 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
I've been looking at this the last few days and can't find a clean fix on top of the existing code. So I'm inclined to go back to the original code as it was after Bug 163964 landed. Here's a summary of the relevant changes in the order they landed: Bug 163964 caused initial discoverallboxes and select URL's to occur in parallel in two connection rather than serially in one connection to speed up startup for users with lots of folders. This change didn't intentionally queue any URLs to run later. Bug 1768121 was caused by the user putting in the wrong password and getting their account locked out. The fix for this was to do the discoverallboxes and queue the select until authentication completes, then run the select. This is where the queuing of URLs began and is the root of the problem. I think if this issue had been a "wontfix" the current problems we are seeing would not be present. Bug 1764770 then required some more fixes for the queuing due to the need to show an override prompt for TLS errors. Bug 1793599 added more queuing and other changes to handle the case where the password is not stored and has to be prompted for. Also handled the case where the first URL has no associated window so you can't prompt for password. Bug 1798161 allows imap PREAUTH not require a password. I think this is OK as-is, but will look closer. I'll look into getting rid of all these changes except for Bug 163964. I think the only downside is that some users may more easily get locked out of their accounts if they use the wrong password as occurred at Bug 1768121. Edited: Bug 1764770 landed before Bug 1793599
I've been looking at this the last few days and can't find a clean fix on top of the existing code. So I'm inclined to go back to the original code as it was after Bug 163964 landed. Here's a summary of the relevant changes in the order they landed: Bug 163964 caused initial discoverallboxes and select URL's to occur in parallel in two connection rather than serially in one connection to speed up startup for users with lots of folders. This change didn't intentionally queue any URLs to run later. Bug 1768121 was found by the user putting in the wrong password and getting their account locked out. The fix for this was to do the discoverallboxes and queue the select until authentication completes, then run the select. This is where the queuing of URLs began and is the root of the problem. I think if this issue had been a "wontfix" the current problems we are seeing would not be present. Bug 1764770 then required some more fixes for the queuing due to the need to show an override prompt for TLS errors. Bug 1793599 added more queuing and other changes to handle the case where the password is not stored and has to be prompted for. Also handled the case where the first URL has no associated window so you can't prompt for password. Bug 1798161 allows imap PREAUTH not require a password. I think this is OK as-is, but will look closer. I'll look into getting rid of all these changes except for Bug 163964. I think the only downside is that some users may more easily get locked out of their accounts if they use the wrong password as occurred at Bug 1768121. Edited: Bug 1764770 landed before Bug 1793599
I've been looking at this the last few days and can't find a clean fix on top of the existing code. So I'm inclined to go back to the original code as it was after Bug 163964 landed. Here's a summary of the relevant changes in the order they landed: Bug 163964 caused initial discoverallboxes and select URL's to occur in parallel in two connection rather than serially in one connection to speed up startup for users with lots of folders. This change didn't intentionally queue any URLs to run later. The request to improve the startup sequence was initiated in Bug 1660672 (eventually marked as duplicate of Bug 163964). Bug 1768121 was found by the user putting in the wrong password and getting their account locked out. The fix for this was to do the discoverallboxes and queue the select until authentication completes, then run the select. This is where the queuing of URLs began and is the root of the problem. I think if this issue had been a "wontfix" the current problems we are seeing would not be present. Bug 1764770 then required some more fixes for the queuing due to the need to show an override prompt for TLS errors. Bug 1793599 added more queuing and other changes to handle the case where the password is not stored and has to be prompted for. Also handled the case where the first URL has no associated window so you can't prompt for password. Bug 1798161 allows imap PREAUTH to not require a password. I think this is OK as-is, but will look closer. I'll look into getting rid of all these changes except for Bug 163964. I think the only downside is that some users may more easily get locked out of their accounts if they use the wrong password as occurred at Bug 1768121. Edited: Bug 1764770 landed before Bug 1793599