179056 - "Check for new message every" stops after hibernate

Status: RESOLVED FIXED

(MailNews Core :: Backend, defect)

Description

[Claudio Cicali]

When resuming Windows from hibernate (on a laptop, HP Omnibook xe4500) Mozilla
Mail does not check my IMAP account every 2 minutes as it does normally. I need
to manually check mail one time and then the feature began to work properly.
Using Mozilla 1.2b/en on a Windows 2000 SP3

Comment 1

[s.marshall]

This still happens on Thunderbird 0.3, Win2k SP4.

Comment 2

[vakos]

Hi
Exactly the same on
winxp sp1 + hotfixes
mozilla 1.6
this is really annoying problem :(

Comment 3

[Greger Cronquist]

I also have this problem on a win2k with Mozilla 1.7b using POP servers.

Comment 4

[Greger Cronquist]

I can confirm this with

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.7.2) Gecko/20040803

on a Win2K laptop (2 POP servers).

Comment 5

[Moritz Strübe]

bug 275727, bug 271183 and bug 179056 are dups. Can someone dup them and confirm
one?

Comment 6

[Mike Cowperthwaite]

*** Bug 271183 has been marked as a duplicate of this bug. ***

Comment 7

[parish]

This also affects Firefox and, I expect the browser component in Mozilla, in
that Live Bookmarks are no longer refreshed every X seconds - as defined by
browser.bookmarks.livemark_refresh_seconds - after hibernate/standby. Perhaps
the Component field of this bug should be set to something more general?

Comment 8

[Moritz Strübe]

Can someone please dup bug 245471.
Thanks

Comment 9

[Moritz Strübe]

Did some digging in lxr and filed two bugs:
bug 282013 is slightly related.
This bug should depend on bug 282014.

Comment 10

[:deanis74]

*** Bug 245471 has been marked as a duplicate of this bug. ***

Comment 11

[Adam Guthrie]

*** Bug 275727 has been marked as a duplicate of this bug. ***

Comment 12

[David Whyte]

This behaviour is not only caused by hibernation, but by putting the PC to sleep
too.

Behaviour is still present in version 1.5 Beta 1 (20050908).  It is very
annoying for RSS feeds!

Comment 13

[Mike Cowperthwaite]

(In reply to comment #12)
> This behaviour is not only caused by hibernation, but by putting the PC to
> sleep too.

I don't see that with any version of TB, 1.0 or 1.5 or 1.6, under Windows 2000.
Bug 179056 is about that situation.  This bug is for hibernate only.

Comment 14

[Jon Baumgartner]

Um, this is bug 179056. And I can confirm it happens with sleep as well on XP.

Comment 15

[Mike Cowperthwaite]

(In reply to comment #14)
> Um, this is bug 179056.

Duh, so it is.  It was bug 275727 that was originally opened for both hibernate 
and standby.

Comment 16

[Jon Baumgartner]

Bug 275727 was marked as a dupe of this one, so shouldn't sleep/standby issues be discussed here?

Comment 17

[R. H. Allen]

I can confirm that this bug is still present in Thunderbird 1.5 RC1. I have a few additional observations that I haven't seen discussed elsewhere -- perhaps somebody will find something useful in them:

* In my experience, hibernate *usually* produces this problem, but not always. There seems to be a correlation between how long the computer hibernates and whether or not the problem occurs. Generally, if I hibernate for only an hour or two, Thunderbird works fine; however, if I hibernate overnight, Thunderbird won't check mail on schedule anymore.

* After returning from hibernate, if I let Thunderbird sit long enough it will sometimes begin checking messages on schedule again. "Long enough" is occasionally only 20 minutes or so, but is usually several hours. I should note that I don't usually let Thunderbird sit long enough to "fix" itself -- if it doesn't work immediately I usually restart it pretty quickly -- so my experience with this is limited.

* I also use Sunbird 0.2 and I have found that after returning from hibernate, if Thunderbird won't check email, Sunbird's alarm function also won't work. The two go hand-in-hand. Perhaps the bug lies in a common piece of code?

Comment 18

[Moritz Strübe]

Well, it is a bit deeper. See my Comment #9; This Bug is dependent on Bug 282014. Unfortunately I don not have the right to set it dependent. Bug 282014 ist XPCOM - so _all_ Products have this Problem. You can eaven reproduce it with FF if you use some extension, which needs a timer.
As you noticed right it got better, though.

Comment 19

[parish]

(In reply to comment #18)
> Bug 282014
> ist XPCOM - so _all_ Products have this Problem. You can eaven reproduce it
> with FF if you use some extension, which needs a timer.

Yes, Live Bookmarks (RSS feeds) stop updating in FF (and TB too I would think, but I don't use TB for RSS) after hibernate.

Comment 20

[Masatoshi Kimura [:emk]]

Attachment: implementing wake/sleep notification

At present, wake/sleep notification is only implemented on Mac. They are not notified at all on other platforms.
This patch implements them on Windows.
A Japanese tester said this fixed the problem.

Comment 21

[Masatoshi Kimura [:emk]]

Attachment: widget: Win32 part

Comment 22

[Masatoshi Kimura [:emk]]

Attachment: XPCOM part

Requesting review to Simon who wrote and reviewed a patch for bug 197863.

Comment 23

[David :Bienvenu]

Comment on attachment 212472 [details] [diff] [review]
implementing wake/sleep notification

I think we'll want this for 1.8.1 (and possibly but less likely, 1.8.0.2, depending on schedule)

Comment 24

[Brian Ryner (not reading)]

Comment on attachment 212472 [details] [diff] [review]
implementing wake/sleep notification

Hm, is the message really dispatched to each window?  That's unfortunate.  I'm a little worried that we could have 0, or more than 1, invisible windows in some cases.  roc would know better than me, can you ask him for review?

Comment 25

[Robert O'Callahan (:roc) (email my personal email if necessary)]

I didn't know, but I checked the code and it seems pretty clear there's always exactly one hidden window.

Comment 26

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

checked-in.

Comment 27

[Masatoshi Kimura [:emk]]

*** Bug 282014 has been marked as a duplicate of this bug. ***

Comment 28

[Masatoshi Kimura [:emk]]

Attachment: branch patch

Bienvenu:
AFAIK, approval‑branch‑1.8.1 flag needs a requestee.
Asking Ere for approval because most of this patch belong to Win32: widget.

Comment 29

[Ere Maijala (slow)]

Comment on attachment 213042 [details] [diff] [review]
branch patch

Sorry, I can't grant approval as I'm not a driver.

Comment 30

[David :Bienvenu]

for the 1.8.1 branch, you just have to be a module owner

Comment 31

[Ere Maijala (slow)]

Comment on attachment 213042 [details] [diff] [review]
branch patch

Oh, good to know. Ok then..

Comment 32

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Comment on attachment 212472 [details] [diff] [review]
implementing wake/sleep notification

ere already approved this

Comment 33

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Kimura-san:

This patch cannot apply to 1.8 branch.

> patching file widget/src/windows/nsWindow.cpp
> Hunk #1 FAILED at 66.
> Hunk #2 succeeded at 4167 (offset 182 lines).
> Hunk #3 succeeded at 4297 (offset 182 lines).
> 1 out of 3 hunks FAILED -- saving rejects to file widget/src/windows/nsWindow.cpp.rej
> (Stripping trailing CRs from patch.)
> patching file widget/src/windows/nsWindow.h
> Hunk #1 FAILED at 262.
> 1 out of 1 hunk FAILED -- saving rejects to file widget/src/windows/nsWindow.h.rej
> (Stripping trailing CRs from patch.)
> patching file xpcom/threads/TimerThread.cpp

Comment 34

[Masatoshi Kimura [:emk]]

Attachment: branch patch

Oops, that was the same as trunk patch.
Here is a correct one. I've confirmed this is not rejected.
Carrying over branch approval (no actual code change).

Comment 35

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

checked-in to 1.8 branch.

Comment 36

[[e-mail]]

My MinGW headers don't have these PBT_ constants, do I need to update?

Comment 37

[Masatoshi Kimura [:emk]]

Attachment: fixing mingw bustage (trunk)

Comment 38

[Masatoshi Kimura [:emk]]

Attachment: fixing mingw bustage (1.8 branch)

Comment 39

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Comment on attachment 213178 [details] [diff] [review]
fixing mingw bustage (trunk)

+#else

unnecessary.

Comment 40

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Comment on attachment 213179 [details] [diff] [review]
fixing mingw bustage (1.8 branch)

but remove the #else.

Comment 41

[Masatoshi Kimura [:emk]]

Attachment: fixing mingw bustage (trunk, updated)

Comment 42

[Masatoshi Kimura [:emk]]

Attachment: fixing mingw bustage (1.8 branch, updated)

Comment 43

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

checked-in to trunk and 1.8 branch.(fixed mingw bustage)

Comment 44

[Cris]

File :

ftp://ftp.mozilla.org/.1/mozilla/pub/mozilla.org/firefox/nightly/latest-mozilla1.8.0/firefox-1.5.0.2.en-US.win32.zip  
6721 KB  	16/03/2006  	17.26.00

has this patch?

Where can I find first realeses with this patch?

Comment 45

[David :Bienvenu]

ftp://ftp.mozilla.org/pub/mozilla.org/thunderbird/nightly/latest-mozilla1.8

however, that's a pre-alpha build. I think it's reasonably stable but hasn't had much testing.

Comment 46

[Cris]

Last firefox 1.5.0.2 still has the problems to take forever to restore from hibernation and reach 100% CPU.

I found interesting the description of bug 325384. It is about SeaMonkey but it
is probably related with the long time to recover from standby.

Comment 47

[Masatoshi Kimura [:emk]]

Attachment: 1.8.0 branch patch

Plaease consider landing this on 1.8.0 branch.
This will fix some of (but not all) CPU 100% freeze after resore in bug 265172.
And this is baked enough long on trunk and 1.8 branch.

Comment 48

[Daniel Veditz [:dveditz]]

Comment on attachment 219440 [details] [diff] [review]
1.8.0 branch patch

approved for 1.8.0 branch, a=dveditz for drivers.

Please land soon!

Comment 49

[Cris]

Sorry, I am not familiar with Mozilla versions.
What does "approved for 1.8.0 branch, a=dveditz for drivers." mean?
Will I see this patch in the next 1.5.0.3 version?
When will it be released?

Comment 50

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

checked-in to 1.8.0 branch. thanks.