1236015 - File menu>Print>Microsoft Print to PDF always results in ERROR

Status: RESOLVED FIXED

(Core :: Printing: Output, defect)

Description

[wvde]

User Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; Trident/7.0; Touch; .NET4.0C; .NET4.0E; .NET CLR 2.0.50727; .NET CLR 3.0.30729; .NET CLR 3.5.30729; Tablet PC 2.0; MASEJS; rv:11.0) like Gecko

Steps to reproduce:

File menu>Print>Microsoft Print to PDF


Actual results:

Error message


Expected results:

Make the PDF file

Comment 1

[YF (Yang)]

WFM on Fx44b4 Win10.

Comment 2

[ced]

Does not work for me with e10s enabled on 01/06/2016 nightly ( Mozilla/5.0 (Windows NT 10.0; WOW64; rv:46.0) Gecko/20100101 Firefox/46.0).

Comment 3

[ced]

Probably this is a known issue. Workaround suggested by Alice in Bug 1189846 is to set security.sandbox.content.level = 0 in about:config

Comment 4

[(inactive) Jim Chen [:jchen] [:darchons]]

Setting 'print.print_via_parent' to true in about:config also works for me (and it's probably safer than setting the sandbox pref).

Comment 5

[ced]

print.print_via_parent is already true (default) but still there's an error when print to PDF is attempted with e10s is enabled. Only security.sandbox.content.level = 0 fixes it.

Comment 6

[Bob Owen (:bobowen)]

Hi both.

Assuming that you were both trying this in Nightly, can you both retest with the latest Nightly.

Thanks.

Comment 7

[ced]

Problem still seen on Nightly (BuildID:20160202030232) with e10s enabled, print.print_via_parent = true, security.sandbox.content.level=2

Comment 8

[ced]

Looks like something is broken. I am always getting an error irrespective of the settings.

Comment 9

[Bob Owen (:bobowen)]

(In reply to ced from comment #8)
> Looks like something is broken. I am always getting an error irrespective of
> the settings.

Thanks for testing this, can you print with Firefox Release or a different application?

Comment 10

[ced]

(In reply to Bob Owen (:bobowen) from comment #9)
> (In reply to ced from comment #8)
> > Looks like something is broken. I am always getting an error irrespective of
> > the settings.
> 
> Thanks for testing this, can you print with Firefox Release or a different
> application?

Update: The issue was seen on Nightly (BuildID:20160202030232). On further testing, I found that the issue has some other dependency. I updated the 'Tab Mix Plus' add-on to the latest Dev version instead of the latest release version and the issue was no longer seen (though a crash with signature https://crash-stats.mozilla.com/report/index/67527034-9eee-429f-8e83-7e9932160202 was seen after saving to PDF). This is puzzling, however, since I observed the issue earlier even in safe mode. Let me see if I can restore the old state and see if the issue still exists.

Comment 11

[Bob Owen (:bobowen)]

That crash looks like a different bug, I'll file that separately thanks.

Comment 12

[ced]

Final Status: 

Could not reproduce issue reported earlier in comment 8.

Problem not seen on latest Nightly (BuildID:20160202030232, UserAgent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:47.0) Gecko/20100101 Firefox/47.0) with e10s enabled, and default settings (i.e. print.print_via_parent = true, security.sandbox.content.level=2)

Comment 13

[Bob Owen (:bobowen)]

(In reply to ced from comment #12)
> Final Status: 
> 
> Could not reproduce issue reported earlier in comment 8.
> 
> Problem not seen on latest Nightly (BuildID:20160202030232, UserAgent:
> Mozilla/5.0 (Windows NT 10.0; WOW64; rv:47.0) Gecko/20100101 Firefox/47.0)
> with e10s enabled, and default settings (i.e. print.print_via_parent = true,
> security.sandbox.content.level=2)

Thanks again.

I'll see if the original poster can reproduce.

Also, I've seen a possible issue while just testing, if I print first (after reboot) with the sandbox enabled on Windows 10.
I'll do some more testing to see if it's a real problem.

Comment 14

[ced]

Bob, I think you are right.

Win10+security.sandbox.content.level=2+computer reboot->issue in comment 8 seen
Win10+security.sandbox.content.level=0+computer reboot->issue in comment 8 not seen

Comment 15

[Bob Owen (:bobowen)]

(In reply to ced from comment #14)
> Bob, I think you are right.
> 
> Win10+security.sandbox.content.level=2+computer reboot->issue in comment 8
> seen
> Win10+security.sandbox.content.level=0+computer reboot->issue in comment 8
> not seen

Thanks for confirming this.

It looks like there is a problem if the first interaction with the print device is from the sandboxed process.
I'll have to debug to confirm.
I'm going to assume that this is what caused the original problem.

At some point I want to stop the child process from accessing the device completely, maybe I'll have to do that sooner rather than later.

Comment 16

[Bob Owen (:bobowen)]

I worked out what is going on here ... on Windows 10 (with 32-bit Firefox) instead of creating a low integrity 64-bit spool helper process the Printer Spooler service creates a medium integrity one, but it still picks up the access token settings from the sandboxed job. This helper then gets used by the main process (and other apps as well) and fails because of the USER_INTERACTIVE settings.

I think I just need to get rid of print device access from the content process

Comment 17

[Bob Owen (:bobowen)]

For this bug, I want to start initiating printing through nsFrameLoader::Print, so that the interaction with the print driver for populating print settings can all happen on the parent.

This will probably mean I can get rid of one of the nested event loops in the child and possibly the sync calls to save the settings as well.

I've got some other things to do before I get to this, but I want to make a change to nsFrameLoader::Print to add the outer window ID I'll need for printing frames.
This function is used by at least one add-on, so I want to make that change now in the same version it was added, so I'll file a blocking bug for this.

Comment 18

[Bob Owen (:bobowen)]

While it doesn't include all of the print set-up changes I described in comment 17, this should have been fixed by bug 1324064.