78664 - [console] Windows widgets must not print to console in opt builds

Status: VERIFIED FIXED

(Core :: XUL, defect, P4)

Description

[cls]

<formletter>It has been decreed (or requested at any rate) that our release
(read non-debug) builds must not print anything to the console when the app is
running.  See bug 76720 for details.  I have done a preliminary tree scouring
and created mini-patches for each module that has bare printfs.   These patches
are not all inclusive as I didn't even think about xul/js output until post
scour so module owners & peers will still need to scour their modules themselves
as well as make sure the preliminary patches do not break anything.</formletter>

Comment 1

[cls]

Attachment: prelim patch

Comment 2

[John Morrison]

There is no console in optimized win32 builds. (Although, there is an 
undocumented '-console' switch).

Naked 'printf' in C++ is (for the most part) wrong, but a lot of bugs are 
brought to an easy resolution simply by bug reporters reporting the error
messages that appear in the console. I think this is a mistake, if the 
intent to disable any and all capability to get console output in optimized
builds.

Comment 3

[cls]

Sorry about the additional spammage but I should clear up a couple of things
before everyone starts replying.

1) I'm just the messenger.  Discussions outside of the specific module/patches
should be discussed in the parent bug ( bug 76720).
2) I have no intention of checking in the patches as is; that's why the bugs are
assigned to someone else ;).
3) The patches are the result of a far & wide-reaching grep across the entire
tree.  They may affect some cases that are not even used and they are far from
optimal.
4) Some platforms/ports will not need the printfs shutoff as they use other
mechanisms to stop the printfs.  That's fine.  Note it in the bug and close it
as invalid(?).  Depending upon the platform/port some people may still be
interested in removing the overhead from the printfs.

Comment 4

[Peter Trudelle]

->dr/0.9.1

Comment 5

[Peter Trudelle]

->0.9.2

Comment 6

[chris hofmann]

lets keep dribbling these [console] bugs into the tree as quick as we
can, but they shouldn't hold up or block 0.9.1 so moving the target milestone
to 0.9.2.

Comment 7

[Dan Rosen]

Attachment: [patch] wraps all uncommented *and* commented printfs in #ifdef DEBUG blocks, cleans up leaky nsString printfs.

Comment 8

[Dan Rosen]

Attachment: [patch] fixes small error, consistent with tabbing in nsWindow.cpp

Comment 9

[jag (Peter Annema)]

r=jag

Comment 10

[David Hyatt]

sr=hyatt

Comment 11

[Dan Rosen]

cc'ing Brendan per trudelle's recommendation. [e-mail] hasn't
reponded since I sent a request for a= last Thursday (6/14). Looking for a quick
 a= before we close for 0.9.2.

Comment 12

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

a=dbaron for trunk checkin (on behalf of (some of) drivers)

Comment 13

[Dan Rosen]

fix checked in.

Comment 14

[John Morrison]

couple of straggling messages from outside windows widgets, but that's a 
separate matter. Ka-chunk!.