Printing: Refactor PrintSettings::Init into a PrintSettingsFactory class

Printing: Refactor PrintSettings::Init into a PrintSettingsFactory class
per-platform.

BUG=none
TEST=none

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=63907

[stuartmorgan, 2010-10-22 17:12:56]

http://codereview.chromium.org/4040001/diff/13001/14003
File printing/print_settings.h (right):

http://codereview.chromium.org/4040001/diff/13001/14003#newcode27
printing/print_settings.h:27: const PageRanges& new_ranges, bool
print_selection_only);
Why is this still here, given that you removed the implementation?

http://codereview.chromium.org/4040001/diff/13001/14007
File printing/print_settings_factory_mac.h (right):

http://codereview.chromium.org/4040001/diff/13001/14007#newcode8
printing/print_settings_factory_mac.h:8: #include <string>
Is this necessary?

http://codereview.chromium.org/4040001/diff/13001/14012
File printing/printing_context.h (right):

http://codereview.chromium.org/4040001/diff/13001/14012#newcode95
printing/printing_context.h:95: return *(settings_.get());
This is now returning a reference to heap-allocated memory, which will become
dangling if any of the settings-changing methods are called again. You should
explicitly document that, and make sure that no current callers violate that
constraint.

(Alternately, you could change the factory to fill an existing object and not
change the type of settings_)

[James Hawkins, 2010-10-25 21:00:33]

http://codereview.chromium.org/4040001/diff/13001/14003
File printing/print_settings.h (right):

http://codereview.chromium.org/4040001/diff/13001/14003#newcode27
printing/print_settings.h:27: const PageRanges& new_ranges, bool
print_selection_only);
On 2010/10/22 17:12:56, stuartmorgan wrote:
> Why is this still here, given that you removed the implementation?

Good catch. I forgot to remove it. Done.

http://codereview.chromium.org/4040001/diff/13001/14007
File printing/print_settings_factory_mac.h (right):

http://codereview.chromium.org/4040001/diff/13001/14007#newcode8
printing/print_settings_factory_mac.h:8: #include <string>
On 2010/10/22 17:12:56, stuartmorgan wrote:
> Is this necessary?

Neg, removed.

http://codereview.chromium.org/4040001/diff/13001/14012
File printing/printing_context.h (right):

http://codereview.chromium.org/4040001/diff/13001/14012#newcode95
printing/printing_context.h:95: return *(settings_.get());
On 2010/10/22 17:12:56, stuartmorgan wrote:
> This is now returning a reference to heap-allocated memory, which will become
> dangling if any of the settings-changing methods are called again. You should
> explicitly document that, and make sure that no current callers violate that
> constraint.
> 
> (Alternately, you could change the factory to fill an existing object and not
> change the type of settings_)

Changed PrintSettingsFactory to PrintSettingsInitializer which takes an existing
PrintSettings object and initializes it.

[stuartmorgan, 2010-10-25 21:55:26]

LGTM with nit.

http://codereview.chromium.org/4040001/diff/28001/29005
File printing/print_settings_initializer_gtk.h (right):

http://codereview.chromium.org/4040001/diff/28001/29005#newcode18
printing/print_settings_initializer_gtk.h:18: // Creates a PrintSettings object
from the provided Gtk printer objects.
s/Creates/Initializes in all the headers.