945245 - Fix misc char16_t/wchar_t mismatches.

Status: RESOLVED FIXED

(Core :: General, defect)

Description

[Jacek Caban]

Attachment: patch.diff

This is a followup of bug 928351 containing simple tree-wide changes. Most of such changes have their own bugs, but if a component has just one or two lines of changes, this seems pointless.

Comment 1

[(no longer active)]

Comment on attachment 8341075 [details] [diff] [review]
patch.diff

Review of attachment 8341075 [details] [diff] [review]:
-----------------------------------------------------------------

::: embedding/components/printingui/src/win/nsPrintDialogUtil.cpp
@@ +687,5 @@
>    HGLOBAL hGlobalDevMode = nullptr;
>  
>    HANDLE hPrinter = nullptr;
>    // const cast kludge for silly Win32 api's
> +  LPWSTR printName = (LPWSTR)aPrintName.get();

Please use reinterpret_cast here.

@@ +805,5 @@
>    if (printerName.IsEmpty()) {
>      GetDefaultPrinterNameFromGlobalPrinters(printerName);
>    } else {
>      HANDLE hPrinter = nullptr;
> +    if(!::OpenPrinterW((wchar_t*)printerName.get(), &hPrinter, nullptr)) {

Ditto.  You may also need a separate const_cast too.

Comment 2

[Jacek Caban]

(In reply to :Ehsan Akhgari (needinfo? me!) from comment #1)
> Comment on attachment 8341075 [details] [diff] [review]
> patch.diff
> 
> Review of attachment 8341075 [details] [diff] [review]:
> -----------------------------------------------------------------
> 
> ::: embedding/components/printingui/src/win/nsPrintDialogUtil.cpp
> @@ +687,5 @@
> >    HGLOBAL hGlobalDevMode = nullptr;
> >  
> >    HANDLE hPrinter = nullptr;
> >    // const cast kludge for silly Win32 api's
> > +  LPWSTR printName = (LPWSTR)aPrintName.get();
> 
> Please use reinterpret_cast here.

The problem here is that reinterpret_cast requires pointer argument and char16ptr_t is not a real pointer on GCC. In this case what we really want is const_cast, but that requires argument to be exactly the expected type, so this would be something like:
const_cast<wchar_t*>(static_cast<const wchar_t*>(aPrintName.get()))
That's why I decided that C-style cast looks better in this case, but I will change it if you like.

Comment 3

[(no longer active)]

(In reply to comment #2)
> (In reply to :Ehsan Akhgari (needinfo? me!) from comment #1)
> > Comment on attachment 8341075 [details] [diff] [review]
> > patch.diff
> > 
> > Review of attachment 8341075 [details] [diff] [review]:
> > -----------------------------------------------------------------
> > 
> > ::: embedding/components/printingui/src/win/nsPrintDialogUtil.cpp
> > @@ +687,5 @@
> > >    HGLOBAL hGlobalDevMode = nullptr;
> > >  
> > >    HANDLE hPrinter = nullptr;
> > >    // const cast kludge for silly Win32 api's
> > > +  LPWSTR printName = (LPWSTR)aPrintName.get();
> > 
> > Please use reinterpret_cast here.
> 
> The problem here is that reinterpret_cast requires pointer argument and
> char16ptr_t is not a real pointer on GCC. In this case what we really want is
> const_cast, but that requires argument to be exactly the expected type, so this
> would be something like:
> const_cast<wchar_t*>(static_cast<const wchar_t*>(aPrintName.get()))
> That's why I decided that C-style cast looks better in this case, but I will
> change it if you like.

I see.  The explicit casts are probably a bit clearer, so it would be great if you could switch to that please.

Comment 4

[Jacek Caban]

OK, I changed that. Thanks for review.

https://hg.mozilla.org/integration/mozilla-inbound/rev/32fcd54fd18b

Comment 5

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/32fcd54fd18b