45797 - Remove deprecated NS_IMPL_QUERY_INTERFACE macro

Status: RESOLVED FIXED

(Core :: XPCOM, defect, P2)

Description

[Chris McAfee]

http://lxr.mozilla.org/mozilla/source/xpcom/base/nsISupportsUtils.h#495
says that NS_IMPL_QUERY_INTERFACE has been depreciated,
we need to yank this.  Other macros in the file are built
with this macro, so those need to change as well.

This bug was filed as a result of looking at bug 39270, removing
one of the usages of NS_IMPL_QUERY_INTERFACE.

Comment 1

[Alec Flett]

Attachment: fix for XPCOM

Comment 2

[Alec Flett]

Attachment: fix for necko

Comment 3

[Doron Rosenberg (IBM)]

settting default qa

Comment 4

[Alec Flett]

Attachment: fixes for intl

Comment 5

[Alec Flett]

Attachment: fix for gfx and widget (Windows)

Comment 6

[Alec Flett]

I've started attaching patches to get rid of people who use NS_IMPL_SUPPORTS()
and friends.
Anyone care to start reviewing so I can check these in? Adding some random
XPCOM-interested people in hopes they will review. :)

These patches can be checked in at any time.. once I've checked them ALL in,
then I can actually remove the old macros.

Comment 7

[Alec Flett]

Attachment: fix for layout

Comment 8

[Alec Flett]

Attachment: fix for embed/docshell/webshell

Comment 9

[Alec Flett]

Attachment: fix for rdf

Comment 10

[Daniel Veditz [:dveditz]]

This is the first I've heard that NS_IMPL_QUERY_INTERFACE is deprecated. This 
sort of thing should be announced more broadly since most folks are going to 
keep on using what they've learned rather than spot obscure comment changes in 
a header file somewhere.

r=dveditz for the patches for xpcom,necko,intl,gfx/widget and layout

In nsFontMetricsWin could you replace the ::QueryInterface with 
NS_IMPL_QUERY_INTERFACE1() ? whatever debug leak stuff they're doing seems to 
be only in the AddRef and Release.

Comment 11

[Alec Flett]

Attachment: fix for xpfe

Comment 12

[Daniel Veditz [:dveditz]]

r=dveditz for the webshell,rdf and xpfe patches added since my last review

Comment 13

[Alec Flett]

Attachment: mailnews patch

Comment 14

[Alec Flett]

adding mscott/sspitzer for mail r=/sr=

Comment 15

[Alec Flett]

the one funky thing in the mail review is that we were doing some INSANE stuff
in addressbook in order to get at the non-interface class nsAbAutoCompleteParam
- I've fixed it by switching to straight casting. The objects in the
nsISupportsArray are only addeded by the autocomplete code, so we always know
what kind of objects they are..

HOWEVER, someone needs to fix it - my change is no less evil than the old code.

Comment 16

[Alec Flett]

just going through the patch for the nth time, I noticed there's a random XUL
change - ignore that, I won't check that in.

Comment 17

[(not reading, please use [e-mail] instead)]

r=sspitzer for the mailnews stuff.

please log a bug on the "HOWEVER, someone needs to fix it - my change is no less
evil..." problem so we remember to fix it.

thanks for the fix.

Comment 18

[Scott MacGregor]

sr=mscott for the mailnews patch.

Comment 19

[Brendan Eich [:brendan]]

Looks like you can nuke kICollationFactoryIID from nsCollation.cpp now:

http://lxr.mozilla.org/mozilla/source/intl/locale/src/nsCollation.cpp#37

Same for kILocaleIID in nsLocale.cpp, for kICollationIID in nsCollationWin.cpp,
and maybe for kIWin32LocaleIID too?  More: kICharsetAliasIID in
nsCharsetAliasImp.cpp, kITextToSubURIIID in nsTextToSubURI.cpp, kIModuleIID in
nsUConvModule.cpp, kIUnicodeDecodeHelperIID in nsUnicodeDecodeHelper.cpp,
kIPlatformCharsetIID in nsWinCharset.cpp, 

That's my review through intl; gfx, etc. next but I need a break.

/be

Comment 20

[Alec Flett]

moving a few 0.9 bugs out to 1.0 to lighten my 0.9 load.

Comment 21

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

There aren't that many left to be fixed:

http://lxr.mozilla.org/seamonkey/search?string=NS_IMPL_ISUPPORTS%28
http://lxr.mozilla.org/seamonkey/search?string=NS_IMPL_QUERY_INTERFACE%28

Comment 22

[Paul Chen]

nav triage team:

Not a 1.0 stopper, pushing out to mozilla1.1

Comment 23

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

I think we should clean these things up for 1.0.  Taking bug.

Comment 24

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Attachment: patch to remove all but a handful of uses

Comment 25

[jag (Peter Annema)]

r=jag with the changes made as suggested in #mozilla.

Comment 26

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Attachment: patch addressing jag's comments on IRC

Comment 27

[Brendan Eich [:brendan]]

rs=[e-mail]

Comment 28

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Attachment: final pass for mozilla tree, not yet tested

Comment 29

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Tested.  Mozilla is fine.  TestArray is fine.

Comment 30

[jag (Peter Annema)]

Comment on attachment 47209 [details] [diff] [review]
final pass for mozilla tree, not yet tested

>Index: mozilla/modules/oji/tests/src/TestLoader/OJITestLoader.cpp
>===================================================================
>RCS file: /cvsroot/mozilla/modules/oji/tests/src/TestLoader/OJITestLoader.cpp,v
>retrieving revision 1.1
>diff -u -d -r1.1 OJITestLoader.cpp
>--- OJITestLoader.cpp	2000/05/03 01:03:53	1.1
>+++ OJITestLoader.cpp	2001/08/27 15:01:13
>@@ -24,9 +24,7 @@
> static NS_DEFINE_IID(kISupportsIID,    NS_ISUPPORTS_IID);
> static NS_DEFINE_IID(kIOJITestLoaderIID, OJITESTLOADER_IID);
> 
>-NS_IMPL_QUERY_INTERFACE(OJITestLoader, kIOJITestLoaderIID);
>-NS_IMPL_ADDREF(OJITestLoader)
>-NS_IMPL_RELEASE(OJITestLoader)
>+NS_IMPL_ISUPPORTS(OJITestLoader, OJITestLoader)
> 
> OJITestLoader::OJITestLoader(void) 
> {

Shouldn't that be:

+NS_IMPL_ISUPPORTS1(OJITestLoader, OJITESTLOADER_IID)
?

Comment 31

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Changed in my tree to:

NS_IMPL_ISUPPORTS1(OJITestLoader, OJITestLoader)

(That file doesn't seem to be built on any platform.)

Comment 32

[jag (Peter Annema)]

Comment on attachment 47209 [details] [diff] [review]
final pass for mozilla tree, not yet tested

r= with that fixed

Comment 33

[Alec Flett]

Comment on attachment 47209 [details] [diff] [review]
final pass for mozilla tree, not yet tested

looks great.
sr=alecf

Comment 34

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Checked in 2001-09-10 12:54 PDT.  Now I need to do the commercial tree, find out
if there are any other users, and remove from nsISupportsUtils.h...

Comment 35

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

I'm currently building patches to complete the removal of
NS_IMPL_QUERY_INTERFACE and related macros (but no numbered macros!).  This
makes the remaining necessary changes to callers and removes the macros from
nsISupports{Impl,Obsolete}.h.  It also moves some non-deprecated macros (the
NS_IMPL_ISUPPORTSn macros) from nsISupportsObsolete.h back into
nsISupportsImpl.h (where the threadsafe forms stayed the whole time).

Comment 36

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Attachment: final (really) pass

Comment 37

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Commercial tree patch is in http://bugscape.netscape.com/show_bug.cgi?id=9952 .

Comment 38

[jag (Peter Annema)]

Comment on attachment 51464 [details] [diff] [review]
final (really) pass

r=jag

Comment 39

[Alec Flett]

Comment on attachment 51464 [details] [diff] [review]
final (really) pass

kick ass!
sr=alecf

Comment 40

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Fix checked in 2001-10-01 20:17/18 PDT.

Comment 41

[Jim Dunn]

I was told that the JPI contains the use of NS_IMPL_QUERY_INTERFACE
and that implementers/ports of the java plugin will have to also
follow this.  Is there any general guideline we are telling people
who port java to their platforms?  Has Sun modified the java source
yet?  

Comment 42

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

NS_IMPL_QUERY_INTERFACE1 should be used instead of NS_IMPL_QUERY_INTERFACE. 
What's the JPI and how do you find out what it uses, anyway?

Comment 43

[Jim Dunn]

Java Plugin Interface.
You know the thing that comes with Java 1.2 so that mozilla
can run java.  

My guess is that java will still work, it is just that Sun
needs to be made aware that this happened, so that they
can change the source for the jpi and remove this.