133801 - nsContentTestNode.cpp breaks 64-bit builds on AIX

Status: VERIFIED FIXED

(Core :: XUL, defect)

Description

[Philip K. Warren]

With the checkin for Bug 118863, content/xul/templates/src/nsContentTestNode.cpp
breaks when building with --enable-debug for 64-bit AIX builds.

Comment 1

[Philip K. Warren]

Attachment: Simple fix - PRInt32->NS_PTR_TO_INT32

Comment 2

[Roland Mainz]

pkw:
Can you post the compiler error message that someone can mark this bug as
"confirmed", please ?

Comment 3

[Roland Mainz]

Comment on attachment 76460 [details] [diff] [review]
Simple fix - PRInt32->NS_PTR_TO_INT32

Looks OK for me...

r=[e-mail]

Comment 4

[Philip K. Warren]

Sorry. Here is the error message:

"/home/pkw/mozilla/content/xul/templates/src/nsContentTestNode.cpp", line 96.31:
1540-0216 (S) An expression of type "nsIContent *" cannot be converted to "PRInt32".
gmake[1]: *** [nsContentTestNode.o] Error 1
gmake[1]: Leaving directory
`/home/pkw/mozilla/obj-debug64/content/xul/templates/src'
gmake: *** [all] Error 2

Comment 5

[Roland Mainz]

Confirming per comment #4 - thanks!

Comment 6

[Brendan Eich [:brendan]]

Comment on attachment 76460 [details] [diff] [review]
Simple fix - PRInt32->NS_PTR_TO_INT32

Easy fix, good enough.	If any 64-bit system needs to log the high 32 bits of a
pointer, this can be changed to use

    aResult += nsPrintfCString("%p", aContext);

or some more efficient formulation using nsPrintfCString.

/be

Comment 7

[Brendan Eich [:brendan]]

Nominating for mozilla1.0.

pkw, Thanks for the fix, please mail [e-mail] asking for approval --
tell them who reviewed, let them know it's <= 0 risk :-).

/be

Comment 8

[Philip K. Warren]

Brendan:

Would your fix be more appropriate in this case? It seems to just be a function
used for debugging, and I would assume printing the entire pointer would be the
desired behavior.

Comment 9

[Brendan Eich [:brendan]]

My fix would be purer, but it's late in 1.0 and we don't really care about DEBUG
only potential high-bit of pointer data loss all that much.  Unless you're
really motivated to do the %p thing and get a new patch reviewed and
super-reviewed, I say go for driver approval with what we have here.

/be

Comment 10

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Comment on attachment 76460 [details] [diff] [review]
Simple fix - PRInt32->NS_PTR_TO_INT32

a=roc+moz

Comment 11

[Jim Dunn]

checked in

Comment 12

[shrirang khanzode]

.