575014 - nsTArray::LastIndexOf triggers "warning C4244: 'return' : conversion from '__int64' to 'nsTArray_base::index_type', possible loss of data" on Win64

Status: RESOLVED FIXED

(Core :: XPCOM, defect)

Description

[Daniel Holbert [:dholbert]]

My FAIL_ON_WARNINGS lading in bug 557566 comment 28 triggered a build failure on 64-bit Windows, due to this warning:
> nsDocLoader.cpp
> e:\builds\moz2_slave\mozilla-central-win64-debug-nightly\build\obj-firefox\dist\include\nsTArray.h(416) : error C2220: warning treated as error - no 'object' file generated
[...]
> e:\builds\moz2_slave\mozilla-central-win64-debug-nightly\build\obj-firefox\dist\include\nsTArray.h(416) : warning C4244: 'return' : conversion from '__int64' to 'nsTArray_base::index_type', possible loss of data

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1277595343.1277596686.26568.gz

I'd run the patch through TryServer, but I guess TryServer doesn't do Win64 (nor does normal tinderbox actually, except to generate a few Win64 nightlies per day.)

Fix coming up.

Comment 1

[Daniel Holbert [:dholbert]]

Attachment: fix: cast pointer-subtraction to be "index_type"

I think this should fix it.

Comment 2

[Daniel Holbert [:dholbert]]

(Ultimately we might want to make nsTArray::index_type be size_t, but for now, the type-casting is a more conservative fix.)

Comment 3

[Daniel Holbert [:dholbert]]

Landed patch with r=timeless and pending-review=bsmedberg, so that the bustage on Win64 nightlies will (hopefully) stop, and also since there's no other way for me to test this :-/  (No Win64 tryservers)

http://hg.mozilla.org/mozilla-central/rev/dc28988730e8

Comment 4

[Daniel Holbert [:dholbert]]

Nice, the next Win64 build (after this patch landed) was green:
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1277607223.1277613689.30944.gz

Comment 5

[Daniel Holbert [:dholbert]]

Resolving this as fixed -- had initially left this open with pending-r=bsmedberg, but timeless reviewed it, and it doesn't need another review.