56623 - Possible typo in gfx/src/nsGraphicsImpl.cpp since rev 1.1

Status: VERIFIED FIXED

(Core Graveyard :: GFX, defect, P3)

Description

[Hallvard B Furuseth]

Gcc warns about the line
  size *= mDev2App;
which actually does
  size *= (nscoord) mDev2App; // cast floating to integer
but looks like (and is maybe meant to do)
  size =  (nscoord) (size * mDev2App);
because size is integral and mDev2App floating.

Anyway, to kill the warning, spell it out as one of these.

Comment 1

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

->Compositor

Comment 2

[Hallvard B Furuseth]

Sorry, I should've know better than to submit a report just before
I rushed out the door...  What happens is exactly the opposite of
what I said.  Don't know how I managed to twist the output to show
the opposite.
Oh well, at least the proposed change does kill off a warning.

Comment 3

[Kevin McCluskey (gone)]

Setting milestone to mozilla1.0.1
Reassigning to Waqar.

Comment 4

[Kevin McCluskey (gone)]

Attachment: Patch to fix warning

Comment 5

[Kevin McCluskey (gone)]

Taking this bug

Comment 6

[Marc Attinasi]

NS_STATIC_CAST instead of direct c-style caset maybe? sr=attinasi

Comment 7

[Kevin McCluskey (gone)]

Attachment: New patch that using NS_STATIC_CAST

Comment 8

[dcone (gone)]

looks good.
r=dcone

Comment 9

[Hallvard B Furuseth]

Last patch, unlike the first one, changes <size>'s new value from
   (nscoord) (size * mDev2App)
to
   size * (nscoord) mDev2App
I hope that was intentional.

This bug has had me wishing for a cancel button for silly Bugzilla reports:-(

Comment 10

[Kevin McCluskey (gone)]

Attachment: New patch for warning which using NS_STATIC and does the calculation correctly

Comment 11

[Kevin McCluskey (gone)]

Fix checked in

Comment 12

[Chris Petersen]

Marking verified in the May 30th build.