666984 - Remove/substitute NS_FloatIsFinite since it's the same as NS_finite

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

One of these is more than enough... CC'ing somebody who might be interested :)

Comment 1

[Ed Morley [:emorley]]

Attachment: Patch v1

http://mxr.mozilla.org/mozilla-central/search?string=NS_FloatIsFinite
http://mxr.mozilla.org/mozilla-central/search?string=NS_finite

First pass at this, let me know if this wasn't the intended approach.

Compiles fine locally, with no warnings on the changed lines (the rest of /content/svg/ is another matter, but don't get me started! More to do in bug 187528, but anyway).

However, slight query about how liberal I should be about #includes ? ie: I've added a few to make it compile, but should I be including nsMathUtils.h in absolutely every file I've added NS_finite into, to prevent it only being included by chance via another file? 

Likewise, in the files where I've s/NS_FloatIsFinite/NS_finite/g, I've removed the existing #include "nsContentUtils.h", on the premise that NS_FloatIsFinite is no longer being used. However it's plausible that nsContentUtils.h is being used for other stuff and it only compiles by chance inclusion elsewhere. 

Basically I don't want to make the bug 557565 / bug 634839 type situation any worse - and don't know what the recommendations are here?

Comment 2

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

The "no bootlegging" principle would suggest including nsMathUtils.h everywhere. For nsContentUtils.h, I'd remove it (unless the file uses an nsContentUtils::* function, of course).

Comment 3

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Comment on attachment 544512 [details] [diff] [review]
Patch v1

>--- a/content/svg/content/src/SVGLength.h
>+++ b/content/svg/content/src/SVGLength.h

>-#include "nsContentUtils.h"

>--- a/content/svg/content/src/SVGPoint.h
>+++ b/content/svg/content/src/SVGPoint.h

>-#include "nsContentUtils.h"

Excellent.

>--- a/xpcom/ds/nsMathUtils.h
>+++ b/xpcom/ds/nsMathUtils.h

>+/*

/**, please.

>+ * Check whether a floating point number is finite (not +/-infinity and not a
>+ * NaN value).
>+ */
> inline NS_HIDDEN_(bool) NS_finite(double d)
> {
> #ifdef WIN32
>     // NOTE: '!!' casts an int to bool without spamming MSVC warning C4800.
>     return !!_finite(d);
> #else
>     return finite(d);
> #endif

Looks great!

Comment 4

[Ed Morley [:emorley]]

Attachment: Patch v2

Thanks Ms2ger :-)

Same as prior patch, other than nsMathUtils.h being included everywhere to avoid bootlegging + the nit for the nsMathUtils.h comment style.

Comment 5

[Ed Morley [:emorley]]

http://dev.philringnalda.com/tbpl/?tree=Try&rev=7e2dec83b318

Comment 6

[Ed Morley [:emorley]]

Ping for review. Thanks :-)

Comment 7

[Jonathan Watt [:jwatt]]

Comment on attachment 544578 [details] [diff] [review]
Patch v2

Sorry, totally forgot about this. r=jwatt although it would be nice if you could tidy up the trailing "\" chars on the lines in nsContentUtils.h.

Comment 8

[Ed Morley [:emorley]]

Attachment: Patch v2.1

Updated to tip, tidied nsContentUtils.h trailing "\" chars; carrying forwards r+.

Thanks Jonathan :-)

Comment 9

[Ed Morley [:emorley]]

Attachment: Patch v2.2

Added missing #include "nsMathUtils.h" in nsSVGIntegerPair.cpp after updating to tip. Carrying forwards r+.

Comment 10

[Boris Zbarsky [:bzbarsky]]

http://hg.mozilla.org/integration/mozilla-inbound/rev/c5e091d2dd38

Comment 11

[Marco Bonardo [:mak]]

http://hg.mozilla.org/mozilla-central/rev/c5e091d2dd38