Bug 231044

Summary:

Allow NaN, infinity, and -infinity in calc

Product:

WebKit

Reporter:

Nikos Mouchtaris <nmouchtaris>

Component:

New Bugs

Assignee:

Nikos Mouchtaris <nmouchtaris>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

clopez, esprehn+autocc, ews-watchlist, glenn, gyuyoung.kim, jonlee, macpherson, menard, seokho, simon.fraser, webkit-bug-importer, youennf

Priority:

Keywords:

InRadar

Version:

WebKit Nightly Build

Hardware:

Unspecified

OS:

Unspecified

Attachments:

Description	Flags
Patch	none
Patch	none
Patch	none
Patch	none
Patch	simon.fraser: review+, ews-feeder: commit-queue-
infinity	none

Description Nikos Mouchtaris 2021-09-30 16:24:39 PDT

Allow NaN, infinity, and -infinity in calc

Comment 1 Nikos Mouchtaris 2021-09-30 16:32:37 PDT

Created attachment 439792 [details]
Patch

Comment 2 Radar WebKit Bug Importer 2021-10-01 10:32:32 PDT

<rdar://problem/83772269>

Comment 3 Nikos Mouchtaris 2021-10-01 14:13:17 PDT

Created attachment 439910 [details]
Patch

Comment 4 EWS Watchlist 2021-10-01 14:14:43 PDT

This patch modifies the imported WPT tests. Please ensure that any changes on the tests (not coming from a WPT import) are exported to WPT. Please see https://trac.webkit.org/wiki/WPTExportProcess

Comment 5 Nikos Mouchtaris 2021-10-01 14:56:42 PDT

Created attachment 439914 [details]
Patch

Comment 6 Nikos Mouchtaris 2021-10-01 15:02:57 PDT

Created attachment 439915 [details]
Patch

Comment 7 Simon Fraser (smfr) 2021-10-01 15:17:35 PDT

Comment on attachment 439915 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=439915&action=review

> Source/WebCore/css/CSSPrimitiveValue.cpp:955
> +    if (m_value.num == -1*std::numeric_limits<double>::infinity())

Spaces around *

> Source/WebCore/css/CSSValueKeywords.in:1369
> +INFINITY

Why is this capitalized?

> Source/WebCore/css/calc/CSSCalcExpressionNodeParser.cpp:100
> +    return { { CSSValuePi, CSSUnitType::CSS_NUMBER, piDouble }, { CSSValueE, CSSUnitType::CSS_NUMBER, std::exp(1.0) }, { CSSValueInfinity, CSSUnitType::CSS_NUMBER, -1*std::numeric_limits<double>::infinity() }, { CSSValueINFINITY, CSSUnitType::CSS_NUMBER, std::numeric_limits<double>::infinity() }, { CSSValueNaN, CSSUnitType::CSS_NUMBER, std::numeric_limits<double>::quiet_NaN() } };

This need wrapping. CSSValueINFINITY is ugly.

Comment 8 Simon Fraser (smfr) 2021-10-01 15:29:57 PDT

Comment on attachment 439915 [details]
Patch

r- for the -infinity confusion

Comment 9 Nikos Mouchtaris 2021-10-01 16:10:16 PDT

Created attachment 439924 [details]
Patch

Comment 10 Simon Fraser (smfr) 2021-10-01 16:14:43 PDT

Comment on attachment 439924 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=439924&action=review

> Source/WebCore/css/calc/CSSCalcExpressionNodeParser.cpp:103
> +    return { { CSSValuePi, CSSUnitType::CSS_NUMBER, piDouble }, { CSSValueE, CSSUnitType::CSS_NUMBER, std::exp(1.0) },
> +        { CSSValueNegativeInfinity, CSSUnitType::CSS_NUMBER, -1 * std::numeric_limits<double>::infinity() },
> +        { CSSValueInfinity, CSSUnitType::CSS_NUMBER, std::numeric_limits<double>::infinity() },
> +        { CSSValueNaN, CSSUnitType::CSS_NUMBER, std::numeric_limits<double>::quiet_NaN() } };

I would wrap this like:

return {
    { CSSValuePi, CSSUnitType::CSS_NUMBER, piDouble }, { CSSValueE, CSSUnitType::CSS_NUMBER, std::exp(1.0) },
    { CSSValueNegativeInfinity, CSSUnitType::CSS_NUMBER, -1 * std::numeric_limits<double>::infinity() },
    { CSSValueInfinity, CSSUnitType::CSS_NUMBER, std::numeric_limits<double>::infinity() },
    { CSSValueNaN, CSSUnitType::CSS_NUMBER, std::numeric_limits<double>::quiet_NaN() },
};

Comment 11 Nikos Mouchtaris 2021-10-01 18:54:01 PDT

Created attachment 439947 [details]
infinity

Comment 12 EWS 2021-10-01 23:23:04 PDT

Committed r283434 (242421@main): <https://commits.webkit.org/242421@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 439947 [details].

Comment 13 Tim Nguyen (:ntim) 2021-11-07 01:38:16 PST

*** Bug 219832 has been marked as a duplicate of this bug. ***