Loading...

XML

Word

Printable

Details

Type: Bug
Resolution: Fixed
Priority: P1: Critical
Fix Version/s: 5.15.12 VxWorks, 6.5.3, 6.6.0 Beta4, 6.7.0 FF
Affects Version/s: 6.5.2, 6.6.0 Beta2, dev
Component/s: Core: Containers and Algorithms
Labels:
None
Environment:
Ubuntu 20.04, GCC 13 (but probably every compiler that natively supports f16).

Epic Link:
consistent-qHash
Story Points:
2
Commits:
c86cf385d (dev), 2d4db1bff (6.5), 3a9bbf235 (6.6), 6da6a17de (dev), e8fb761f7 (6.6), eb5be509d (6.5), 0f19cafc3 (dev), 8e14e13c1 (6.6), 3974778dc (6.5), fdfbb6106 (6.5.3), 096a11225 (6.6.0)
Sprint:
Foundation Sprint 90

Description

This code worked fine on GCC 13 up to Qt 6.4, but breaks in Qt 6.5:

auto h = qHash(qfloat16(-42.f));

In Qt 6.5, it breaks with the following error:

tst.cpp:5:19: error: call of overloaded ‘qHash(qfloat16)’ is ambiguous
    5 |   return int(qHash(qfloat16(-42.0f)));
      |              ~~~~~^~~~~~~~~~~~~~~~~~

It goes on to list all the qHash() overloads, incl. the three FP ones: float, double, long double. It seems it no longer prefers the float overload.

Attachments

- Sort By Name
- Sort By Date
- Ascending
- Descending
- Thumbnails
- List

tst.cpp
0.1 kB
15 Aug '23 09:29
tst.pro
0.0 kB
15 Aug '23 09:29

Issue Links

relates to

QTBUG-116077 Inconsistent hashing between different FP types

Reported

QTBUG-116076 qHash(qfloat16(x), seed) != qHash(float(x), seed) unless seed == 0

Closed

QTBUG-116080 Inconsistent hashing between qint{32,64} on 32-bit platforms

Closed

QTBUG-126674 Inconsistent hashing between bool and integral overloads

Closed

Gerrit Reviews

- Issue Only
- Show All Reviews
- Show Open Reviews
- Show All Issues
- Show Open Issues

For Gerrit Dashboard: QTBUG-116064
#	Subject	Branch	Project	Status	CR	V
496753,3	tst_QHashFunctions: extend the consistency() test with int/FP types	dev	qt/qtbase	Status: MERGED	+2	0
497081,2	tst_QHashFunctions: extend the consistency() test with int/FP types	6.6	qt/qtbase	Status: MERGED	+2	0
497082,2	tst_QHashFunctions: extend the consistency() test with int/FP types	6.5	qt/qtbase	Status: MERGED	+2	0
503188,4	Fix qHash(qfloat16) to match Qt 6.4 behavior	dev	qt/qtbase	Status: MERGED	+2	0
503546,2	Fix qHash(qfloat16) to match Qt 6.4 behavior	6.6	qt/qtbase	Status: MERGED	+2	0
503577,2	Fix qHash(qfloat16) to match Qt 6.4 behavior	6.5	qt/qtbase	Status: MERGED	+2	0
504567,3	[docs] Fix \since for qHash(qfloat16)	dev	qt/qtbase	Status: MERGED	+2	0
505193,2	[docs] Fix \since for qHash(qfloat16)	6.6	qt/qtbase	Status: MERGED	+2	0
505395,2	[docs] Fix \since for qHash(qfloat16)	6.6.0	qt/qtbase	Status: MERGED	+2	0
505396,2	[docs] Fix \since for qHash(qfloat16)	6.5	qt/qtbase	Status: MERGED	+2	0
505438,2	[docs] Fix \since for qHash(qfloat16)	6.5.3	qt/qtbase	Status: MERGED	+2	0