[QGIS-Developer] Replacing QLiteralString() and QLatin1String() by qt 6.4 u""_s and ""_L1 ?

[David Koňařík]

Hi,
I'm glad you've opened this discussion. I'm personally in favour of 
moving from the macros to the new string literal operators, since I 
think that both have next to no value for a programmer reading the code, 
so I favour the less intrusive option.

Though as far as I know, we still support building QGIS with Qt 5, so 
this will have to wait until we're officially Qt 6-only.

David Koňařík

On 12/21/25 19:50, Even Rouault via QGIS-Developer wrote:
> Hi,
> 
> I personally find that QLiteralString() and QLatin1String()  add 
> significant visual noise when reading QGIS source code. Qt 6.4 adds a 
> Qt::Literals::StringLiterals namespace (https://doc.qt.io/qt-6/qt- 
> literals-stringliterals.html) with operator"" _s(...) and 
> operator""_operator ""_L1(...)
> 
> So code like
> 
>    wallProperties.insert( QStringLiteral( "geometryModifier" ), 
> WALL_EXPRESSION );
>    wallProperties.insert( QStringLiteral( "symbolType" ), 
> QStringLiteral( "Fill" ) );
>    if ( renderer->type() == QLatin1String( "25dRenderer" ) ) { ... }
> 
> can be simplified as
> 
>    wallProperties.insert( u"geometryModifier"_s, WALL_EXPRESSION );
>    wallProperties.insert( u"symbolType"_s, u"Fill"_s );
>    if ( renderer->type() == "25dRenderer"_L1 ) { ... }
> 
> For the replacement of QStringLiteral() by ""_s is is a bit unfortunate 
> though that we have to put the 'u' prefix to indicate this is a Unicode 
> string, but u""_s is still shorter than QStringLiteral( "" )
> 
> Thoughts on doing a mass replacement to those new ways?
> 
> Even
>