[FEATURE] Allow comparison with null in `<f:if>` ViewHelper

[simonschaufi]

It is not possible to use a condition with the <f:if /> ViewHelper to check directly against null value, like it is known from the PHP world.

For example, following php-code

  // strict comparison
  if ($variable === null) {}
  if ($variable !== null) {}

  // loose comparision
  if ($variable == null) {}
  if ($variable != null) {}

can now be written with fluid as

  <!-- /* strict comparison */ -->
  <f:if condition="{var} === null"></f:if>
  <f:if condition="{var} !== null"></f:if>

  <!-- /* loose comparison */ -->
  <f:if condition="{var} == null"></f:if>
  <f:if condition="{var} != null"></f:if>

which extinguish the need of a custom isNull ViewHelper in projects.

[simonschaufi]

In my opinion null should be added as general language construct and not parsed from string at occasional places.

I have no idea what exactly you mean by "general language construct". Can you given an example? Boolean things are parsed in the BooleanParser.

[bnf]

There needs to be something like a TYPO3Fluid\Fluid\Core\Parser\SyntaxTree\NumericNode TYPO3Fluid\Fluid\Core\Parser\SyntaxTree\TextNode but for null values.

=> TYPO3Fluid\Fluid\Core\Parser\SyntaxTree\NullNode.

So that null is a language construct that is understood in all expressions, not just in boolean.

[simonschaufi]

@bnf I've started the implementation but now I'm stuck. I tried to copy what the NumericNode does but maybe I'm going the wrong way here. I also don't really understand what the BooleanNode does as it requires a comparison and not just "true" or "false". Can you help me please?

[bnf]

Sorry, I can't provide further guidance without implementing it myself (which is currently not possible due to priorities).

[s2b]

In my opinion null should be added as general language construct and not parsed from string at occasional places. Supporting it in conditions, but not contexts like parameters makes no sense from a developer experience and makes it hard to grasp what fluid actually supports.

Further discussion about this can happen in #911. I'm closing this PR, since we're taking a different approach.