Actually, if you put a CAST after a relation (!= ``is`` or ``is_instance_of``),
the checker raises an error saying that the etype (inside the CAST) must be
after a ``is`` or ``is_instance_of`` relation. In fact the CAST case was tested too late.

This allow to use CAST outside the selection section, so:
Closes #79232

Added tag rql-version-0.32.0, rql-debian-version-0.32.0-1, rql-centos-version-0.32.0-1 for changeset 843843e76fac

Allow rql to work with yapps 2.2.0 which is properly released on pypi, and remove the custom dependency link for it.

Closes #124962
Closes #245540

Allow multiple (non "=") operators for uid variables.

It just degrades the readability for no apparent benefit.

Closes #190394.

Added tag rql-centos-version-0.31.4-1, rql-version-0.31.4, rql-debian-version-0.31.4-1 for changeset 5cad47e1ffbf

memory statistics have been removed from the search engines.

And use it in the Debian packaging to make sure we don't end up with a
disfunctional package.
Closes #159003

Related to #159003

Looks like they changed enums to functions.

Closes #159002

In some cases (eg, windows without Visual Studio installed),
distutils/setuptools build_ext.run() may raise exceptions other than
CompileError. By catching all exceptions, we can make sure rql's
setup.py never fails, which should help dumb package managers such as
pip.

When the variable type is defined by an IN() function, ``add_type_restriction``
removes all types in it that don't match the specified type. Since the list
of types is modified inplace, the iteration must be done on a copy, not on
the list itself.

There was a test for "add_type_restriction + IN()" case but since there
was only 2 types defined in the IN function, the problem was not exposed.

make it look more like our other modules