mirror of
https://gitlab.com/freepascal.org/fpc/source.git
synced 2025-10-27 08:31:33 +01:00
37b5c061e3
via {$modeswitch finalfields} (on by default on the JVM target). The
meaning is the same as in Java: a final (class) field can only be set
in a (class) constructor of the class it's defined in, and can only be
written once there (and *must* be set there). They are currently only
supported for external classes since that basically turns them into
constants, since for non-external classes we need full dataflow analysis
o refactored pdecobj.parse_object_members() a bit in the process to reduce
the amount of repetition (which would have been further increased for
the support for final fields)
o made error message about "wrong use of absolute" for fields etc generic,
so it gives a proper error depending on which token was used (it had
to be made generic for "final" support, but already was used for other
things that were wrongly reported as "absolute" misusages)
git-svn-id: branches/jvmbackend@18398 -
|
||
|---|---|---|
| .. | ||
| errorct.msg | ||
| errord.msg | ||
| errorda.msg | ||
| errordu.msg | ||
| errore.msg | ||
| errores.msg | ||
| errorf.msg | ||
| errorfi.msg | ||
| errorhe.msg | ||
| errorheu.msg | ||
| errorid.msg | ||
| erroriu.msg | ||
| errorn.msg | ||
| errorpl.msg | ||
| errorpli.msg | ||
| errorpt.msg | ||
| errorptu.msg | ||
| errorr.msg | ||
| errorru.msg | ||
| errorues.msg | ||