2019-03-03 Rico Tzschichholz Release 0.40.14 2019-02-28 Rico Tzschichholz girwriter: Always use get_gir_name() for TypeSymbols poppler-glib: Update to 0.74 girwriter: Support transfer-ownership="container" for arrays gobject-2.0: Add SignalHandler.disconnect_by_data() codegen: Respect array_length_cname attribute for global fields tests: Add "custom types" struct tests to increase coverage tests: Add "constructor chain up" generics test to increase coverage tests: Add "null-terminated array" delegate test to increase coverage tests: Add "struct in array" test to increase coverage tests: Add "to_string" enum test to increase coverage tests: Add regression test for modulo operation on double and float https://bugzilla.gnome.org/show_bug.cgi?id=610660 girparser: Correctly set array_length_type for delegates returning an array Fixes https://gitlab.gnome.org/GNOME/vala/issues/754 libvaladoc: Properly support the output of async constructors Fixes https://gitlab.gnome.org/GNOME/vala/issues/753 clutter-1.0: Drop deprecated from Actor.pick() The signal "pick" was deprecated with 1.12, but not the virtual method. Fixes https://gitlab.gnome.org/GNOME/vala/issues/625 valadoc: Report warning if resources for doclet could not be copied vapi: Drop gdk-x11-3.0 from GIDL_BINDINGS gtk+-unix-print-*.0: Fix PrintJob.get_page_ranges(), Printer.get_hard_margins() codegen: Pass lambda expression of variable initializer to signal connect gdk-3.0: Update to 3.24.4+a0129f55 g-i: Fix return value on error in start_discriminator() libvaladoc: Don't change the size of an array in-parameter Avoid a possible memory issue while the resize of an array reference is possible, the new length can't be propagated back to the caller. Issue an warning for unexpected string lengths. vala: Parser doesn't ensure "protected" to be applied on TypeSymbol members Probably ObjectTypeSymbol should be enforced. vala: Check inferred variable_type of LocalVariables codegen: Properly handle and catch inner-error of finally-block If all inner-errors are caught there is no jump-out of the current finally- block therefore the control-flow analyzer is happy. To make this work the surrounding inner-error must not be used or changed here. Fixes https://gitlab.gnome.org/GNOME/vala/issues/742 codegen: Replace some unfortunate asserts with internal error reports