Bernd, watch out on your current branch....

[neroden]

The signal placement patch introduced a failure to compile with GCC 4.4 which I *can't debug*.  It dies at linker time with

Code Select Expand


gui/signal_spacing.o: In function `non-virtual thunk to signal_spacing_frame_t::action_triggered(gui_action_creator_t*, value_t)':
/home/neroden/programming/simutrans/gui/../ifc/gui_fenster.h:41: multiple definition of `non-virtual thunk to signal_spacing_frame_t::action_triggered(gui_action_creator_t*, value_t)'
gui/signal_spacing.o:/home/neroden/programming/simutrans/gui/../ifc/gui_fenster.h:41: first defined here
gui/signal_spacing.o: In function `signal_spacing_frame_t::action_triggered(gui_action_creator_t*, value_t)':
/home/neroden/programming/simutrans/gui/signal_spacing.cc:58: multiple definition of `signal_spacing_frame_t::action_triggered(gui_action_creator_t*, value_t)'
gui/signal_spacing.o:/home/neroden/programming/simutrans/gui/signal_spacing.cc:58: first defined here
gui/signal_spacing.o:(.data+0x2): multiple definition of `signal_spacing_frame_t::replace'
gui/signal_spacing.o:(.data+0x2): first defined here
gui/signal_spacing.o:(.data+0x1): multiple definition of `signal_spacing_frame_t::remove'
gui/signal_spacing.o:(.data+0x1): first defined here
gui/signal_spacing.o:(.data+0x0): multiple definition of `signal_spacing_frame_t::signal_spacing'
gui/signal_spacing.o:(.data+0x0): first defined here
gui/signal_spacing.o: In function `signal_spacing_frame_t':
/home/neroden/programming/simutrans/gui/signal_spacing.cc:19: multiple definition of `signal_spacing_frame_t::signal_spacing_frame_t(spieler_t*, wkz_roadsign_t*)'
gui/signal_spacing.o:/home/neroden/programming/simutrans/gui/signal_spacing.cc:19: first defined here
gui/signal_spacing.o: In function `signal_spacing_frame_t':
/home/neroden/programming/simutrans/gui/signal_spacing.cc:19: multiple definition of `signal_spacing_frame_t::signal_spacing_frame_t(spieler_t*, wkz_roadsign_t*)'
gui/signal_spacing.o:/home/neroden/programming/simutrans/gui/signal_spacing.cc:19: first defined here
collect2: ld returned 1 exit status
make: *** [sim] Error 1


(EDIT: No idea why.) Anyway, it's beyond my C++ expertise.

EDIT: Mods, this belongs in the experimental forum.  It's getting late.  :-(

[prissi]

I would say there is a virtual missing before one of the overlaid definitions. Or this is a way to warn before a virtual function call, if there is no action creator?

[neroden]

I would say there is a virtual missing before one of the overlaid definitions. Or this is a way to warn before a virtual function call, if there is no action creator?

There was a "virtual" missing before one of the two definitions of get_resizemodes, but adding it didn't fix the problem.

There's something very weird about the warning message.
gui/signal_spacing.o:(.data+0x2): multiple definition of `signal_spacing_frame_t::replace'
gui/signal_spacing.o:(.data+0x2): first defined here

Notice that both of the "multiple definitions" are in the *same location*.

That and the fact that this is a linker error rather than a compiler error.

If this were templated code I would guess it was a problem with the same template being instantiated independently in two different compilation units (files), creating separate definitions for the same class in two compilation units.  But it's not templated code.

EDIT: Maybe there's a double definition of the resizemodes enum....

[jamespetts]

Although the signal spacing patch is incorporated in Experimental, it was originally from a patch submitted by Dwachs for Standard, and there is an up-to-date Standard compatible patch that would, I assume, have the same issues...

[neroden]

Duplicate entry in the Makefile confused things.     Fixed on my jp-devel branch.