Re: Exception propagation problem on IA-64/Linux

[Eric Botcazou <ebotcazou at adacore dot com>]

> Yes, that's wrong.  I'd expect that to be a front-end bug, but if it
> doesn't happen on all platforms, then, maybe it's not?

I think it can happen for all targets that use DW_EH_PE_indirect incoding.  
And it happens in Ada too because, like in C++, local exceptions are not 
expected to be visible outside the compilation (translation) unit so they are 
not uniquified.

I think the bug is simply in dw2_force_const_mem.  Here's a more verbose 
excerpt of the assembly files:

 .LLSDACSE2:
        data1   0x1
        data1   0x0
        .align 8
        data8.ua        @gprel(DW.ref._ZTIZ3foovE1S#)
.LLSDATT2:

[...]

        .section        .rodata
        .align 8
        .type   _ZTIZ3foovE1S#, @object
        .size   _ZTIZ3foovE1S#, 16
_ZTIZ3foovE1S:
        data8   _ZTVN10__cxxabiv117__class_type_infoE#+16
        data8   _ZTSZ3foovE1S#
        .align 8
        .type   _ZTSZ3foovE1S#, @object
        .size   _ZTSZ3foovE1S#, 10
_ZTSZ3foovE1S:
        stringz "Z3foovE1S"
        .hidden DW.ref._ZTIZ3foovE1S
        .weak   DW.ref._ZTIZ3foovE1S#
        .section        .gnu.linkonce.s.DW.ref._ZTIZ3foovE1S,"aws",@progbits
        .align 8
        .type   DW.ref._ZTIZ3foovE1S#, @object
        .size   DW.ref._ZTIZ3foovE1S#, 8
DW.ref._ZTIZ3foovE1S:
        data8   _ZTIZ3foovE1S#

Note that the type_info object itself (_ZTIZ3foovE1S) is local.  What is not 
local is the indirect reference to it through DW.ref._ZTIZ3foovE1S.  So, 
while the 2 DW.ref._ZTIZ3foovE1S symbols are advertised as being identical, 
their contents would *not* be identical at link-time.

-- 
Eric Botcazou