Re: An ICE that does not generate a full report

[Alice Osako]

    Oh, and note that the
      constants in CardMath.mod are not used for their own
      values as such, but rather (AFAICT) are used to compute the number
      of bits in a CARDINAL on a given platform, and the
      maximum number of decimal digits it can represent. The ones in Size.def are used to determine the
        bitwidth for the largest scalar supported by the
        compiler/system.

Alice Osako:

>       
>       Gaius Mulley:
> > I'm now seeing it compile:
> >
> > $ gm2 -g -c BasicFileSys.mod -I../../../../Sandpit/gm2-outside-tree-tests/build/source/m2pp/src/ -fsources
> > Compiling: BasicFileSys.mod
> > Pass 0: lexical analysis, parsing, modules and associated filenames
> >    Module SYSTEM               : /home/gaius/opt/lib/gcc/x86_64-pc-linux-gnu/14.0.1/m2/m2cor/SYSTEM.def [m2cor]
> >    Module M2RTS                : /home/gaius/opt/lib/gcc/x86_64-pc-linux-gnu/14.0.1/m2/m2pim/M2RTS.def [m2pim]
> >    Module RTExceptions         : /home/gaius/opt/lib/gcc/x86_64-pc-linux-gnu/14.0.1/m2/m2pim/RTExceptions.def [m2pim]
> >    Module BasicFileSys         : ../../../../Sandpit/gm2-outside-tree-tests/build/source/m2pp/src/BasicFileSys.def
> >    Module BasicFileSys         : BasicFileSys.mod
> >    Module COROUTINES           : /home/gaius/opt/lib/gcc/x86_64-pc-linux-gnu/14.0.1/m2/m2pim/COROUTINES.def [m2pim] (for C)
> >    Module Size                 : ./Size.def
> >    Module stat                 : ./stat.def
> >    Module stdio                : ./stdio.def (for C)
> >    Module unistd               : ./unistd.def (for C)
> >    Module fcntl                : ./fcntl.def (for C)
> >    Module stat0                : ./stat0.def (for C)
> >    Module SysTypes             : ./SysTypes.def
> > Pass 1: scopes, enumerated types, imports and exports
> > Pass 2: constants and types
> > Pass C: aggregate constants
> > Pass 3: quadruple generation
> > Pass 4: gcc tree generation
> >         symbols to gcc trees
> >         statements to gcc trees
> >         gcc trees given to the gcc backend
> >       
>       
>       Thank you!
>       
>       However, when I try to compile it, I get the following error:
>       
>       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>       gm2 -g -fiso -freport-bug
>         -fdebug-function-line-numbers -fsources -Isrc/ -c
>         src/imp/BasicFileSys.mod \
>         -o obj/BasicFileSys.o
>         Compiling: src/imp/BasicFileSys.mod
>         Pass 0: lexical analysis, parsing, modules and associated
>         filenames
>            Module SYSTEM               :
> /home/schol-r-lea/opt/lib/gcc/x86_64-pc-linux-gnu/14.0.1/m2/m2iso/SYSTEM.def
>         [m2iso]
>            Module M2RTS                :
> /home/schol-r-lea/opt/lib/gcc/x86_64-pc-linux-gnu/14.0.1/m2/m2iso/M2RTS.def
>         [m2iso]
>            Module RTExceptions         :
> /home/schol-r-lea/opt/lib/gcc/x86_64-pc-linux-gnu/14.0.1/m2/m2pim/RTExceptions.def
>         [m2pim]
>            Module BasicFileSys         : src/BasicFileSys.def
>            Module BasicFileSys         : src/imp/BasicFileSys.mod
>            Module Size                 : src/Size.def
>            Module stat                 : src/stat.def
>            Module stdio                : src/stdio.def (for C)
>            Module unistd               : src/unistd.def (for C)
>            Module fcntl                : src/fcntl.def (for C)
>            Module stat0                : src/stat0.def (for C)
>            Module SysTypes             : src/SysTypes.def
>         Pass 1: scopes, enumerated types, imports and exports
>         Pass 2: constants and types
>         Pass C: aggregate constants
>         Pass 3: quadruple generation
>         Pass 4: gcc tree generation
>                 symbols to gcc trees
>         cc1gm2: error: In definition module ‘BasicFileSys’: constant
>         ‘BW8’ should not be reassigned
>         cc1gm2: error: constant ‘BW128’ should not be reassigned
>         cc1gm2: error: constant ‘BW120’ should not be reassigned
>         cc1gm2: error: constant ‘BW112’ should not be reassigned
>         cc1gm2: error: constant ‘BW104’ should not be reassigned
>         cc1gm2: error: constant ‘BW96’ should not be reassigned
>         cc1gm2: error: constant ‘BW88’ should not be reassigned
>         cc1gm2: error: constant ‘BW80’ should not be reassigned
>         cc1gm2: error: constant ‘BW72’ should not be reassigned
>         cc1gm2: error: constant ‘BW64’ should not be reassigned
>         cc1gm2: error: constant ‘BW56’ should not be reassigned
>         cc1gm2: error: constant ‘BW48’ should not be reassigned
>         cc1gm2: error: constant ‘BW40’ should not be reassigned
>         cc1gm2: error: constant ‘BW32’ should not be reassigned
>         cc1gm2: error: constant ‘BW24’ should not be reassigned
>         cc1gm2: error: constant ‘BW16’ should not be reassigned
>       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>       
>       The constants in question are all defined in Size.def, and are not redefined or even
>       imported in either BasicFileSys.def
>       or BasicFileSys.mod; however, they
>       are defined slightly differently in CardMath.mod,
>       which also gives the same error even though it doesn't import Size.def. 
>       
>       I checked that the -fiso option was
>       not the deciding factor in this, and it wasn't. I'm not sure why
>       it is behaving differently on my system.
>       
>