[Bug-apl] Deal (?) reports assertion error rather than APL error message

[Frederick Pitts]

Hello Jürgen,

Should GNU APL report an APL error message, say "DOMAIN ERROR",
rather then an assertion diagnostic for `¯1 ? 5'? The reason I ask is
`3 ? ¯1' produces "DOMAIN ERROR".

Details follow:

***@centari.modelling.org[fred] $ apl -v
BUILDTAG:
---------
Project: GNU APL
Version / SVN: 1.7 / 1022M
Build Date: 2017-11-23 18:07:12 UTC
Build OS: Linux 4.13.13-300.fc27.x86_64 x86_64
config.status: 'RATIONAL_NUMBERS_WANTED=yes'
'CORE_COUNT_WANTED=syl'
Archive SVN: 989

***@centari.modelling.org[fred] $ apl
¯1 ? 5

=======================================================================
=======
Assertion failed: items
in Function: at
in file: Simple_string.hh:286

Call stack:

----------------------------------------
-- Stack trace at Simple_string.hh:286
----------------------------------------
0x7F40C62DD03A __libc_start_main
0x448C25 main
0x5BB01D Workspace::immediate_execution(bool)
0x49A424 Command::process_line()
0x49ABDD Command::do_APL_expression(UCS_string&)
0x49A48F Command::finish_context()
0x4A5658 Executable::execute_body() const
0x557350 StateIndicator::run()
0x4E9494 Prefix::reduce_statements()
0x4E6BAD Prefix::reduce_END_B__()
0x557CAF StateIndicator::statement_result(Token&, bool)
0x5B6D85 Value::print(std::ostream&) const
0x50A7DD PrintBuffer::PrintBuffer(Value const&,
PrintContext const&, std::ostream*)
0x45B5E8 do_Assert(char const*, char const*, char const*,
int)
========================================

SI stack:

Depth: 0
Exec: 0xc3a1a0
Safe exec: 0
Pmode: ◊ ¯1 ? 5
PC: 4 (5) RETURN_STATS
Stat: ¯1 ? 5
err_code: 0x0


=======================================================================
=======


Regards,

Fred