This was earlier covered by "?->", but this is inconsistent with the
"[?" index variant and the future "(?" program indexing variant.

Usage of "?->" will start emitting a deprication warning is some
future major version of Pike.

Manual backport from master: 19582df8,
because the conflict was too big to bother with cherry-pick.

It has accepted 8:s for quite some time, which is somewhat
confusing and could not possible be intentional.

Right?

'abcd' is now the same as ('a'<<24)|('b'<<16)|('c'<<8)|'d'.

this prevents left shifts of negative numbers, which is undefined

This allows one more syntax for int-ranges in types:

(Xbit): This reprensents the range (0..(1<<X)-1).

So, string(8bit) is identical to string(0..255) and string(16bit) is
string(0..65535).

The same is true for integers, int(3bit) is int(0..7).

This does not conflict with any existing code because identifiers can
not start with a number.

It has now been renamed to [?] since that does not conflict with any
other syntax in pike.

As an example:

int foo( array(int) a ) { return a[?10]; }

?-> can be used to index something that is either indexable or null.
?: is an alias for ||

Very useful for things like:

C c = a && a->b && a->b->c;

which can now be rewritten as

C c = a?->b?->c;

The code in language.yacc also supports ?. and ?[], but there are
syntax conflicts with ?: for those:

?. conflicts with expr?.Module:...
?[] conflicts wtih expr?[softcast]:...

?: was added mostly to be compatible with other 'modern' C-like
languages.

More pain than they are worth.

Rev: src/lexer.h:1.86

bignum.

Rev: src/lexer.h:1.85
Rev: src/stralloc.c:1.231
Rev: src/stralloc.h:1.110

Rev: src/lex.c:1.123
Rev: src/lexer.h:1.84

Rev: src/cpp.c:1.176
Rev: src/lex.c:1.122
Rev: src/lex.h:1.38
Rev: src/lexer.h:1.83
Rev: src/preprocessor.h:1.95

Rev: src/lexer.h:1.82

Rev: src/language.yacc:1.439
Rev: src/lexer.h:1.81

Rev: src/lexer.h:1.80
Rev: src/testsuite.in:1.832

Rev: src/lexer.h:1.79
Rev: src/testsuite.in:1.831

Rev: src/lexer.h:1.78