[v3] more istream::ignore cleanups

[Benjamin Kosnik <bkoz at redhat dot com>]

Yo Paolo, happy new year.

I think you can be a bit more aggressive about the istream::ignore
cleanups. Let me know what you think of the following patch.

Also, I am mystified by:

		  if (__n == numeric_limits<streamsize>::max()
		      && !traits_type::eq_int_type(__c, __eof)
		      && !traits_type::eq_int_type(__c, __delim))
		    _M_gcount = numeric_limits<streamsize>::min();

this resetting of _M_gcount. Any chance you can explain?

tested x86/linux

best,
benjamin

Attachment: p.20050103-2
Description: Text document