[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]
I Just Talked to Stroustrup (was "Third Party Suicide?")
- To: <tribble>
- Subject: I Just Talked to Stroustrup (was "Third Party Suicide?")
- From: Mark S. Miller <mark>
- Date: Mon, 20 Nov 89 20:30:47 PST
- Cc: <marcs>, <xanadu>, <xtech>
- In-reply-to: <Eric>,40 PST <8911202212.AA25578@xanadu>
Date: Mon, 20 Nov 89 14:12:40 PST
From: tribble (Eric Dean Tribble)
Will they consider implementations of 'not implemented' features as
well as bug fixes? Do we need any of the not-implemented features?
(I assume so).
All the issues we need for garbage collection are documented
"unimplemented"s. We know of only one compiler bug from AT&T, and
that one doesn't affect garbage collection.
AT&T's stand is that 2.1 will try to fix the bugs they know about (I
will make sure they know of the one we found (pureVirtualBug)), but
2.1 will not fix *any* of the documented "unimplemented" cases.
Unless we provide them with a fix--then it seems quite possible (no
guarantee of course).
They also mentioned that the current compilers (of which 2.1 will be
an instance) were not meant to be production quality software
originally, but rather research vehicles. (Even without this, I would
say our having only hit one bug is very impressive. Given this, I'd
say it's fantastic!) They are working on a re-written compiler which
they guess would be out at the end of 1990. As Heh points out, that
probably means in the next 2 or 3 years.