Digital Mars - c++ - MSVC++.NET RC3 Standards Compliance

↑ ↓ ← → Richard Grant <fractal clark.net> writes:

In article <b4bahe$2i05$1 digitaldaemon.com>, Mark Evans says...
According to this post, the final release is due in April.  It builds Boost.


MSVC had better build boost. If I had all the boost developers doing compiler
specific code in what seems like every lib, every 100 code lines or so - just
for MSVC - and it didn't compile, I would be *very* concerned.

Richard

Mar 08 2003

↑ ↓ ← → Garen <garen_nospam_ wsu.edu> writes:

Richard Grant wrote:

 
 MSVC had better build boost. If I had all the boost developers doing compiler
 specific code in what seems like every lib, every 100 code lines or so - just
 for MSVC - and it didn't compile, I would be *very* concerned.
 


They use it internally for testing.  The 7.1 compiler has been pretty much
done for quite awhile now too, it's only being released in april to 
coincide with other things in the "Everett" release (sadly). If you have
an MSDN account you can download a beta/release-candidate online.

Mar 09 2003

↑ ↓ ← → "Jeff Peil" <jpeil bigfoot.com> writes:

"Garen" <garen_nospam_ wsu.edu> wrote in message
news:b4go60$26s9$1 digitaldaemon.com
 They use it internally for testing.  The 7.1 compiler has been pretty
 much done for quite awhile now too, it's only being released in april
 to
 coincide with other things in the "Everett" release (sadly). If you
 have
 an MSDN account you can download a beta/release-candidate online.



There are differences between "pretty much done" and "done" the last
compiler we accepted into Everett was in mid febuary which is less than a
month ago.

-- 
Jeff Peil

Mar 10 2003

↑ ↓ ← → Garen <garen_nospam_ wsu.edu> writes:

Jeff Peil wrote:

 There are differences between "pretty much done" and "done" the last
 compiler we accepted into Everett was in mid febuary which is less than a
 month ago.


Whats your point?

Mar 10 2003

↑ ↓ ← → "Jeff Peil" <jpeil bigfoot.com> writes:

"Garen" <garen_nospam_ wsu.edu> wrote in message
news:b4jhqo$788$1 digitaldaemon.com
 Jeff Peil wrote:

 There are differences between "pretty much done" and "done" the last
 compiler we accepted into Everett was in mid febuary which is less
 than a month ago.


 Whats your point?



It appears I left out a word which might have clarified "last compiler we
accepted into Everett" should be "last compiler fix we accepted into
Everett"

You seemed to be implying that we were merely sitting on the compiler and
that it had been "pretty much done for quite awhile now"

The reality is that with the fixes in mid-Febuary (including a compiler
fix,) we had to do another release candidate.  RC3, which then requires
adequate testing time.

Believe it or not, we really wanted to be finished much earlier.

-- 
Jeff Peil

Mar 10 2003

↑ ↓ ← → Garen <garen_nospam_ wsu.edu> writes:

Jeff Peil wrote:

 It appears I left out a word which might have clarified "last compiler we
 accepted into Everett" should be "last compiler fix we accepted into
 Everett"
 
 You seemed to be implying that we were merely sitting on the compiler and
 that it had been "pretty much done for quite awhile now"
 


Yeah, thats the sentiment I've run into several times from MSers -- that
it's release was being deferred to be sync'd with a whole group of other
products. 

Is that not so?

 The reality is that with the fixes in mid-Febuary (including a compiler
 fix,) we had to do another release candidate.  RC3, which then requires
 adequate testing time.
 
 Believe it or not, we really wanted to be finished much earlier.


Me too. I really wanted 7.1 about 4+ years ago. :-)

Mar 10 2003

↑ ↓ ← → "Jeff Peil" <jpeil bigfoot.com> writes:

"Garen" <garen_nospam_ wsu.edu> wrote in message
news:b4k13r$g0i$1 digitaldaemon.com
 Yeah, thats the sentiment I've run into several times from MSers --
 that it's release was being deferred to be sync'd with a whole group
 of other products.

 Is that not so?


Yes and no.  Yes in the sense that the feature work for the compiler has
been done for a rather long time, and even in the sense that there were
periods of time that I would have been confident in shipping the compiler
because enough time had passed without finding a serious issue while we were
looking, but at those same points in time I would not have been confident
shipping some other part of Visual Studio (such as some piece of the IDE.)
The fact is that syncing up everything in the Visual Studio box leads to
more bake time than would otherwise be necessary (and so that extra time
inevitably leads to the discovery of serious issues that really need to be
fixed which leads to more bake time, etc.)  In the end the extra time
required to put multiple languages/tools in the box really does lead to
higher quality releases.

-- 
Jeff Peil

Mar 11 2003

↑ ↓ ← → Garen <garen_nospam_ wsu.edu> writes:

Jeff Peil wrote:

 Yes and no.  Yes in the sense that the feature work for the compiler has
 been done for a rather long time, and even in the sense that there were
 periods of time that I would have been confident in shipping the compiler
 because enough time had passed without finding a serious issue while we were
 looking, but at those same points in time I would not have been confident
 shipping some other part of Visual Studio (such as some piece of the IDE.)
 The fact is that syncing up everything in the Visual Studio box leads to
 more bake time than would otherwise be necessary (and so that extra time
 inevitably leads to the discovery of serious issues that really need to be
 fixed which leads to more bake time, etc.)  In the end the extra time
 required to put multiple languages/tools in the box really does lead to
 higher quality releases.
 


Where's the 'no' part? :-)  Seems I wasn't misleading by your description. 
Except that the 'syncing' part isn't quite so artficial; as it depends on 
other interrelated stuff.  I think thats what you were trying to get at 
anyway.

Mar 11 2003

↑ ↓ ← → "Jeff Peil" <jpeil bigfoot.com> writes:

"Garen" <garen_nospam_ wsu.edu> wrote in message
news:b4mc2k$1147$1 digitaldaemon.com
 Jeff Peil wrote:

 Where's the 'no' part? :-)  Seems I wasn't misleading by your
 description. Except that the 'syncing' part isn't quite so artficial;
 as it depends on other interrelated stuff.  I think thats what you
 were trying to get at anyway.


The no part is that I wouldn't really consider things like the ide, or
libraries like the crt, stl, mfc, or atl as seperate products.  For example,
if the crt isn't in a shape where it can ship, then the compiler isn't
either (particularly in the crt case since the compiler itself uses the
crt.)

-- 
Jeff Peil

Mar 12 2003

↑ ↓ ← → Garen <garen_nospam_ wsu.edu> writes:

Jeff Peil wrote:

 The no part is that I wouldn't really consider things like the ide, or
 libraries like the crt, stl, mfc, or atl as seperate products.  For example,
 if the crt isn't in a shape where it can ship, then the compiler isn't
 either (particularly in the crt case since the compiler itself uses the
 crt.)


Ok, well I follow that. Seems we were just talking in circles over 
differing interpretatons of terms.  In any case, looking forward to
the next release.

Mar 12 2003