Re: VS 2005

Tech-Archive recommends: Repair Windows Errors & Optimize Windows Performance

From: Hendrik Schober (SpamTrap_at_gmx.de)
Date: 01/14/05


Date: Fri, 14 Jan 2005 13:06:34 +0100

Ronald Laeremans [MSFT] <ronaldl@online.microsoft.com> wrote:
> >> So we always try to give users 1 release time
> >> to
> >> adjust to the changed behavior before making it the default in cases
> >> where
> >> that is an option and where the change of behavior is as big as in the 2
> >> /Zc
> >> options.
> >
> > I don't see the point in doing this. All
> > it does is making your customers write
> > even more lines that will give problems.
>
> It allows them to adopt the new version of the compiler and then decide on
> their won time to do the work to their code base to make the changes.

  This sounds good in theory. In reality,
  however, changes are introduced steadily
  with each new version.
  So when your customers adopt version X of
  the compiler there's a change that was
  already introduced in X-1, but which,
  since X-1 defaulted to the behaviour of
  X-2, they never notized. Now they either
  loose time by changing the code they had
  written between X-1 and X (atop of older
  code, that is) or they just have throw
  the compatibility switch.
  The first will most likely steal them
  time and prevents them from realizing
  that there are more changes introduced
  with X which are agin hidden by the
  compiler which uses the old behaviour by
  default. So they'll face the very same
  problem with the next version.
  The second they could have done just as
  well when X-1 was released.

  Sorry, but I still fail to see the
  advantage.

> Having
> the option, but having the default immediately work the other way should
> accomplish the same, but in our experience means that an inordinate number
> of customers get so scared by the initial build failures they just decide
> not to upgrade and many don't even do a before further investigation.

  So why not have a all-in-one switch for
  compatibility with X-1, and maybe even
  X-2? Whoever wants can throw that and go
  on compiling and hacking old-style code.
  If well advertized, everyone able to
  read should be able to do that. I mean,
  everyone using this product is dealing
  with a language that has abstract virtual
  base classes, class template partial
  specialization -- while missing function
  template partial specialization! --, and
  comes with exceptions, but no GC, except
  if they throw a switch -- by which they
  get a very different language!
  Heck, you could even bring in a wizard
  that sets up or imports old projects so
  that the code compiles.

> Ronald

  Schobi

-- 
SpamTrap@gmx.de is never read
I'm Schobi at suespammers dot org
"The presence of those seeking the truth is infinitely
 to be prefered to those thinking they've found it."
 Terry Pratchett


Relevant Pages

  • Re: SEARCH ALL
    ... > so that it DOES always return a consistent result when there are duplicate ... What criteria should the *compiler* apply unconditionally ... from customers who thought the performance of SEARCH ALL could benefit from ... The requested solution was "We want SEARCH ALL unconditionally to produce ...
    (comp.lang.cobol)
  • Re: And now my thoughts on Delphis survival
    ... as it is for many C++ Builder customers. ... large parts of the VCL go unused (data binding, web services, XML ... they need a mixed mode and Win64 compiler ASAP ... to remain competitive with VS. Compiler development is expensive and there ...
    (borland.public.delphi.non-technical)
  • Re: The 64bit future
    ... But it doesn't preserve compatibility with code that assumes long is ... do the users punish themselves? ... Say you're a compiler vendor, depending on customers giving you ...
    (comp.lang.c)
  • Re: Casting the return value of malloc() ?
    ... include build scripts which invoke a compiler for the correct language ... you _small_fortune_ for the code explicitly requiring that it is C code. ... Remember that neither customer will care about the other customers ...
    (comp.lang.c)
  • Re: Operator overloading in C
    ... AND, since it is not hidden in the compiler, you can CUSTOMIZE the ... You cannot both avoid burden on C implementations and get all the ... optimization that an assembly programmer would have missed. ... seat to additions the implementors' customers actually want. ...
    (comp.std.c)