[boost] Re: (For me a) showstopper [was: boost thread static library]

From: Michael Glassford (glassfordm_at_[hidden])
Date: 2004-02-18 07:04:07

"Stefan Slapeta" <stefan_at_[hidden]> wrote in message
news:c0vg2l$so7$1_at_sea.gmane.org...
> Michael Glassford wrote:
> > It is no longer supported: the tss class requires Boost.Thread to
be
> > in a dll to be able to clean up after itself properly (i.e. to
prevent
> > leaks). It would be possible to build a version that doesn't
include
> > support for this, but in a future version of Boost.Thread
(currently
> > in the thread_dev branch of CVS), I believe several new features
> > depend on this support being available so it may not be worthwhile
to
> > do so.
> >
>
> There's one more big "surprise": the need for the shared runtime
> libraries! I'm very frustrated at the moment because I don't know
which
> strategy to choose for boost.thread at all:
>
> - IMO, it's not a 'little fish' to change the runtime libraries of
all
> projects that use Boost.Thread. This and the necessity to ship an
> additional boost.thread dll is, at least for me and our company, at
the
> moment an absolute showstopper for moving to a future release of
boost!
> It is not easy to explain the costs for deploying the shared runtime
> library on some thousand machines.

I'm aware of this cost as I have the same problem myself. As I
mentioned in another post, it is still possible, if you are aware of
the issues involved and set up an appropriate build yourself, to use
Boost.Thread in a static library.

For what it's worth, I'm hoping to investigate what it would take to
make Boost.Thread support a static library option again in the future.
If such a thing ever happened (no promises), it would have to be a
version in which features that require the dll are disabled.

> Personally, I'm very disappointed as it seems to be possible in
boost,
> that a whole library becomes unusable for me in a future release.
[Not
> the best example of backward compatibility!]
>
> - There is NO WORD about that in the whole documentation.

True, and this should be fixed.

> If there is one, you hid it very well!

I'm not the one who made the change, only the messenger who's telling
you about them.

> What is even worse: there is no word about
> ANY CHANGES in boost thread in the release notes!

That's because this change is quite old. If I remember correctly, it
was made in the 1.30 release.

>
> Sorry for this critisism. I know about all your efforts and
appreciate
> them very much. This time it seems that the need for new
functionality
> broke too much of existing code for my taste!
>
>
> br, Stefan
>
> _______________________________________________
> Unsubscribe & other changes:
http://lists.boost.org/mailman/listinfo.cgi/boost
>