Re: [boost] [1.34] [showstopper] autolink doesn't work when not specifyingtoolset version
Hi, If this is a showstopper we better figure out what we do about it. My question is who is doing what and when? Any help with this issue will be greatly appreciated. Thanks Thomas On May 7, 2007, at 12:37 PM, David Abrahams wrote:
The following message is a courtesy copy of an article that has been posted to gmane.comp.lib.boost.user as well.
on Sun May 06 2007, "John Maddock" <john-AT-johnmaddock.co.uk> wrote:
I don't think there's anything that autolinking can do to fix this: it requires that libraries are named consistently, and there's no way around that.
Yes, that's not what I meant to imply.
So we either need to update the docs, or preferably fix Boost.Build to detect the MSVC version. As Dave says it does this in other contexts already, so it should be possible.
Exactly.
BTW, this is a showstopper for 1.34 release I believe.
Absolutely.
-- Dave Abrahams Boost Consulting http://www.boost-consulting.com
Don't Miss BoostCon 2007! ==> http://www.boostcon.com
-- Thomas Witt witt@acm.org
on Mon May 07 2007, Thomas Witt <witt-AT-acm.org> wrote:
Hi,
If this is a showstopper we better figure out what we do about it. My question is who is doing what and when?
I think Volodya should fix it, since only he understands most of the code in msvc.jam, and, I guess, right away, since we want to ship 1.34 -- Dave Abrahams Boost Consulting http://www.boost-consulting.com Don't Miss BoostCon 2007! ==> http://www.boostcon.com
on Tue May 08 2007, David Abrahams <dave-AT-boost-consulting.com> wrote:
on Mon May 07 2007, Thomas Witt <witt-AT-acm.org> wrote:
Hi,
If this is a showstopper we better figure out what we do about it. My question is who is doing what and when?
I think Volodya should fix it, since only he understands most of the code in msvc.jam, and, I guess, right away, since we want to ship 1.34
I "fixed" the problem by inserting a caveat in the documentation. -- Dave Abrahams Boost Consulting http://www.boost-consulting.com Don't Miss BoostCon 2007! ==> http://www.boostcon.com
On Tuesday 08 May 2007 03:49, Thomas Witt wrote:
Hi,
If this is a showstopper
I suppose that you, as release manager, get to to decide of this "if". Calling "showstopper" something that does not prevent getting working binaries is a bit liberal as far as I'm concerned.
we better figure out what we do about it.
Apparently, no concrete proposal was made.
My question is who is doing what and when?
Assuming the concrete proposal is: Whenever no explicit version of msvc is provided, run compiler, extract version from the output, and use that. and I'm asked to implement that in code, it would require access to all versions of msvc for which that behaviour is desired, plus a couple of days. - Volodya
Vladimir Prus wrote:
Assuming the concrete proposal is:
Whenever no explicit version of msvc is provided, run compiler, extract version from the output, and use that.
and I'm asked to implement that in code, it would require access to all versions of msvc for which that behaviour is desired, plus a couple of days.
Personally, I'd rather we take a few days to fix this, rather than ship "as is". Vladimir: if there's anything I can do to aid testing of any changes then let me know. I believe that if we have "using msvc ;" in user-config.jam then it automatically detects the available versions and sets the most recent as the default? Can we rig it so that --toolset=msvc has the same effect? The only issue is what happens if the user invokes bjam from within an msvc command prompt. Should we expect --toolset=msvc to invoke the compiler that's currently in the path and detect it's version correctly? My gut feeling is this may be too much to ask for, but if so we should say so in the docs (Dave?). Regards, John.
John Maddock wrote:
Vladimir Prus wrote:
Assuming the concrete proposal is:
Whenever no explicit version of msvc is provided, run compiler, extract version from the output, and use that.
and I'm asked to implement that in code, it would require access to all versions of msvc for which that behaviour is desired, plus a couple of days.
Personally, I'd rather we take a few days to fix this, rather than ship "as is".
Why didn't this show up earlier ? What prevents any such fix to introduce a new series of regressions, delaying the release even further ? This is really depressing... :-( Regards, Stefan -- ...ich hab' noch einen Koffer in Berlin...
on Tue May 08 2007, Vladimir Prus <ghost-AT-cs.msu.su> wrote:
On Tuesday 08 May 2007 03:49, Thomas Witt wrote:
Hi,
If this is a showstopper
I suppose that you, as release manager, get to to decide of this "if". Calling "showstopper" something that does not prevent getting working binaries is a bit liberal as far as I'm concerned.
we better figure out what we do about it.
Apparently, no concrete proposal was made.
My question is who is doing what and when?
Assuming the concrete proposal is:
Whenever no explicit version of msvc is provided, run compiler, extract version from the output, and use that.
and I'm asked to implement that in code, it would require access to all versions of msvc for which that behaviour is desired, plus a couple of days.
I can easily give you access to everything but VC-7.0, and with some work, I can give you access to vc-7.0 too. -- Dave Abrahams Boost Consulting http://www.boost-consulting.com Don't Miss BoostCon 2007! ==> http://www.boostcon.com
On May 7, 2007, at 9:46 PM, Vladimir Prus wrote:
Assuming the concrete proposal is:
Whenever no explicit version of msvc is provided, run compiler, extract version from the output, and use that.
and I'm asked to implement that in code, it would require access to all versions of msvc for which that behaviour is desired, plus a couple of days.
I think it's too late in the game for extensive surgery. Dave can we get a quick doc fix for this, even if the docs aren't at fault? Thanks Thomas -- Thomas Witt witt@acm.org
on Tue May 08 2007, Thomas Witt <witt-AT-acm.org> wrote:
On May 7, 2007, at 9:46 PM, Vladimir Prus wrote:
Assuming the concrete proposal is:
Whenever no explicit version of msvc is provided, run compiler, extract version from the output, and use that.
and I'm asked to implement that in code, it would require access to all versions of msvc for which that behaviour is desired, plus a couple of days.
I think it's too late in the game for extensive surgery. Dave can we get a quick doc fix for this, even if the docs aren't at fault?
What should it say? "if you're using any toolset other than gcc, be sure to include its version number explicitly?" P.S. Don't doc changes need testing, too? ;-) -- Dave Abrahams Boost Consulting http://www.boost-consulting.com Don't Miss BoostCon 2007! ==> http://www.boostcon.com
on Tue May 08 2007, Thomas Witt <witt-AT-acm.org> wrote:
On May 7, 2007, at 9:46 PM, Vladimir Prus wrote:
Assuming the concrete proposal is:
Whenever no explicit version of msvc is provided, run compiler, extract version from the output, and use that.
and I'm asked to implement that in code, it would require access to all versions of msvc for which that behaviour is desired, plus a couple of days.
I think it's too late in the game for extensive surgery. Dave can we get a quick doc fix for this, even if the docs aren't at fault?
Done and checked in. -- Dave Abrahams Boost Consulting http://www.boost-consulting.com Don't Miss BoostCon 2007! ==> http://www.boostcon.com
David Abrahams wrote:
on Tue May 08 2007, Thomas Witt <witt-AT-acm.org> wrote:
I think it's too late in the game for extensive surgery. Dave can we get a quick doc fix for this, even if the docs aren't at fault?
Done and checked in.
Thanks a lot! Thomas -- Thomas Witt witt@acm.org
Thomas, Is it too late to add a link from the docs to printer-friendly PDF's for the regex/type_traits/tr1/math libraries? Thanks, John.
participants (5)
-
David Abrahams -
John Maddock -
Stefan Seefeld -
Thomas Witt -
Vladimir Prus