From: Vladimir Prus (ghost_at_[hidden])
Date: 2004-05-12 02:15:37
Michael Stevens wrote:
> On Tuesday 11 May 2004 16:28, Vladimir Prus wrote:
> > The last version I've played with is attached. Usual disclaimers apply --
> > I have *some* V2 tests work with it, but otherwise it's completely
> > untested.
> > Of course, if you could share some expertise with the intel toolset, that
> > would be most appreciated.
> Thanks for mailing that. It got me working and I have merged it with my own
> experimental code. I have attached my version.
> It is a little more complete. I call it "intellinux.jam" which I find
> visually simpler when specifying properties. Does anyone have any strong
> preferences for naming?
> It add few useful features.
> An comprehensive init which allow path, compiler name and compile and link
> options to be specified in the user-config. These are all conditioned on
> the version which is therefore mandatory.
> A default compiler location. This should probably be extended to the new
> scheme that version 80 uses.
thanks a lot!
Let me start with some nit-picking ;-)
1. How the version is to be specified? As "80" or as "8.0". At least all
other toolset use the latter variant.
2. My copy is installed at /opt/intel_cc_80/bin/icc and you toolset assumes
different path. Is that the "new" scheme you mention above?
3. The path passes to the 'init' rule is required to end in slash, while other
toolsets are happy either way.
4. Finally, I think the the first declared version should become default one.
And now the really good news. I've run all V2 tests and see only three
I don't know why it fails, but the test is a bit fragile and often fails on
Here, the compiler emits the following warning:
icc: Command line warning: ignoring option '-O'; no argument required
it happens only for "optimization=space", where the command line looks like:
/opt/intel_cc_80/bin/icc -c -xc++ -Os -g -Ob0 -c -o
library_order. Any ideas what to do?
This fails since now I have to specify "intellinux-80" as toolset when
invoking tests, and the test system does not like it. I suspect this means
that this newly added test is broken whenever you specify toolset with
version, and is going to look into it.
This means that intel toolset is ready to be added! We only need to decide
what to do with the glitches I've listed above.
> To make everything perfect gcc.jam should be modifed with
> flags gcc.link OPTIONS <debug-symbols>off : -Wl,--strip-all ;
> as icc does not understand '-s'.
I've comitted that change to gcc.jam.
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk