Boost Testing :
From: D. Kevin McGrath (dmcgrath_at_[hidden])
Date: 2005-11-22 12:03:09
Ok, I tried running the regressions with sun studio 11, but it fails
when trying to build process_jam_log. I'm attaching the output from
regression.py that pertains to process_jam_log, to give you an idea
of what is going wrong. I'll look into this later today.
On a good note though, bjam built without a hitch.
On Nov 21, 2005, at 10:04 PM, Beman Dawes wrote:
> "D. Kevin McGrath" <dmcgrath_at_[hidden]> wrote in message
>> I'm trying to get this running on the sparc platform (solaris 10),
>> but it's slow going so far. If anyone wants to take this off my
>> hands, feel free to do so. I've got the the compiler and the OS up
>> and running, but am having problems getting a usable python install,
>> which I believe is required for the regression tests, no?
>> The system I'll be running on is an older Ultra5 system, with the
>> goal of running regressions once a day. I'll post an update if/when
>> I get them running.
>> Unfortunately, I don't have access to the other platforms (linux x86/
>> x64, solaris x86/x64).
>> If I do get this up and running, what exactly would be required of
>> me? If you can lay out the responsibilities, that would be greatly
> The reason I brought it up was that it seems there is a disconnect
> the OSL4 tests, which have many, many failures, and the assertions
> of folks
> at Sun that they do a pretty good job of compiling Boost code.
> I'm reluctant to ask Doug Gregor (listed as the OSL4 tester) to dig
> this because he has put in so much work recently on Boost release
> that I hate to ask him for more effort. Doug has also done a
> dynamite job
> recently of hacking GCC to handle the Indiana C++ Concepts proposal,
> allowing the C++ committee to try out a complex language addition
> before it
> goes in the standard. He needs a break:-)
> So what I was hoping for was a volunteer to start testing the
> actual Sun
> latest release compiler, just like any other compiler. Fix any
> detected with the Boost toolset, report any problems with the compiler
> itself to Sun (ask me privately for the email address to do this),
> Boost developers to fix any problematic Boost code, and generally
> keep an
> eye on the tests until they settle down.
> I'm not a Sun user myself, but Sun is a big enough player in the
> business that I (and others) have wanted additional Sun test
> coverage for a
> long time now.
> Boost-Testing mailing list