Re: [boost] Any sunpro experts out there?

From: Timo Geusch (timo_at_[hidden])
Date: 2006-06-15 01:58:58

John Maddock wrote:
> Timo Geusch wrote:
>   
>> Well, I'm not an sunpro "expert" but I use the compilers at work and
>> am
>> sad enough to have a Sun workstation running Sol10 and the latest
>> compiler here at home. I've been playing with boost on that for a
>> while
>> so I should be able to help.
>>
>> I'm just about to update the compiler on that box to the latest patch
>> level so I'll give it a try against 1.34 RC on the weekend. I've seen
>> the is_enum breakage on 1.33 and was planning to investigate shortly
>> before all my spare time disappeared, but the code made my head hurt a
>> tad...
>>     
>
> OK, if you could try the type_traits tests with the very latest sunpro 
> patches applied that would be great: Sun claim to have fixed this, so if the 
> latest patches do indeed make the problem go away, then that's an easy fix 
> :-)
>   

Finally managed an attempt at building 1.34 RC with the latest patch set 
for SUN Studio 11 and run the tests against type_traits. For both build 
& test I used bjam v1, not v2 as I'm not yet familiar with v2 and the 
sunpro-tools.jam file I'm using is a tad different from the one in the 
distribution. My jam file based on a bunch of patches that Simon 
Atanasyan published in his blog for 1.32 and generally achieves better 
build and test results than the one that comes with Boost as that one is 
missing a couple of important flags. I'll prepare a patch tonight and 
would hope that it'll make it into 1.34...

Anyway, the test results don't look that good - 11 failures and 15 
targets skipped because it failed to build a couple of tests 
(test_extent and rank_test failed to build, a quick trawl through the 
test output suggests that several of the is_<something>_test case fail 
as well, namely is_object_test, is_member_func_test and 
is_reference_test. I'll try to provide full output in the next couple of 
days.

Regards,
Timo