|
|
Boost Testing : |
From: Markus Schöpflin (markus.schoepflin_at_[hidden])
Date: 2005-05-10 08:21:50
John Maddock wrote:
> OK, back to square one, first off here's my test case that illustrates
> the problem:
<snip>
> You should get a compiler error if you try and compile this (tru64 only,
> it's perfectly legal code).
Here is what I get when doing this:
> cxx -version V6.5-042 -model ansi -D__USE_STD_IOSTREAM -I../../..
foo1.cpp && ./a.out
cxx: Error: foo1.cpp, line 24: a namespace name is not allowed
std::cout << "Checking case conversion for regex_replace formatting." <<
std:
---------------------------------------------------------------------------^
cxx: Error: foo1.cpp, line 24: expected a ";"
std::cout << "Checking case conversion for regex_replace formatting." <<
std:
------------------------------------------------------------------------------^
cxx: Error: foo1.cpp, line 31: specific definition of inline template
function must precede its first use
char global_lower(char c)
-----^
cxx: Error: foo1.cpp, line 41: specific definition of inline template
function must precede its first use
char global_upper(char c)
-----^
cxx: Info: 4 errors detected in the compilation of "foo1.cpp".
> Now here's the test case that should reproduce the regression failure,
> it should pass with current cvs: does for me with:
<snip>
Here is what I get when doing this:
> cxx -version V6.5-042 -model ansi -D__USE_STD_IOSTREAM -I../../..
foo2.cpp && ./a.out
cxx: Warning: ../../../libs/regex/src/regex_traits_defaults.cpp, line 350:
nested comment is not allowed
regex_constants::escape_type_identity, /*/*/
------------------------------------------------------^
cxx: Warning: ../../../libs/regex/src/regex_traits_defaults.cpp, line 545:
nested comment is not allowed
regex_constants::syntax_char, /*/*/
---------------------------------------------^
> Finally, I've looked into my attempts to build the regression test
> program, and note that as well as a linker out of memory problem, some
> of the test source files did not build (compiler complains about
> std::collate), can you check the log to verify that this is not a
> problem for you?
When building the regression library tests manually, I don't get a single
error from the compiler or linker.
HTH, Markus