On Jan 29, 2010, at 3:16 AM, Joel Fielder wrote:
> Yeah, I guess. Regex is TR1 though so it's a standard - was Vlad's
> alternative TR1 based?
No, v2 was not: 100% STL only.
- I switched from accepting true regexes for test names to simple
*.?-wildcards. For wildcards I added a non-recursive matcher. I suppose I could
have coded regex engine from scratch, too, but decided to draw the line at
wildcards since they usually have enough expressive power for this particular
case.
- switched from shared_ptr's to manual delete's in destructors and such
- for boost::program_options, I naively thought I'd spare the future saps like
myself the trouble and clone the functionality that allows easy expanding the
set of cmd line options and auto-generates usage string:
utils::OptParser options ("usage: " + utils::file_basename (std::string
(av [0])) + " [options],\nwhere options are:");
options.add_options ()
("suite", "s", "test suite to run (default: 'default')")
("include", "i", "(*,?)-filter for tests to include (default:
all)")
("exclude", "x", "(*,?)-filter for tests to exclude (default:
none)")
("maxtesttime", "t", "global test time constraint [ms] (default:
disabled)")
("help", "h", "get this usage information", true);
nvm = options.parse (ac, av);
if (nvm.count ("help") > 0)
{
std::cout << options << std::endl;
return 1;
}
- in my misguided generosity, the OptParser will accept any option prefix as
long as it's unambiguous (i.e., -inc, -incl, -include will all work) so I added
a small trie to support such lookups.
- added utils.h with replacements for boost::path::leaf() and similar
This weekend I'll try to find time and do the git version of the patch. After
that, I am heading out the door. What's the point of being on "dev" list of a
project that considers itself "100% complete" ? By definition, there will be
no dev.
As for "boost" vs "no boost": I routinely intersect with people who don' t know
boost and can't seem to overcome the energy barrier required to become
proficient. Yes, they code their own hashtables. Yes, those are buggy and slow.
I deal with it.
I don' t like this direction but C++ is firmly evolving towards a lot more
metaprogramming because that's how the language manages to remain current and
replicate what's available in others (notably, Java). I think eventually C++
will collapse under the weight of its own complexity. Until then, when I have
to use C++ I play with the cards that were dealt to me and boost is a massive
productivity multiplier in those situations.
Cheers,
Vlad
------------------------------------------------------------------------------
The Planet: dedicated and managed hosting, cloud storage, colocation
Stay online with enterprise data centers and the best network in the business
Choose flexible plans and management services without long-term contracts
Personal 24x7 support from experience hosting pros just a phone call away.
http://p.sf.net/sfu/theplanet-com
_______________________________________________
unittest-cpp-devel mailing list
unittest-cpp-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/unittest-cpp-devel
