Hi! Unfortunately not. You can use either gcov or BullseyeCoverage to measure coverage. With Kind Regards, Ovanes On Wed, Jun 9, 2010 at 2:40 PM, wrote:

Does Boost.Test have any way to report coverage statistics? I have a requirement to provide coverage statistics for the code I am testing. Thanks, Herb Miller

[Please do not mail me a copy of your followup] boost-users@lists.boost.org spake the secret code <20100609080130.21006u389wm3wotm@webmail.hiwaay.net> thusly:

I am testing class A. However, class A makes some calls to class B. Does Boost.Test provide any "test harness" capability so that the responses from class B can be canned. I really only want to test class A, not A and B together. I hope the question makes sense.

You need to decouple the classes somehow. The easiest way is to use an interface to separate them, in C++ terms a pure virtual base class for B. In my 5-part tutorial on using Boost.Test, you can see the use of an interface to decouple a class from its collaborator in part 4. http://legalizeadulthood.wordpress.com/2009/07/05/c-unit-tests-with-boost-te... There are lots of ways to decouple two collaborators, but I like "Extract Interface" the best. If for some reason your code is too performance sensitive for an interface sitting in the middle, you can decouple the collaborators with templates. I give an example of how to decouple your code from the Win32 API using templates here: http://legalizeadulthood.wordpress.com/2007/12/17/faking-out-the-win32-api-w... The book "Working Effectively With Legacy Code" provides lots of good ideas on how to decouple your system under test from its dependencies. I wrote a review of it here: http://legalizeadulthood.wordpress.com/2007/04/11/working-effectively-with-l... -- "The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download http://legalizeadulthood.wordpress.com/the-direct3d-graphics-pipeline/ Legalize Adulthood! http://legalizeadulthood.wordpress.com

I haven't seen that, but I have seen a '\0' passed into an assertion ending up in the XML as-is. I don't even know if that's valid XML or not, but Hudson's XUnit plugin cannot handle it. Of course I could get around it, but would it be reasonable to request that the default behavior of that formatter be to show the hex codepoint or something rather than the character itself for those characters that are iscntrl or some such? On Wed, Jun 9, 2010 at 2:26 PM, Gennadiy Rozental wrote:

Gaurav N Shah writes:

...

Hi,

I have written my own struct to customize logging from struct "boost::unit_test::output::xml_log_formatter"

However, the XML file generated by log Formatter is Invalid. it adds invalid node end = "" , by the end of each test case logging.

Is it possible your formatter does this?

How your struct customizes xml_log_formatter? It does not really have any means of customization. Did you inherit from it?

Gennadiy

_______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users