Julian Foad wrote on Mon, Apr 23, 2012 at 17:18:22 +0100:
> I (Julian Foad) wrote:
>
> > There isn't currently an easy build switch (such as NDEBUG) to disable
> > SVN_ERR_ASSERT completely at compile time. That's just a side issue. If
> > you want such a switch, just ask; we can easily create one. Or if you
> > think we
> > need two levels of assertions -- one for quick tests and another for slow
> > tests
> > -- and want to be able to compile-out the slow ones independently of the
> > quick
> > ones, just ask. But implying we should use 'assert' for slow tests and
> > 'SVN_ERR_ASSERT' for quick tests is the Wrong Way.
>
> We can also introduce run-time control of whether the conditions are
> evaluated: test a global 'assertions enabled?' variable or function before
> evaluating the condition. For example:
>
> Index: subversion/include/svn_error.h
> ===================================================================
> #define SVN_ERR_ASSERT(expr) \
> do { \
> - if (!(expr)) \
> + if (svn_error__assertions_enabled && !(expr)) \
> SVN_ERR(svn_error__malfunction(TRUE, __FILE__, __LINE__, #expr)); \
> } while (0)
>
> +/* ... */
> +extern svn_boolean_t svn_error__assertions_enabled;
>
That doesn't sound right. Surely we don't want to allow disabling _all_
uses of SVN_ERR_ASSERT() this way? (Remember that some of them
translate to segfaults (possibly corruptions?) if the condition doesn't
hold)
Perhaps we should have two flavours of the SVN_ERR_ASSERT() macros --
one unconditional and one only for developers.
Daniel
> - Julian
