Re: RFC: GLib testing framework
- From: muppet <scott asofyet org>
- To: Gtk+ Developers <gtk-devel-list gnome org>
- Subject: Re: RFC: GLib testing framework
- Date: Fri, 9 Nov 2007 08:52:42 -0500
On Nov 9, 2007, at 4:16 AM, Mikkel Kamstrup Erlandsen wrote:
That is certainly another possibility. However a design goal of my
code was to only expose one public symbol per simple type and let
the rest of the public methods be implemented in macros wrapping
these symbols. I am not sure that is viable with your above solution.
These symbols would appear in a library used only for unit testing,
not user applications. I don't think having a full suite of symbols
for test matches in a speacial-purpose library is a bad thing.
And, besides, all of g_assert_<type>_<op>() symbols will be
preprocessor macros, so that they can interpolate __FILE__, __LINE__,
and __func__, anyway. And i think you want to write at least some of
them by hand in order to tweak the messages that they produce.
Doing a good job around here is like wetting your pants in a dark
suit; you get a warm feeling, but no one notices.
] [Thread Prev