I'm glad people have been looking *at the code I posted.*

On Thu, Aug 6, 2009 at 6:15 PM, Numbeast . <[email protected]> wrote:

> Quinn,
>
> Thank you for your welcome, I'm glad to have found such a cool open source
> game.
>
> Leo,
>
> Thank you for your encouragement and advice, as you can see forking worked
> fine and I was able to use my knowledge of git to figure out how to use
> mercurial. I'm glad people have been looking
>
> Martin,
>
> I've been wondering who wrote usl, so I could tell you how amazing I think
> it is. While it's obviously still in development I'm impressed at how much
> work has been put into it. I've glanced at inprepreter.cpp and I'm fairly
> sure that the return value will not be garbage collected but I'll test that,
> thanks for the heads up.
>
> I've also just pushed a few more changes into my public repo. The existing
> test has been changed to use Boost::Test and there's now a build target
> which compiles and runs all unit tests, "scons test=1". I still need to
> write something which checks that boost_unit_test_framework(-mt) is
> installed but it's not required unless you specify test=1.
>
> By the way I'm a beginner at scons, if anyone could look at my changes and
> tell me if I did anything wrong/offer suggestions on how to do it better
> please speak up. Specifically I would like to include any object files which
> are required by the unit tests, as opposed to manually adding every one to
> SConscript.
>
>
> On Thu, Aug 6, 2009 at 11:35 AM, Martin Voelkle 
> <[email protected]>wrote:
>
>> Hello Brian,
>>
>> On Wed, Aug 5, 2009 at 8:16 AM, Numbeast .<[email protected]> wrote:
>> > After a very brief conversation with giszmo3 I decided I would try
>> tackling
>> > adding unit testing to Glob2. I started with libusl but quickly ran into
>> a
>> > problem, USL provides no method of receiving the return value of a
>> script.
>> > The value is dumped to cout but not saved for later. In order to test
>> for
>> > the proper execution of scripts, it would be nice to know that a script
>> > returns (what the result of executing the last line is). I therefore
>> made
>> > this small (2 additional lines) patch for intrepreter.cpp and
>> intrepreter.h
>> > so that after USL step()s through a thread until it's state is
>> state:STOP it
>> > saves the return value. After asking on irc i was told that this would
>> be a
>> > good place to ask for this small diff to be committed (the diff is
>> > attached).
>>
>> It's great to see you playing with usl!
>> Your patch seems right, except that you have to make sure that the
>> return value is not garbage collected: in Thread::markForGC(), call
>> returnValue->markForGC() if returnValue != 0.
>> I invented the language and implemented it with Steph, feel free to
>> ask questions here, as we don't hang much on IRC.
>>
>> > On another subject, after a little thought I decided to use Boost:Test
>> as
>> > the testing framework. Your wiki says you don't want any extra
>> dependencies
>> > but in your mailing lists there have been a few comments to the effect
>> of if
>> > something makes programming simpler you'll accept it, and I believe
>> using
>> > Boost:Test is much easier and faster than implementing your own library.
>> You
>> > probably allready have it installed considering you use Boost:Thread.
>> Any
>> > feedback on this choice would be greatly appreciated.
>>
>> Since it's for unit testing, you can make it optional. The tests would
>> be compiled and run at build-time only if boost:test is available.
>>
>> Martin
>>
>>
>> _______________________________________________
>> glob2-devel mailing list
>> [email protected]
>> http://lists.nongnu.org/mailman/listinfo/glob2-devel
>>
>
>
_______________________________________________
glob2-devel mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/glob2-devel

Reply via email to