On Wed, 2006-03-01 at 07:27 -0600, John Fultz wrote:
>
> I may not look at this again until the weekend. If anybody else wants to
> try to replicate this and go further, the non-wstring case should be pretty
> easy to set up given the info I've already provided. The wstring case
> requires a number of detailed steps which I won't go into right now.
Yay! It'd be cool if lots of people can try compiling and running
quackle in the non-wstring case (just look at the top of quackle/uv.h
and make sure USE_WSTRING_FOR_USER_VISIBLE is not defined). If it
crashes, if you can get a backtrace I'd be really appreciative.
Make sure you run quacker from within the quacker/ directory. By default
it looks for data files starting from ../data. I should change this to
be smarter in the future.
[temporal shift]
>
> Good news. I've gotten everything to compile and link under Windows with
> the wstring stuff in place (and also without...but I'm finding no
> difference between the two in terms of serious bugs).
I just put in a major fix for the nonwstring case so quackletest and
quacker actually look correct and deal with player names properly! so
quackle is just as usable with or without wstring enabled. Please
update.
>
> Right now, quackletest crashes immediately on startup...I've not
> investigated. The quackle interface itself seems to function (with minor,
> but very noticeable issues with the string handling of player names). The
> engine doesn't seem to be working. A simulation of a perfectly reasonable
> rack suggests about 10 different exchanges, and the computer never plays
> (probably for the same reason). And pressing the Commit button results in
> a crash.
A backtrace for crashes would be awesome if it's possible to get one.
And as I noted above make sure ../data points somewhere with the lexica
and stuff.
>
> Nothing investigated further cause it's way past time for me to go to bed.
>
> In summary, building with wstring support required a number of tweaks to
> the build system. Without wstring support required very few tweaks...most
> significantly, I had to replace some backslashes with forward slashes in
> various pathname specifications in the QT-generated makefiles. Odd, since
> I would have expected QT to generate makefiles that work out of the box.
> But it's not hard to do.
Ah that's odd.
>
> Aside from that, I have had to make three source code changes to get mingw
> to compile (the first two were only required to use STLport to get wstrings
> to work). They are...
okay I fixed these, thanks.
thankyou# thankyou#
Jason
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/quackle/
<*> To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
