On Wed, Oct 01, 2008 at 08:21:09PM -0400, Mike Caron wrote:
> Ralph Versteegen wrote:
> >2008/10/2 James Paige <[EMAIL PROTECTED]>:
> >>On Thu, Oct 02, 2008 at 12:28:05PM +1300, Ralph Versteegen wrote:
> >>>2008/10/2 Mike Caron <[EMAIL PROTECTED]>:
> >>>>James Paige wrote:
> >>>>>Eeep!
> >>>>>
> >>>>>1) I was about to cut the branch for Xocolatl, but now I am not so 
> >>>>>sure.
> >>>>>Do I cut from revision 2302? Do I delay another week or two for further
> >>>>>testing and more features?
> >>>>I've tested this thoroughly. Seriously. Let me add the other sprite 
> >>>>loading
> >>>>functions, then I'll call this feature complete.
> >>>>
> >>>>If you want, cut it at 2302, but I suspect a bunch of people would be
> >>>>annoyed if they learned what they just missed. ;)
> >>>>
> >>>>>2) Spiiiiiify. There is so cool! I am glad you went with a load/free
> >>>>>model. Much better than a fixed id model. I'll be retrofitting strings 
> >>>>>for a
> >>>>>load/free model soon after Xocolatl.
> >>>>I've always advocated this format. I mean, just look at the rest of the 
> >>>>code
> >>>>I write.
> >>>NO!!!!!
> >>>I want REAL strings as part of HamsterSpeak, without requiring manual
> >>>memory management. I was going to mention on the talk page for script
> >>>arrays that we should look at how we should improve strings first, and
> >>>see that array semantics would be much the same. (And therefore that I
> >>>don't like James' arrays suggestion either). Hint: it probably
> >>>involves adding either static or dynamic typing to the language.
> >>I do think real strings would be a tasty language feature. However, the
> >>current string implementation has to stick around to satiate the daemon
> >>beast backwardscompatelzebub. What I plan is to add a .used member to
> >>the PlotString type. It will default to ON for strings 0-31. Then I will
> >>add a "create string" command which will find an unused string (and
> >>make the plotstring array bigger if necessary) and I will add a "free
> >>string" command that would set .s="" and .used=NO
> >>
> >>So the difference from the current string system would be very small,
> >>alll old scriipts would work the same as usual, and I don't think any
> >>extra work will have been created for potential future implementation of
> >>better strings.
> >
> >But complicating strings further by having three different string
> >systems seems undesirable.
> >
> >>>>Well, I wasn't going to add an option at this point. However, there's no
> >>>>reason we can't add a "Z order", and then combine the two systems into 
> >>>>one.
> >>>>But, that's for another day.
> >>>>
> >>>I'm surprised that both aren't being added at once: I always
> >>>considered them the same thing.
> >>Me too, actually, which is why I hadn't tackled it sooner.
> >
> >So, right now there is no layering-control support at all? When you
> >strip away layering, grouping and attaching, it seems like this must
> >be quite diffiferent from what we want to end up with.
> 
> Not at all. Though I'm not sure what you mean by grouping, the other 
> stuff can be accomplished reasonably easily. Layering is a matter of 
> adding a layer member to the UDT, and calling draw_plotsprites more than 
> once. Attaching just requires two members in the UDT, and a bit of math 
> in draw_plotsprite.
> 
> All of which I was planning to do (and still am going to do), once I get 
> the base grunt work done.

Grouping will be a big deal. It means that layers will really behave 
like a tree rather than a list.

---
James
_______________________________________________
Ohrrpgce mailing list
[email protected]
http://lists.motherhamster.org/listinfo.cgi/ohrrpgce-motherhamster.org

Reply via email to