ID:2180072
 
This was a sort of weird development week, where I was looking at a number of things at once along with a few bug reports. Today's new release 511.1365 addresses a few bugs, one of them quite an old one that turned out to be a very difficult find.

Why hasn't 511 been buttoned up as the stable build yet? Well to be honest, the main big hurdle I'm still facing is that I want a better way to handle systems with video card oddities, that don't behave properly when NVidia and Intel work together. As part of this, the new build adds a couple of exported variables that should tell AMD and NVidia cards to kick things into high-performance mode, but I'm still wondering if there's a better way to handle video card selection--preferably without adding to any GUIs.

512 has been on the radar also. A number of feature requests are still pending that I'd like to get into 512, and there are some new ones like raw strings (mainly for regular expressions) that don't use escape characters, which I think would be a good idea. This is what's on the List of features being considered:
  • Raw string format
  • Fix assignment chaining of a = b[x] = c for complex cases (the current fix only covers simple cases)
  • Look into proc/list . operator chaining to determine feasibility
  • Mouse position tracking overhaul
  • New method to produce a flat combined icon from an appearance
  • Filters
  • background-color in maptext
  • text-shadow and outline in maptext
  • Improved preload_rsc support
  • More backend Steam API support
  • Game-specific mute
  • Icon info routines
  • Run as guest from Dream Maker
  • libcurl?
Filters are actually fairly high on my own wish list, so I've been trying to figure out a good way of dealing with them. The trick here is, filters are really a complex topic. The simplest ones, like a blur, would have parameters that need to be managed, and more complex ones should be able to do composites, masks, etc. and specify output images. It's not an easy thing to figure out how to come up with a friendly syntax for that. And of course I would want filter params to be animatable somehow. The main reason filters didn't hit 511 was the syntax problem, which has not been resolved yet.

The operator chaining has been on everybody's list for a long time, and that's something I'm not sure about for a whole lot of reasons. Internally, the way DM parses something like a.b.c is special, and I'm not sure how a.b.c().d would fit into that. But it's high time I did a feasibility study.

Christmas is just over three weeks away! Get your wish lists up to date and roll out some cookies. Cyber week is almost over, so be sure to get your shopping in. Thanks to everyone who donated to BYOND or became a Member last month, and for those who haven't yet, remember Santa Claus is watching.
Lol I want Peppermints
I'm so excited right now. q-q
It's a party at your house that's why you excited I am coming too
When you say video card oddities, do you mean the sort that I reported here a while ago?
na, the ones he addressed have to do with laptops that have a low power video card for normal desktop usage, and a high powered video card for 3d/games. There is a way you can tell these hybrid video card systems that your program should be treated as a game and use the higher powered system, he added this indicator to the client.
maptext improvements and filters are my two personal favorites on this list.

depending on what you do with the filters, I can see them becoming super useful in alot of games.
In response to IchiroKeisuke
IchiroKeisuke wrote:
maptext improvements and filters are my two personal favorites on this list.

Look into proc/list . operator chaining to determine timeframe for release.

I will wait until the end of time bet
Awesome, more steam api options would be lovely.
D3DCompileFromFile for game-defined shaders when?