Fleury
|
|
« Reply #80 on: April 05, 2013, 11:13:44 AM » |
|
Are you using immediate mode? Of course; it's not common to store scene models in a 2D game (especially when they are sprite based). A good example of this is Direct2D, which is an immediate mode graphics API. You likey meant "why glBegin/glEnd instead of glDrawArray or glDrawElements" (note: both are still immediate mode). It's been on my TODO list for a while, and I do use glDrawElements on iOS (wherein I compose my entire scene in a single call), I'll likely get around to it at some point (since it's likely to help performance on systems lacking a prudent OpenGL driver).
|
|
« Last Edit: April 05, 2013, 11:30:19 AM by Fleury »
|
Logged
|
|
|
|
Fleury
|
|
« Reply #81 on: April 10, 2013, 10:07:19 PM » |
|
Oops, I seem have to have been forgetting about TigSource. So here's an update: I've written an asynchronous wrapper over my A* pathfinding implementation so that long paths an be computed in a background thread. As a result, a lengthy auto-walk from one side of the map to the other side is entirely feasible in the game-client (and without stalling the UI thread while the path is computed).
|
|
|
Logged
|
|
|
|
|
|
Fleury
|
|
« Reply #84 on: May 23, 2013, 11:47:50 PM » |
|
|
|
|
Logged
|
|
|
|
Fleury
|
|
« Reply #85 on: May 25, 2013, 01:17:52 AM » |
|
|
|
|
Logged
|
|
|
|
devi ever
Guest
|
|
« Reply #86 on: May 25, 2013, 01:50:33 AM » |
|
Just wanted to drop in and say that this project looks really fantastic! Keep up the great work.
|
|
|
Logged
|
|
|
|
HyMyNameIsMatt
Level 1
Vernal Edge
|
|
« Reply #87 on: May 25, 2013, 02:43:52 PM » |
|
Who does your enviornmental art and tilesets?
|
|
« Last Edit: May 25, 2013, 02:52:38 PM by HyMyNameIsMatt »
|
Logged
|
|
|
|
Fleury
|
|
« Reply #88 on: May 26, 2013, 06:46:35 AM » |
|
Back in 2006 I failed to get CJK text support working across all platforms due to an issue with Mono Framework's Cairo implementation of GDI ignoring many GDI properties: InterpolationMode, SmoothingMode, and TextRenderingHint. I've revisited this issue lately, and came up with a workaround: rendering the characters in advance (so that they appear identical on all platforms) and storing (all 25,000+ of) them in RAM (but not VRAM) for quick random retrievals. @HyMyNameIsMatt There's been a few artists that have done contract work for me since starting this project. If you're looking to pay an artist to draw you pixel art, I can forward a message of yours to the one I'm presently working with.
|
|
« Last Edit: May 26, 2013, 06:52:58 AM by Fleury »
|
Logged
|
|
|
|
|
|
jeremic
Level 0
|
|
« Reply #91 on: May 28, 2013, 07:09:19 AM » |
|
Wouldn't this be build 240, according to the bottom right corner of the game? Nice game you've made over the years. It runs terribly slow on my laptop, but it is 5 years old so its maybe not surprising.
Do you go by any other alias Fleury? I don't seem to recall your nick from the old OT or Tibia development forums?
|
|
« Last Edit: May 28, 2013, 07:15:33 AM by jeremic »
|
Logged
|
|
|
|
Fleury
|
|
« Reply #92 on: May 31, 2013, 10:09:08 AM » |
|
May 20, 2013 - Build 237 Released - The new world map has been applied, but it is largely incomplete. Further patches will the new quests and monsters, and tweak various zones.
- Fixed a missing packet-size check on the server.
- Giant Scarabs have a new sideways walking animation.
- Bows and Crossbows (as-well as their respective ammunition) now have the same attack values.
- Boats, Phalanxes, Red Potions, Blue Potions, Purple Potions, and Scarabs now have a new sprite.
@jeremic Does it have Intel GMA integrated graphics? Intel GMA has notoriously bad OpenGL support. I may end up writing a DirectX renderer for laptop users, since I've read roughly 70% of laptops use Intel GMA.
|
|
|
Logged
|
|
|
|
jeremic
Level 0
|
|
« Reply #93 on: May 31, 2013, 10:15:58 AM » |
|
Yes, it indeed does. This must be the problem then. Greets from Nudan
|
|
|
Logged
|
|
|
|
|
sunnder
|
|
« Reply #95 on: June 06, 2013, 03:26:14 PM » |
|
Didn't this game used to have a different name years ago? Something with "Frost" or... yeah. I can't believe it's still in development! Very impressive dedication that you have. Nice work, man!
|
|
|
Logged
|
"In any case you mustn't confuse a single failure with a final defeat." - F. Scott Fitzgerald
|
|
|
Fleury
|
|
« Reply #96 on: June 08, 2013, 07:37:11 AM » |
|
After much pondering what Pixel Champions would be like with a more conventional MMO UI (floating panels), I decided to just spend a few days implementing one. Hacking away at the code to allow the various UI elements to float around was pretty straightforward, but I got stuck trying to figure out how to handle UI scaling. WoW, FlyFF, and all the popular MMOs allow the UI to be scaled so that panels look roughly the same size at 800x600 as they do at 1980x1200. But they arn't scaling pixel art. @sunnder Yah, "Frostwinds", and thanks.
|
|
|
Logged
|
|
|
|
Fleury
|
|
« Reply #97 on: June 20, 2013, 12:07:26 AM » |
|
After alot of research into the Intel GMA GPU issue, it would appear there's an easy fix I can implement: default to BGRA for all GPUs.
Just about every OpenGL tutorial I've seen recommends glTexImage2D/glTexSubImage2D be given GL_UNSIGNED_BYTE combined with either GL_RGBA or GL_RGB for uploading bitmap data to VRAM. However, some rather common Intel GMA cards will perform dreadfully slow unless the parameters used are exactly GL_UNSIGNED_INT_8_8_8_8_REV with GL_BGRA. This I've known for quite a while.
However, what I did not know was that I could default to using Intel GMA's preferred format (BGRA), and not negatively affect performance on AMD and nVIDIA GPUs; there are some notable OpenGL applications and software libraries that defaulted to BGRA for that reason.
So Build-242 will use BGRA, and that should make the game run much better on most laptops.
|
|
« Last Edit: June 20, 2013, 12:14:15 AM by Fleury »
|
Logged
|
|
|
|
Fleury
|
|
« Reply #98 on: July 05, 2013, 12:49:07 PM » |
|
More OpenGL "gotchas". Apparently ATI Radeon HD 5800 Series graphic cards have a bug in their driver software wherein glIsEnabled(GL_SCISSOR_TEST) is hard-coded to always return true, even to this day (with the latest driver version).
I've managed to implement a workaround though. Those affected by this should see it resolved in Build 244.
|
|
|
Logged
|
|
|
|
|
|