V39c up! [MONKEY NEWS]
Monkey Forums/Monkey Programming/V39c up! [MONKEY NEWS]
| ||
| Hi, New version 39c is now available - mainly just more tweaks to hopfully get us back to v38! ***** V39c ***** Tools: MServer now uses 'etags' for detecting modified files. Trans (1.12): Fixed parenthesis mismatch error with multiline expressions. Trans compiler back to debug build for internal err stack dump. Fixed extern overriding extern causing internal err bug. Added radix 10 param to parseInt for JS/AS translators. Fixed method munging bug. |
| ||
| All working here - although not affected (with my project at least) by these issues. |
| ||
| Still using v29, i think i'll update to this one, next stop might be v49. |
| ||
I don't know if this is "by design" and expected, but with this version, you can still compile this:Interface MyInterface Method Draw:String() Method DoThis:int() end Class MyClass Extends List<String> Implements MyInterface End Function Main() Local A:=New MyClass End |
| ||
| Hi, This is the aggressive dead code elimination kicking in. Since you never invoke MyInterface.Draw, it doesn't really exist. Ditto DoThis. This is probably not ideal - and in general, I think a way to semant ALL decls would be useful. But I also think it's well worth keeping the code elimination as efficient as possible. This however wont compile: Interface MyInterface Method Draw:String() Method DoThis:Int() End Class MyClass Extends List<String> Implements MyInterface End Function Main() Local a:=New MyClass Local i:=MyInterface( a ) i.Draw End |
| ||
When testing HTML5 and Flash with the mserver, I have to kill the mserver before it picks up any changes or I get this:1:2> GET C:/BlitzMonkey/MonkeyPro39c/tmp/untitled1.build/flash/MonkeyGame.swf (200 OK) 1:2> GET C:/BlitzMonkey/MonkeyPro39c/tmp/untitled1.build/flash/MonkeyGame.html (304 Not Modified) 1:2> 400 Bad Request: HEAD /MonkeyGame.swf HTTP/1.1 1:2> BYE And of course I dont see my changes... |
| ||
| Hi, That looks correct to me - it indicates that only the swf has changed, not the html. Are you sure it's not working? Can you try another browser? |
| ||
| This is the aggressive dead code elimination kicking in. Ok, this make perfect sense, I just wanted to be sure there was not something unexpected. |
| ||
| Are you sure it's not working? For IE9 it doesnt work... Can you try another browser? For Chrome it works :) |
| ||
| ^^ Stunned silence =] |
| ||
| XNA, glfw, Flash, HTML5 and Andriod all seem to compile and run so far.:) I am using FireFox 4 (system default browser) and no problems with update/changes (see therevills post). L8r, |
| ||
| Works fine here with me projects :D (Firefox 4 :) ) |
| ||
| yes works well here too thanks |
| ||
| This is the aggressive dead code elimination kicking in. *facedesk* This is probably not ideal - and in general, I think a way to semant ALL decls would be useful. While I agree that the code elimination is a great idea (the Java compiler removes dead code too), it'd still be nice for the compiler to do a full scan over the entire source file (and imported modules) to make sure it's valid before choosing exactly what should be in the final binaries. Even if a full scan is a command line option (although I think it should be mandatory, just like Strict should be mandatory... -_- ) |
| ||
| Not sure if its some odd settings of mine, but the Rotation Sickness demo seems to crawl now with this particular update. |
| ||
I get (GLFW)build/glfw/xcode/../main.cpp:2666: error: cannot convert 'bb_gamestate_GameState*' to 'gc_object*' for argument '1' to 'void gc_mark_q(gc_object*)' GameState is an interface placed in a seperate module file. I'm sorry I don't have the time to try to reproduce. But maybe it rings a bell with someone? It seems like the garbage collector is trying to free an interface? Maybe interface does not inherit from gc_object? Update: Removing use of interfaces makes it compile again. Also I get a lot of wierd errors that seems to go away after some time. Maybe cache issues? |
| ||
| I think a way to semant ALL decls would be useful. The more I think about it, the more I think this is a very good idea. Specially if we want the chance to support reflection anytime. Additionally, The scenario when you make, as instance, a function call, and you get an error anywhere else in the code, just becouse it was there, but you were not using that code portion, is very weird. I can live with it, and I'm happy living with it, but if it could be improved, I would be even happier... :D |
| ||
| Not sure if its some odd settings of mine, but the Rotation Sickness demo seems to crawl now with this particular update. This demo crawls in HTML5, Plus every thing is colored green (images) in all the targets I have (change in setcolor?). I tested it in v36 (the demo source from that package) and runs fine in html5. I am looking over the source now but I have not seen a problem. I am using FireFox 4. Let ya know if I find more like this. |
| ||
| Moved from v38 to v39c today (skipped v39 & v39b) ... Current project complies for iOS, Android & GLFW (Mac) without incident ... No changes in performance compared to previous versions ... Have not tried any of the new features, but so far ... Thumbs Up! |
| ||
| This demo crawls in HTML5, Plus every thing is colored green (images) in all the targets I have (change in setcolor?). I tested it in v36 (the demo source from that package) and runs fine in html5. I am looking over the source now but I have not seen a problem. The problem is that it sets the colour to green to draw the ground rect and then doesn't reset it to 255,255,255 before drawing the images. |
| ||
| HI, > Additionally, The scenario when you make, as instance, a function call, and you get an error anywhere else in the code, just becouse it was there, but you were not using that code portion, is very weird. Not sure I follow you there...could you clarify a bit? |
   |