The latest SVN bugs thread

You discovered a bug in the engine, and you are sure that it is not a problem of your code? Just post it in here. Please read the bug posting guidelines first.
Nadro
Posts: 1648
Joined: Sun Feb 19, 2006 9:08 am
Location: Warsaw, Poland

Re: The latest SVN bugs thread

Post by Nadro »

Mipmaps issue is already fixed in the latest trunk.
Library helping with network requests, tasks management, logger etc in desktop and mobile apps: https://github.com/GrupaPracuj/hermes
Simson
Competition winner
Posts: 95
Joined: Wed Nov 30, 2005 8:53 am
Location: France : midi pyrénées

Re: The latest SVN bugs thread

Post by Simson »

Yessss, thank you Nadro :)
Nadro
Posts: 1648
Joined: Sun Feb 19, 2006 9:08 am
Location: Warsaw, Poland

Re: The latest SVN bugs thread

Post by Nadro »

Was It solve your problems visible on the screen from page 14?
Library helping with network requests, tasks management, logger etc in desktop and mobile apps: https://github.com/GrupaPracuj/hermes
Simson
Competition winner
Posts: 95
Joined: Wed Nov 30, 2005 8:53 am
Location: France : midi pyrénées

Re: The latest SVN bugs thread

Post by Simson »

No that doesn't solve my problem of filtering but that solve the mip map problem...
For the filtering bug, I apply a material then made first render pass in render to texture and after I apply another material then new render pass ... maybe that can help you.
If I have the time, I'll try to reproduce this bug in small example this week.
Thanks
Simson
Competition winner
Posts: 95
Joined: Wed Nov 30, 2005 8:53 am
Location: France : midi pyrénées

Re: The latest SVN bugs thread

Post by Simson »

Well, I have tested my class individually and that work perfect...
I think there is a bug in MY code that does not appear on the previous version of irrlicht for unknown reason at this time.
I will try to clean my render passes.
Thank for your attention Nadro.
hendu
Posts: 2600
Joined: Sat Dec 18, 2010 12:53 pm

Re: The latest SVN bugs thread

Post by hendu »

@Nadro

COpenGLDriver.cpp:771
Quad2DVertices[0] =
Quad2DVertices[0] = ...
Quad2DVertices[0] = ...
Quad2DVertices[0] = ...

;)
Nadro
Posts: 1648
Joined: Sun Feb 19, 2006 9:08 am
Location: Warsaw, Poland

Re: The latest SVN bugs thread

Post by Nadro »

OMG! Thanks for a report :) It'll be fixed at wednesday.

Update: Fixed in the latest trunk.
Library helping with network requests, tasks management, logger etc in desktop and mobile apps: https://github.com/GrupaPracuj/hermes
greenya
Posts: 1012
Joined: Sun Jan 21, 2007 1:46 pm
Location: Ukraine
Contact:

Re: The latest SVN bugs thread

Post by greenya »

Hello.
Trying to build latest svn trunk (rev. 4515) and fail with
error C2664: 'irr::core::cutFilenameExtension' : cannot convert parameter 1 from 'irr::core::stringc' to 'irr::io::path &' in \source\Irrlicht\COBJMeshWriter.cpp at line 68
I build with _IRR_WCHAR_FILESYSTEM flag.

P.S.: changing type of "name" to "io::path" removes error:

Code: Select all

    ...
    // write OBJ MESH header
 
    io::path name;
    core::cutFilenameExtension(name,file->getFileName()) += ".mtl";
    file->write("# exported by Irrlicht\n",23);
    file->write("mtllib ",7);
    file->write(name.c_str(),name.size());
    file->write("\n\n",2);
    ...
hybrid
Admin
Posts: 14143
Joined: Wed Apr 19, 2006 9:20 pm
Location: Oldenburg(Oldb), Germany
Contact:

Re: The latest SVN bugs thread

Post by hybrid »

Yeah, right. That seems to have changed recently. Fixed now.
greenya
Posts: 1012
Joined: Sun Jan 21, 2007 1:46 pm
Location: Ukraine
Contact:

Re: The latest SVN bugs thread

Post by greenya »

Cannot compile latest svn trunk rev.4525.

Code: Select all

error C2057: expected constant expression in \source\Irrlicht\CMeshManipulator.cpp at line 1083
error C2057: expected constant expression in \source\Irrlicht\CMeshManipulator.cpp at line 1092
Seems like maybe its C++11 standard features, but i'm not sure.
I use \examples\BuildAllExamples_vc10.sln for compiling.
zerochen
Posts: 273
Joined: Wed Jan 07, 2009 1:17 am
Location: Germany

Re: The latest SVN bugs thread

Post by zerochen »

hi,

doesnt work under 2012 as well because something like this isnt allowed

Code: Select all

 
int a = getAInt();
char s[a];
 
should be

Code: Select all

 
int a = getAInt();
char* s = new char[a];
...
delete[] s;
 
here is a patch for that. CMeshManipulator.cpp.patch

i also removed the goto's in that function O.o
hope that i didnt make an error

regards
zerochen
hendu
Posts: 2600
Joined: Sat Dec 18, 2010 12:53 pm

Re: The latest SVN bugs thread

Post by hendu »

I can't download that, I see no download button:

Image

Please use the SF.net tracker.

I can't comment on how you removed the gotos until I can see the patch, but with nested loops they are often the best solution.
zerochen
Posts: 273
Joined: Wed Jan 07, 2009 1:17 am
Location: Germany

Re: The latest SVN bugs thread

Post by zerochen »

4 downloads already...
however, here we go.
https://sourceforge.net/p/irrlicht/bugs/416/

also i think if you split the func into seperate funcs it will become more readable. have not much time right now.
hendu
Posts: 2600
Joined: Sat Dec 18, 2010 12:53 pm

Re: The latest SVN bugs thread

Post by hendu »

OK, NAK.

- many unneeded changes like indent and bracing, or using new. If you want to change the brace style, or new, each should be a separate patch
- the goto removal appears to break the logic, make the code more complex (two more levels of indent) and make it slower. Not acceptable.
It also added a possible crash, if a too complex mesh was given.

Please test if my patch lets it compile on VS.
zerochen
Posts: 273
Joined: Wed Jan 07, 2009 1:17 am
Location: Germany

Re: The latest SVN bugs thread

Post by zerochen »

many unneeded changes like indent and bracing, or using new.
the new's are necessary otherwise you cant compile irrlicht under visual studio.
indeed i followed irrlichts code style so i changed some brackets
- the goto removal appears to break the logic, make the code more complex (two more levels of indent) and make it slower. Not acceptable.
It also added a possible crash, if a too complex mesh was given.
have you a test case for that? it was late yesterday maybe i've made ​​an error.

maybe it s little bit slower but that doesn matter here because it s now safer (no goto).
Post Reply