nel-all
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

OOPS! Re: [Nel] problems compiling.


From: HERNANDEZ CORDOBA RODRIGO JOSE
Subject: OOPS! Re: [Nel] problems compiling.
Date: Tue, 02 Nov 2004 10:45:30 -0600

Sorry, use pserver instead of ext, and you might want to try
compression, not sure if it works though

cvs -d:pserver:address@hidden:/cvsrepo co nevrax

or

cvs -z3 -d:pserver:address@hidden:/cvsrepo co nevrax

----- Mensaje Original -----
De: HERNANDEZ CORDOBA RODRIGO JOSE <address@hidden>
Fecha: Martes, Noviembre 2, 2004 10:26 am
Asunto: Re: [Nel] problems compiling.

> 
> Well, if you dont mind downloading the whole tree, the code can be
> retrieved with cvs like so:
> 
> cvs -d:ext:address@hidden:/cvsrepo co nevrax
> 
> use a blank password (just hit return when prompted)
> 
> This is work in progress, and currently does not compile on either 
> MSYSor Linux (I am adding the inline asm blocks to the Linux code 
> as well)
> and there are some issues on VC7, have not tried VC6, the guess is 
> theremight be some issues there as well.
> 
> like I said, nelmisc.dll is the only one that does get compiled so 
> far.
> Cheers!
> 
> ----- Mensaje Original -----
> De: Vianney Lecroart <address@hidden>
> Fecha: Martes, Noviembre 2, 2004 10:28 am
> Asunto: Re: [Nel] problems compiling.
> 
> > Hi Rodrigo,
> > 
> > Thank you for your patch but we can't apply it since it modifies 
> > to much 
> > unused files (Entries CVS directory, $Id and others).
> > It seems that you did a fork of nel with your modifications. If 
> > you give 
> > us the modification, we'll be able to add then on a main tree 
> and 
> > avoid 
> > bad effect of forked CVS.
> > 
> > Vianney Lecroart
> > 
> > Rodrigo Hernandez wrote:
> > > 
> > > I have created a patch (attached) however I do not believe 
> this 
> > is the 
> > > optimal form, as it is 2 mb uncompressed and I am sure I did 
> not 
> > made 
> > > that many changes, its not entirely functional (the code still 
> > does not 
> > > entirely compile under MSYS) and has many '$Id:' changes due 
> to 
> > the 
> > > import that I don't think are right.
> > > 
> > > I am submitting this patch with the hopes that it may be 
> useful 
> > for 
> > > example to get the VC++ 7.1 projects and solution files to a 
> > more 
> > > portable state, since they have relative paths included in 
> > themselves, 
> > > and some other changes related to MSYS compilation (autotools 
> > and 
> > > defines mostly).
> > > 
> > > Also so I can get some help with the problems I described 
> before :)
> > > 
> > > Once I get the source tree to properly compile I will create a 
> > proper 
> > > patch.
> > > 
> > > Thanks :)
> > > 
> > > At 06:49 PM 10/22/2004, you wrote:
> > > 
> > >> if you could make .patch files and throw them at the mailing 
> > list that 
> > >> would be handy indeed
> > >>
> > >> then vianney lecroart can patch the NeL cvs tree with ones 
> that 
> > are 
> > >> approved
> > >>
> > >>
> > >> ----- Original Message ----- From: "HERNANDEZ CORDOBA RODRIGO 
> > JOSE" 
> > >> <address@hidden>
> > >> To: "Developer's list for the NeL platform" <address@hidden>
> > >> Sent: Friday, October 22, 2004 5:04 PM
> > >> Subject: Re: [Nel] problems compiling.
> > >>
> > >>
> > >>> Hi Andi,
> > >>>
> > >>> Great! I updated my cvs yesterday, but before I did was 
> still 
> > having the
> > >>> replicate symbols errors on VC++7.1.
> > >>>
> > >>> also on the clamp asm right at the top I realized after I 
> sent 
> > the email
> > >>> that instead of:
> > >>>
> > >>> mov eax, v
> > >>> mov ebx, v
> > >>>
> > >>> You should do
> > >>>
> > >>> mov eax, v
> > >>> mov ebx, eax
> > >>>
> > >>> since moving data from register to register is faster than 
> > memory to
> > >>> register, in any case in both cases eax and ebx end up with 
> > the same
> > >>> value, you can check my copy of the cvs tree to see the 
> inline gas
> > >>> version, I also changed some other asm around and added a 
> > check for
> > >>> 3DNow! on cpu_info.cpp I think it is.
> > >>>
> > >>> Cheers! ;)
> > >>>
> > >>>
> > >>> ----- Mensaje Original -----
> > >>> De: "Andi [ Debug ]" <address@hidden>
> > >>> Fecha: Viernes, Octubre 22, 2004 4:06 pm
> > >>> Asunto: Re: [Nel] problems compiling.
> > >>>
> > >>>> i wlil have a look maybe i need to update the projects 
> based on
> > >>>> the new cvs
> > >>>> tree i do this only once a week
> > >>>> on fridays
> > >>>>
> > >>>> oh thats today :)
> > >>>>
> > >>>>
> > >>>> ----- Original Message ----- From: "HERNANDEZ CORDOBA 
> RODRIGO 
> > JOSE" 
> > >>>> <address@hidden>
> > >>>> To: <address@hidden>
> > >>>> Sent: Friday, October 22, 2004 2:53 PM
> > >>>> Subject: [Nel] problems compiling.
> > >>>>
> > >>>>
> > >>>> > Hello all,
> > >>>> > I am having some troubles compiling, some I am not sure 
> if its
> > >>>> due to my
> > >>>> > changes or due to the latest changes on CVS, hope you can 
> > help me:
> > >>>> >
> > >>>> > On VC++7.1 I get tons of already defined symbol errors, 
> > here is
> > >>>> a sample:
> > >>>> >
> > >>>> > nlmisc_d.lib(debug.obj) : error LNK2005: _RtlSecureZeroMemory
> > >>>> already> defined in nlmisc_d.lib(log.obj)
> > >>>> > nlmisc_d.lib(debug.obj) : error LNK2005: 
> _InterlockedAnd64 
> > already>>>> > defined in nlmisc_d.lib(log.obj)
> > >>>> > nlmisc_d.lib(debug.obj) : error LNK2005: _InterlockedOr64 
> > already>>>> > defined in nlmisc_d.lib(log.obj)
> > >>>> > nlmisc_d.lib(debug.obj) : error LNK2005: 
> _InterlockedXor64 
> > already>>>> > defined in nlmisc_d.lib(log.obj)
> > >>>> > nlmisc_d.lib(debug.obj) : error LNK2005: 
> > _InterlockedIncrement64>>>> already> defined in 
> nlmisc_d.lib(log.obj)> >>>> > nlmisc_d.lib(debug.obj) : error 
> LNK2005: 
> > _InterlockedDecrement64>>>> already> defined in 
> nlmisc_d.lib(log.obj)> >>>> > nlmisc_d.lib(debug.obj) : error 
> LNK2005: _InterlockedExchange64
> > >>>> already> defined in nlmisc_d.lib(log.obj)
> > >>>> > nlmisc_d.lib(debug.obj) : error LNK2005: 
> > _InterlockedExchangeAdd64>>>> > already defined in 
> > nlmisc_d.lib(log.obj)>>>> >
> > >>>> > Those seem to be all windows functions so I have no clue 
> about> >>>> what is
> > >>>> > wrong.
> > >>>> >
> > >>>> > Then on device_manager.cpp there is the line
> > >>>> >
> > >>>> > #include "nel/misc/device_manager.h"
> > >>>> >
> > >>>> > However the file device_manager.h is no where to be 
> found, the
> > >>>> file only
> > >>>> > has an empty constructor so I asume everything is inlined 
> > in the
> > >>>> header> or the class is empty, I created an empty class 
> > inside the
> > >>>> cpp file and
> > >>>> > removed the include line, but I dont know if this is ok 
> or not.
> > >>>> >
> > >>>> > I am making changes so it is possible to compile NEL on MSYS
> > >>>> with MinGW.
> > >>>> >
> > >>>> > I also found a bug on the fastClamp8(sint &v) function in
> > >>>> algo.h, the
> > >>>> > asm reads:
> > >>>> > // clamp v in 0..255 (no cond jmp)
> > >>>> > __asm
> > >>>> > {
> > >>>> > mov esi, v
> > >>>> > mov eax, [esi]
> > >>>> > mov ebx, eax
> > >>>> > // clamp to 0.
> > >>>> > add eax, 0x80000000
> > >>>> > sbb ecx, ecx
> > >>>> > not ecx
> > >>>> > and ebx, ecx
> > >>>> > // clamp to 255.
> > >>>> > add eax, 0x7FFFFF00
> > >>>> > sbb ecx, ecx
> > >>>> > or ebx, ecx
> > >>>> > and ebx, 255
> > >>>> > // store
> > >>>> > mov [esi], ebx
> > >>>> > }
> > >>>> >
> > >>>> > but it should be:
> > >>>> >
> > >>>> > __asm
> > >>>> > {
> > >>>> >    mov eax, v
> > >>>> >    mov ebx, v
> > >>>> >    // clamp to 0.
> > >>>> >    add eax, 0x80000000
> > >>>> >    sbb ecx, ecx
> > >>>> >    not ecx
> > >>>> >    and ebx, ecx
> > >>>> >    // clamp to 255.
> > >>>> >    add eax, 0x7FFFFF00
> > >>>> >    sbb ecx, ecx
> > >>>> >    or ebx, ecx
> > >>>> >    and ebx, 255
> > >>>> >    // store
> > >>>> >    mov v, ebx
> > >>>> > }
> > >>>> >
> > >>>> > Anyway, I have a public CVS where I keep my imported copy of
> > >>>> Nel, if
> > >>>> > someone wants to take a look the CVSROOT is
> > >>>> >
> > >>>> > :pserver:address@hidden:/cvsrepo
> > >>>> >
> > >>>> > the module is 'nevrax' just hit enter for the password, 
> and be
> > >>>> warned, I
> > >>>> > added the game data to it, so its not quite dial up friendly,
> > >>>> you should
> > >>>> > be able to run a diff against the official CVS, though I 
> > dont really
> > >>>> > know how yet :)
> > >>>> >
> > >>>> > Thanks for any help.
> > >>>> >
> > >>>> >
> > >>>> >
> > >>>> >
> > >>>> >
> > >>>> > _______________________________________________
> > >>>> > Nel mailing list
> > >>>> > address@hidden
> > >>>> > http://lists.nongnu.org/mailman/listinfo/nel-all
> > >>>> >
> > >>>> >
> > >>>>
> > >>>>
> > >>>>
> > >>>>
> > >>>> _______________________________________________
> > >>>> Nel mailing list
> > >>>> address@hidden
> > >>>> http://lists.nongnu.org/mailman/listinfo/nel-all
> > >>>
> > >>>
> > >>>
> > >>>
> > >>> _______________________________________________
> > >>> Nel mailing list
> > >>> address@hidden
> > >>> http://lists.nongnu.org/mailman/listinfo/nel-all
> > >>>
> > >>
> > >>
> > >>
> > >>
> > >> _______________________________________________
> > >> Nel mailing list
> > >> address@hidden
> > >> http://lists.nongnu.org/mailman/listinfo/nel-all
> > > 
> > > 
> > > Rodrigo Hernandez, lonewolf programmer
> > > Aeon Games
> > > http://www.aeongames.com
> > > 
> > > 
> > > ---------------------------------------------------------------
> --
> > -------
> > > 
> > > _______________________________________________
> > > Nel mailing list
> > > address@hidden
> > > http://lists.nongnu.org/mailman/listinfo/nel-all
> > 
> > -- 
> > Commit du soir, espoir. Update du matin, chagrin.
> > 
> > 
> > 
> > _______________________________________________
> > Nel mailing list
> > address@hidden
> > http://lists.nongnu.org/mailman/listinfo/nel-all
> > 
> 
> 
> 
> _______________________________________________
> Nel mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/nel-all
> 





reply via email to

[Prev in Thread] Current Thread [Next in Thread]