So I have a somewhat odd windows bug. Not sure if learn is the right place to ask this but....

I have a pragma to link in ws2_32.lib in my code for windows builds
and I have code that makes use of htonl(). When I compile code that doesn't use htonl() but makes use of sockets the code compiles and runs fine.  When I use code that uses htonl() but no sockets the code compiles and runs fine.
But when I try to compile code using both sockets and htonl() I get the following error.

C:\dmd\lib\phobos.lib(winsock)  Offset 1D966H Record Type 00C3
 Error 1: Previous Definition Different : _htonl@4


I'm using the build tool(bud.exe) to do my compiling and linking for me and it's not let me down yet so I'm hoping it's not build but I'm really not sure what to do about this error.  I opened up phobos.lib with vim and looked around abit and the string _htonl@4 appears 3 seperate places in the lib file... then again reading lib files isn't all that informative due to the whole not human readable thing.

Cheers,

Dan

On Tue, 12 Dec 2006 01:33:51 -0500, Dan Abretske wrote:

> So I have a somewhat odd windows bug. Not sure if learn is the right place to ask this but....
> 
> I have a pragma to link in ws2_32.lib in my code for windows builds
> and I have code that makes use of htonl(). When I compile code that
> doesn't use htonl() but makes use of sockets the code compiles and runs
> fine.  When I use code that uses htonl() but no sockets the code
> compiles and runs fine.
> But when I try to compile code using both sockets and htonl() I get the
> following error.
> 
> C:\dmd\lib\phobos.lib(winsock)  Offset 1D966H Record Type 00C3
>   Error 1: Previous Definition Different : _htonl@4
> 
> I'm using the build tool(bud.exe) to do my compiling and linking for me and it's not let me down yet so I'm hoping it's not build but I'm really not sure what to do about this error.  I opened up phobos.lib with vim and looked around abit and the string _htonl@4 appears 3 seperate places in the lib file... then again reading lib files isn't all that informative due to the whole not human readable thing.

I've asked about this error message from time to time and still yet to receive an answer on it. However, sometimes it goes away if you use the -full switch on the Bud command line.

-- 
Derek
(skype: derek.j.parnell)
Melbourne, Australia
"Down with mediocrity!"
12/12/2006 5:38:15 PM

Heh,

Glad to know I'm not alone in this :)  As for -full it looks like it's not gonna save the day this time :)

Thanks though :)

Dan

Derek Parnell wrote:
> On Tue, 12 Dec 2006 01:33:51 -0500, Dan Abretske wrote:
> 
>> So I have a somewhat odd windows bug. Not sure if learn is the right place to ask this but....
>>
>> I have a pragma to link in ws2_32.lib in my code for windows builds
>> and I have code that makes use of htonl(). When I compile code that doesn't use htonl() but makes use of sockets the code compiles and runs fine.  When I use code that uses htonl() but no sockets the code compiles and runs fine.
>> But when I try to compile code using both sockets and htonl() I get the following error.
>>
>> C:\dmd\lib\phobos.lib(winsock)  Offset 1D966H Record Type 00C3
>>   Error 1: Previous Definition Different : _htonl@4
>>
>> I'm using the build tool(bud.exe) to do my compiling and linking for me and it's not let me down yet so I'm hoping it's not build but I'm really not sure what to do about this error.  I opened up phobos.lib with vim and looked around abit and the string _htonl@4 appears 3 seperate places in the lib file... then again reading lib files isn't all that informative due to the whole not human readable thing.
> 
> I've asked about this error message from time to time and still yet to
> receive an answer on it. However, sometimes it goes away if you use the
> -full switch on the Bud command line.
> 

Dan Abretske schrieb am 2006-12-12:
> So I have a somewhat odd windows bug. Not sure if learn is the right place to ask this but....
>
> I have a pragma to link in ws2_32.lib in my code for windows builds
> and I have code that makes use of htonl(). When I compile code that
> doesn't use htonl() but makes use of sockets the code compiles and runs
> fine.  When I use code that uses htonl() but no sockets the code
> compiles and runs fine.
> But when I try to compile code using both sockets and htonl() I get the
> following error.
>
> C:\dmd\lib\phobos.lib(winsock)  Offset 1D966H Record Type 00C3
>   Error 1: Previous Definition Different : _htonl@4

(I don't have a Window box at hand)

Does ws2_32.lib define a _htonl@4 symbol?
If it does it clashs with std.c.windows.winsock.htonl and you should
file a Phobos bug.

Thomas

Alright I'll check that.

The odd resolution to my problem was to move my main functions to separate files and then run build on those files but import the original file. I'm still not sure what's wrong but at least I got it working :)

Dan

Thomas Kuehne wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Dan Abretske schrieb am 2006-12-12:
>> So I have a somewhat odd windows bug. Not sure if learn is the right place to ask this but....
>>
>> I have a pragma to link in ws2_32.lib in my code for windows builds
>> and I have code that makes use of htonl(). When I compile code that doesn't use htonl() but makes use of sockets the code compiles and runs fine.  When I use code that uses htonl() but no sockets the code compiles and runs fine.
>> But when I try to compile code using both sockets and htonl() I get the following error.
>>
>> C:\dmd\lib\phobos.lib(winsock)  Offset 1D966H Record Type 00C3
>>   Error 1: Previous Definition Different : _htonl@4
> 
> (I don't have a Window box at hand)
> 
> Does ws2_32.lib define a _htonl@4 symbol?
> If it does it clashs with std.c.windows.winsock.htonl and you should
> file a Phobos bug.
> 
> Thomas
> 
> 
> -----BEGIN PGP SIGNATURE-----
> 
> iD8DBQFFfyL9LK5blCcjpWoRAjLcAJ9wISBfmEi21vNcOoODzWq7EERnwwCfaY/V
> 9kaVAEFIEPV1nfa7bLA5UHU=
> =62XU
> -----END PGP SIGNATURE-----