mingw-cross-env-list
[Top][All Lists]
Advanced

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

Re: [Mingw-cross-env-list] New package proposal: libircclient


From: Lothar May
Subject: Re: [Mingw-cross-env-list] New package proposal: libircclient
Date: Thu, 26 May 2011 21:53:42 +0200

Hi,

sorry for the delay, I was busy moving...

2011/3/13 Tony Theodore <address@hidden>:
> On 11 March 2011 12:59, Lothar May <address@hidden> wrote:
>> Hi,
>>
>> please find enclosed the proposed new package libircclient.
>> Since "make install" is broken, I install the required files by hand
>> (the same files as with libircclient 1.3 on Ubuntu/Debian are
>> installed, in the same structure, and they are sufficient for
>> developing).
>> The Makefile is also broken, a patch is included.
>
> That seems reasonable, though I wonder if it's also worth fixing the
> install target if you're patching Makefile.in anyway.

I just tried for over an hour to do this. However, the build of
libircclient seems to be totally broken, even the prefix is unknown in
this makefile, although set in configure. I gave up and left the
install commands in the .mk file.

>> IPv6 support suffers
>> from a buffer overflow issue and works only partly, is therefore also
>> disabled.
>
> I'm not sure this is actually disabled, the only place config.h is
> used is in portable.c, and then only for non-win32 builds.

I've added a patch to force it to be disabled.

>> I can use the resulting lib for PokerTH, so in my case it works fine.
>>
>> I'm unsure about the define $(PKG)_UPDATE-part (was copied from
>> another package).
>
> This looks okay and returns the right result.
>
>> I'm also unsure about the "rcu" options to ar, they were also copied
>> because it did not work without them.
>
> You could get by with just "cr", but a better option is to pass
>
> --host='$(TARGET)'

I've changed this, thanks for the hint.

>
> to configure. That way you can simply call make without CC/AR etc.
>
>> One question remains: The example program, if build in strict ANSI
>> mode, requires -DWIN32 because libircclient considers WIN32 instead of
>> _WIN32. Should I create a patch for libircclient and remove this
>> define?
>
> I'd say so, it's a fairly simple patch (you should also remove the U__*).

I've added a patch, and also removed the U__... option.

Attached are the new files. Any more feedback is appreciated.

Regards,
Lothar

Attachment: libircclient.mk
Description: Binary data

Attachment: libircclient-1-makefile.patch
Description: Binary data

Attachment: libircclient-2-disable-ipv6.patch
Description: Binary data

Attachment: libircclient-3-win32def.patch
Description: Binary data

Attachment: libircclient-test.cpp
Description: Text Data


reply via email to

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