[Zlib-devel] Second zlib 1.2.4 candidate

[Vincent Torri]

On Sun, 14 Mar 2010, Cosmin Truta wrote:

> On Sat, Mar 13, 2010 at 9:49 PM, Mark Adler wrote:
>>   http://alumnus.caltech.edu/~madler/zlib-1.2.4.tar.gz
>
> Works well for me so far, but, Mark, could you please put back the
> following files:
> contrib/masmx64/gvmat64.obj
> contrib/masmx64/inffasx64.obj
> contrib/masmx86/gvmat32.obj
> contrib/masmx86/inffas32.obj
>
> They used to be there in the previous zlib distributions, for the sake
> of the users who lack MASM and still want to use the asm routines.

is there a reason not to write portable asm files (using yasm or nasm as 
compiler) ? performance ?

Vincent Torri

> The readme files in the two directories actually do imply the
> existence of these files. For example, here is a quote from
> contrib/masmx86/readme.txt:
>  Use instructions
>  ----------------
>  Copy these files into the zlib source directory, then run the
>  appropriate makefile, as suggested below.
>
> I realize other people might disagree on this, because nowadays MASM
> is available as a free download. If that's the case, fine by me.
>
> **
>
> Another thing, not necessarily a stop-ship, but a good thing to do in
> case this gets another revision:
> In gzguts.h at line 32, the substitution vsnprintf --> _vsnprintf
> should only be done under _MSC_VER, because that one isn't fully ISO C
> compatible. So replace #ifdef WIN32 with #if defined(_MSC_VER).
>
> **
>
> One last thing: the occurence of _MSV_VER in zutil.h is a typo, isn't it?
> _MSV_VER --> _MSC_VER
>
> Best regards,
> Cosmin
>
> _______________________________________________
> Zlib-devel mailing list
> Zlib-devel at madler.net
> http://mail.madler.net/mailman/listinfo/zlib-devel_madler.net
>
>