Re: [ECOS] No zlib crc32 support?

[Andrew Lunn <andrew at lunn dot ch>]

On Tue, Dec 09, 2003 at 02:04:05PM -0700, Steve Strublic wrote:
 
> Would there be any objection to including it (meaning the 'official'
> zlib version of crc32) in the zlib package anyhow?  In my case (and
> I would think many others), it makes what I'm doing completely
> portable between Linux and eCos without having to ifdef for eCos.

I took this and several other crc32 implementations out from various
bits of eCos simply because an image would end up with many
implementations in the same image. This is particularly bad because
each one had its own table of coefficients which is quite large and
totally redundant. eCos wants to be small so this was an obvious
optimization.

There was also the name space pollution problem. One of the crc32()
functions actually implemented a different variation on the crc32()
algorithm which was incompatible with all the other crc32 algorithms.
Plus some "user" code may want to yet again implement its own crc32
algorithm. So sorting this mess out i decide it was better to stick to
the standard cyg_ prefix convention.

I would be reluctant to put back the crc32 function in zlib. Its a
step back towards chaos, bigger images and potential naming problems.

Either do as Gary suggested, or call cyg_crc32(). 

       Andrew

-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss