[Zlib-devel] zlib 1.2.3.6 available for testing
Török Edwin
edwintorok at gmail.com
Wed Jan 20 09:29:09 EST 2010
On 01/20/2010 03:41 PM, Enrico Weigelt wrote:
>
>> yes, it is small. And it's not necesary to use the autotools. The
>> current system is working. But a change in the files (change of names,
>> adding or removing files, options passed to the proprocessor, compiler,
>> linker, etc...) mean a change in more than 10 Makefiles, that is, a high
>> probability to make mistakes when updating them.
>>
>
> True. But we could easily move common parts to includes or
> splitted templates for auto-generation - step by step. And so
> we can write code that can be easily understood by people who
> are not autoconf experts.
>
>
Except for msdos and win32 don't we already have a single Makefile.in
for all OSes?
I like zlib's configure better than autoconf: it doesn't do any useless
checks that autoconf does,
is fast, and has obviously worked very well so far.
autoconf and libtool are good for larger projects, which need to check
for availability
of lots of system functions, headers, features, etc.
IMHO autoconf and libtool in zlib would only add complexity and wouldn't
solve any real problem.
I guess that the current approach of configure+Makefile.in could be used
for the msdos/win32 makefiles too.
Best regards,
--Edwin
More information about the Zlib-devel
mailing list