Re: [lmi] Do we have zlib already?

[Greg Chicares]

On 2016-06-29 11:51, Vadim Zeitlin wrote:
[...]
>  I have, of course, completely forgotten about the conflict between the old
> version of libxml2 and newer (but basically all of them by now, as "newer"
> means "less than 11 years old") version of zlib included in MinGW 4.9.1
> that already gave me some grief before, see
> 
>       https://lists.gnu.org/archive/html/lmi/2016-01/msg00026.html
> 
> The solution chosen back then was to disable zlib support and was done in
> 91a19b246d553712e7f92606a88fb6487dd15621, but re-enabling it makes the
> problem appear again, of course, so now we need some other solution.

Curiously enough, it sounds like we always had the ability to use zlib
with libxml2 when we were using gcc-3.4.5, and we just never used it.
But we had problems building this ancient libxml2 version with gcc-4.9.1,
so we added these options:
  '--without-threads' in commit 64dd4f93da611a51c74158da253e29609d7f42f7
  '--without-zlib'    in commit 91a19b246d553712e7f92606a88fb6487dd15621
as a workaround.

>  I still think that updating libxml2 to a modern version would be the best
[...alternatively, we could patch the ancient version, but let's not...]
> Please let me know what would you prefer and I'll make the appropriate
> changes in install_libxml2_libxslt.make.

We should update libxml2. I hope we can then use both threads and zlib
with it. It seems only prudent to update libxslt simultaneously.

> GC> >  I'd just like to know if compressing the files above and checking if an
> GC> > illustration can be created is a sufficient test or whether I should 
> try to
> GC> > compress/check something more/else?
> GC> 
> GC> I believe that's not merely sufficient, but plenary and conclusive.
> 
>  FWIW I can confirm that after
> 
> 1. Applying the libxml2 patch above
> 2. Removing --without-zlib
> 3. Compressing all XML files
> 
>       % for f in 
> /opt/lmi/data/sample.{database,funds,policy,rounding,strata}; do
>               gzip -c $f > $f.tmp && mv $f.tmp $f
>       done
> 
> lmi still seems to work correctly and, notably, illustration creation
> works.
> 
>  But, again, to make this work out of the box, we need to do something
> about our ancient libxml2: either patch or upgrade it.

Presumably upgrading the xmlsoft libraries will work at least as well as
(1.) above. If you would please provide a patch, I'll test everything.