taglib on windows with visual studio

[Roel Vanhout]

Scott Wheeler wrote:
> Recent versions of GCC like to have something similar (explicit symbol
> visibility), so I'm not opposed to adding a macro for that in the near
> future.

Aha very cool, I didn't know about that. For those interested, see eg 
http://gcc.gnu.org/wiki/Visibility.

> I don't have such a version of GCC for testing, but if you ask really 
> nicely
> you might be able to convince Michael to do some testing if you come up 
> with a patch for such.  :-)

I do have a couple of Linux of boxes that I could test on but I'd have 
to check their gcc versions. Who is Michael? It would be nice to have as 
many people as possible testing this.

> If all of that happens in the next few days it might make into 1.4.

I won't make that.

> *) Headers are structured the way that they are in the source because, 
> well,
> that's easier to work with when developing the library.  That won't 
> change. 
> They're all installed to the same location.  (By
> default /usr/local/include/taglib.)

Ok so I could make a windows 'devel' distribution with all the headers 
in one directory, too. That would suffice: I could provide a msvc 
project file (sort of like a build script for Visual Studio) with all 
paths coded in for people who want to compile the dll and then users 
would only have to add one directory to their header include directories 
for referencing the headers from their code. Which is what other 
libraries require too, so that seems acceptable.

> Now for the part where I'm going to be annoying -- I don't have any 
> desire to
> have Windows patches just so that TagLib can compile on Windows; what it
> needs for me to accept these patches is a Windows maintainer that will keep
> up the patches, package the releases, etc.  Perhaps you and Stefan working
> together on that would be good if you're both willing.  You seem to have a
> stronger Windows development background than he does.  (I can't imagine
> Stefan really seeing that as an insult.  ;-) )

I hope that not that many patches will be necessary after a first build 
system is in place. As long as the code remains fairly platform-agnostic 
(as it is now) that should be not a problem. Maybe it would be a good 
idea to distribute a separate Windows release of 1.4 after it's released 
and see if there's any interest at all. If Stefan and I are just doing 
it for ourselves (and even the two of use are using different 
environments) then there's not really a need for use to spend our time.

As a sidenote, are you interested in unit tests for taglib? I'll 
probably write some for my own use to see if the windows version works 
as it should, and I'd use cppunit, unless you're interested in including 
them with the sources (I see that you don't use a unit testing framework 
in toolkit-test.cpp) and don't want to introduce dependencies.


cheers,

roel


> 
> Cheers,
> 
> -Scott
> -- 
> Audience Member: "What was the hardest part of building TeX?"
> Donald Knuth: "It was all pretty easy."
> _______________________________________________
> taglib-devel mailing list
> taglib-devel at kde.org
> https://mail.kde.org/mailman/listinfo/taglib-devel
>