Re: [Mingw-cross-env-list] mingw-w64 and mingw-cross-env

[Tony Theodore]

On 26/03/2012, at 9:10 PM, Volker Grabsch wrote:

> Tony Theodore schrieb:
[...]
> 
>> 1. Implicit rules when not specified and define rules for failing packages
>> 2. Explicit rules always and skip failures with a missing (or empty) rule
>> 
>> 1 is less code and looks cleaner, but I'm leaning back towards 2 for
>> the extra clarity.
> 
> I alse have a strong feeling that 2) is the better approach
> in the long run.

Indeed, in practice it does seem to work out better - I've pushed the latest 
version here:

https://github.com/tonytheodore/mxe/commit/7930a673b3ac681f39e7beaad149cae7fe510797

[...]

>>> - Most packages use the same build code for all toolchains,
>>>  with a small conditional about static/dynamic builds.
>> 
>> What do you think of something like a LINK_STYLE variable [5]? I
>> imagine this would work in the simple cases.
> 
> Looks great!

The only problem is that it looks a little out of place among the other --* 
options, but my sed skills aren't up to reordering them.

Perhaps we should extend the idea to a CONFIGURE_OPTIONS variable that also 
sets --host, --prefix, and possibly --build?

Otherwise, the i686-static-mingw32 builds everything successfully, and there 
are 34 failing packages for the x86_64-static-mingw32 target. Many of these 
seem related to either pthreads or libgcrypt, so I imagine once those are 
sorted out we'll be a lot closer to full support. Also, some of the test 
programs seem to require a different standard than -ansi. I still haven't 
gotten around to setting up a win64 machine again, so I can't verify that these 
do anything apart from building.

Cheers,

Tony