[vlc-devel] [PATCH 5/6] contrib/bootstrap: Simplify host checks for mingw

[Martin Storsjö]

On Wed, 5 Jun 2019, Steve Lhomme wrote:

> On 2019-06-05 13:10, Marvin Scholz wrote:
>> ---
>>   contrib/bootstrap | 7 ++-----
>>   1 file changed, 2 insertions(+), 5 deletions(-)
>> 
>> diff --git a/contrib/bootstrap b/contrib/bootstrap
>> index fc6fbcfb2b..02a958ca79 100755
>> --- a/contrib/bootstrap
>> +++ b/contrib/bootstrap
>> @@ -360,11 +360,10 @@ case "${OS}" in
>>   		add_make_enabled "HAVE_WIN32"
>>   		case "${HOST}" in
>>   			*winphone*|*windowsphone*)
>> +				add_make_enabled "HAVE_WINSTORE"
>>   				add_make_enabled "HAVE_WINDOWSPHONE"
>>   				;;
>> -		esac
>> -		case "${HOST}" in
>> -			*winphone*|*windowsphone*|*winrt*|*uwp*)
>> +			*winrt*|*uwp*)
>>   				add_make_enabled "HAVE_WINSTORE"
>
> I think it's better to be able to associate directly what a define 
> corresponds to.
>
>>   				;;
>>   		esac
>> @@ -372,8 +371,6 @@ case "${OS}" in
>>   			amd64*|x86_64*)
>>   				add_make_enabled "HAVE_WIN64"
>>   				;;
>> -		esac
>> -		case "${HOST}" in
>
> armv7a and win64 are not in the same 'family' of defines. Another arm 
> variant could have win64 (and it should?) and another arm define. It's 
> better to keep them separate.

Yes, HAVE_WIN64 is/should be set for windows/aarch64 as well. It is set by 
main.mak though. I'm not sure of the division between setting it in 
bootstrap and main.mak though, both should probably be done consistently 
in either of them (or both?).

// Martin