[109887] users/mojca/wxports/x11/xchm/Portfile

[Ryan Schmidt]

On Aug 22, 2013, at 08:40, Mojca Miklavec wrote:

> On Thu, Aug 22, 2013 at 5:14 AM, Ryan Schmidt wrote:
>> 
>>> Modified: users/mojca/wxports/x11/xchm/Portfile
>>> ===================================================================
>>> --- users/mojca/wxports/x11/xchm/Portfile     2013-08-21 13:22:03 UTC (rev 109886)
>>> +++ users/mojca/wxports/x11/xchm/Portfile     2013-08-21 13:42:51 UTC (rev 109887)
>>> @@ -52,11 +52,12 @@
>>>    app.name                xCHM
>>>    app.executable          xchm
>>>    app.icon                ${worksrcpath}/art/${icon_name}.png
>>> -    # TODO: ImageMagick doesn't need to be universal
>>> -    # it's only needed to convert the image to png
>>> +    # TODO: ImageMagick is only needed to convert the image to png
>>>    # if there is a more lightweight program to do the same, that one should be used
>>> -    # or the icon as png should be fetched from elsewhere
>>> +    # or the icon as png/icns should be fetched from elsewhere
>>>    depends_build-append    port:ImageMagick
>>> +    depends_skip_archcheck-append \
>>> +                            port:ImageMagick
>> 
>> It must be only the port name, not the full depspec.
>> 
>>    depends_skip_archcheck-append \
>>                            ImageMagick
> 
> Now that I went through ten iterations anyway: how could I properly
> fetch the icon file from
>    http://sourceforge.net/p/xchm/code/ci/master/tree/art/xchm-128.png?format=raw
> instead of converting it on the fly, so that dependency on ImageMagic
> could be avoided altogether?

If you're asking how to deal with a master_sites URL that needs a query string, or more generally a master_sites URL that does not end with the filename, we have a recipe for that:

https://trac.macports.org/wiki/PortfileRecipes#fetchwithgetparams

You'd combine that with tags on the master_sites and distfiles (to indicate which distfile is from which master_site) and extract.only to indicate which files should be extracted.