On 2025/11/28 15:43, Johannes Thyssen Tishman wrote:
> +GH_ACCOUNT = openscad
> +GH_PROJECT = openscad
> +GH_COMMIT =  b0f18e9b4d61288b2cd6c659135f7e77def179e0
>  
> -SITES =                      https://files.openscad.org/
> -EXTRACT_SUFX =               .src.tar.gz
> +MCAD_COMMIT =        1ea402208c3127ffb443931e9bb1681c191dacca
> +SITES.mcad = https://github.com/${GH_ACCOUNT}/MCAD/archive/
>  
> -MODULES =            devel/qmake \
> +DISTFILES.mcad =     
> MCAD-${MCAD_COMMIT:C/(........).*/\1/}{${MCAD_COMMIT}}${EXTRACT_SUFX}

i have some nits with these bits and associated in post-extract, but
they can be handled after the update is in

> -BUILD_DEPENDS =              devel/bison \
> +# Match devel/boost
> +COMPILER =           base-clang ports-gcc

missing COMPILER_LANGS

> +BUILD_DEPENDS =              cad/clipper2 \

this is in both BUILD_DEPENDS and LIB_DEPENDS, should just be the latter

> +TEST_DEPENDS =               graphics/ImageMagick
> +
> +# A test requires this
> +TEST_DEPENDS +=              shells/bash

both of those are already in BUILD_DEPENDS, and you can't get to
testing if you didn't build

> +# The openscad script wrapper breaks debug packages
> +#DEBUG_PACKAGES =    ${BUILD_PACKAGES}

what's up here? there's no reason for a script to break debug packages.

> +# We don't have Pillow
> +CONFIGURE_ARGS +=    -DUSE_IMAGE_COMPARE_PY=OFF

we do, graphics/py-Pillow if it's useful for something

> +     #find ${WRKSRC} -type f -exec sed -i 
> '1s,^#!/bin/sh,#!${PREFIX}/bin/bash,' {} +

commented-out; either remove or if it's needed somewhere then should
be targetted, running sed -i on all of extracted source isn't ideal

> +# OpenSCAD uses the program_location (see patch-src_openscad_cc) to
> +# figure out resource paths, but on OpenBSD that doesn't work when
> +# launched from the PATH So, add a shell script shim to fix up argv[0]
> +# to be absolute so our patch works.

(not related to the port, but it's such a pain that we don't have this,
it is widely expected by many things in ports these days)

Reply via email to