On 2024年2月15日 19:58:31 GMT+01:00, Omar Polo <o...@omarpolo.com> wrote:
>Hello :)
>
>On 2024/02/14 21:56:52 -0500, Jag Talon <jag@aangat.lahat.computer> wrote:
>> Quick update, I just realized that I added a shell/bash dependency when
>> it's not necessary. Attached is the updated archive.
>
>still have to properly play with it, but here's some early comments on
>the port:
>
> - please don't start COMMENT with "a"
> - could use DIST_TUPLE (+ make makesum)
> - I'd prefer if packages names are lowercase
> - revision starts empty (i.e. no revision at all), only after changes
> in the port it's bumped to 0 and then incremented.
> - why set COMPILER? Haven't tested with gcc4, but a quick scan didn't
> reveal something that it shouldn't handle. Maybe there's a chance
> it'll work :)
> - sdl2 and sdl2-image are only LIB_DEPENDS, no need to set them
> BUILD_DEPENDS
> - it still needs bash as build depends
> - we can simplify do-build by using ALL_TARGET
> - ports are expected to respect CC and CFLAGS:
> - CC can be passed as COMPILER in MAKE_ENV
> - CFLAGS needs some patching, and drop the hardcoded -O2
> - by patching a script, all tests are passing, so enable them :)
> - in the readme, use TRUEPREFIX rather than LOCALBASE:
> - TRUEPREFIX is for this very port, LOCALBASE for where *other* ports
> are installed
> - in the readme suggesting to run firefox /usr/local/share/... is not
> going to work: most browsers can only read /tmp/ and ~/Downloads by
> default on OpenBSD. I'm personally using python3 -m http.server as a
> quick way to access these HTML files, not sure we can suggest it in
> the README however... Not a huge deal however.
>
>upstream install scripts would need some heavy work so let's keep the
>custom do-install.
>
>(a bit sad that needs a bdep on vim for xxd...)
>
>I'm attaching a diff against your makefile and an updated tarball with
>the other patches. With this it builds fine and tests are fully
>passing. I can also run the program and some of the installed examples.
>Still have to really dive in it however :)
>
>
>Thanks!
>
>Omar Polo
>
>
>--- Makefile.orig Thu Feb 15 19:47:12 2024
>+++ Makefile Thu Feb 15 19:47:57 2024
>@@ -1,11 +1,8 @@
>-COMMENT = a multimedia platform inspired by HyperCard
>+COMMENT = multimedia platform inspired by HyperCard
>
>-GH_ACCOUNT = JohnEarnest
>-GH_PROJECT = Decker
>-GH_TAGNAME = v1.38
>+DIST_TUPLE = github JohnEarnest Decker v1.38 .
>+PKGNAME = ${DISTNAME:L}
>
>-REVISION = 0
>-
> CATEGORIES = graphics
>
> HOMEPAGE = https://beyondloom.com/decker/
>@@ -15,25 +12,19 @@
> # MIT
> PERMIT_PACKAGE = Yes
>
>-COMPILER = base-clang ports-gcc
>-COMPILER_LANGS = c
>+WANTLIB += SDL2 SDL2_image c m
>
>-BUILD_DEPENDS = editors/vim \
>- devel/sdl2 \
>- devel/sdl2-image \
>-
> USE_GMAKE = Yes
>
>-NO_TEST = Yes
>+BUILD_DEPENDS = editors/vim \
>+ shells/bash
>
> LIB_DEPENDS = devel/sdl2 \
>- devel/sdl2-image \
>+ devel/sdl2-image
>
>-do-build:
>- cd ${WRKSRC} && \
>- gmake decker && \
>- gmake lilt && \
>- gmake docs
>+ALL_TARGET = decker lilt docs
>+
>+MAKE_ENV = COMPILER="${CC}"
>
> do-install:
> .for bin in decker lilt
>
ah, I forgot, WANTLIB was also missing. As the Makefile.template suggest, you
can use `make port-lib-depends-check' and copy-paste[0] its output in the
makefile after having done `make fake'.
[0]: there are some exceptions, but works 99% of the times :P
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
