science/gramps: Update to 4.2.3

[woodsb02]

Summary

science/gramps: Update to 4.2.3

• Take maintainership
• Convert to USE_GITHUB (new location of main upstream repository)
• Update dependencies for new version (namely python 3.2+ and GTK3) [1]
• Convert "strongly recommended" options to dependencies (removing option) [1]
• Add "optional" dependencies as OPTIONS [1]
• Remove pkg-plist and convert build to USE_PYTHON=autoplist distutils
• Remove pre-extract upgrade warning as it is ineffective for pkg upgrades and now only applies for systems which have not been upgraded in 1.5 years.
• Add post-stage hack to fix staging support, copied from Arch Linux [2]
• Delete files/patch-src_gramps.py as the offending file has been removed
• Add files/patch-setup.py to allow "--single-version-externally-managed" [3]
• Add files/patch-setup.py to fix double slashes in generated by autoplist [4]
  
  Changes:
  
  https://github.com/gramps-project/gramps/blob/v4.2.2/NEWS
  
  [1] https://github.com/gramps-project/gramps/blob/v4.2.3/README [2] https://projects.archlinux.org/svntogit/community.git/tree/trunk/PKGBUILD?h=packages/gramps&id=ee1c7fae15872481d8f55868f4211efcbce15314 [3] https://github.com/gramps-project/gramps/commit/ab67c786b03552ce4b0ac8f0ed9d7af228d46fe4 [4] https://github.com/gramps-project/gramps/pull/96

PR: 207005

Test Plan

poudriere testport science/gramps

[mat]

science/gramps/Makefile

6 ↗

(On Diff #16536)

Is there a <tab> between = and v ?

33 ↗

(On Diff #16536)

Unless I'm reading it badly, this is the default and should be removed.

43–44 ↗

(On Diff #16536)

Shouldn't those be *_LIB_DEPENDS ?

49–50 ↗

(On Diff #16536)

This should most certainly be post-install.

[adamw]

science/gramps/Makefile

22 ↗

(On Diff #16536)

This is the default and should be removed.

30 ↗

(On Diff #16536)

If you add USES=localbase do you still need those lines?

33 ↗

(On Diff #16536)

I suspect it's there because of the DISTVERSIONPREFIX.

43–44 ↗

(On Diff #16536)

And the ${LOCALBASE}/lib shouldn't be necessary.

[mat]

science/gramps/Makefile

33 ↗

(On Diff #16536)

No, this is using USE_GITHUB, it gets overwritten anyway.

[woodsb02]

Address comments from mat and adamw:

• Add tab after DISTVERSIONPREFIX
• Do not set GH_PROJECT to its otherwise default value
• Add USES=localbase to replace common Makefile lines
• Do not set WRKSRC to its otherwise default value
• Convert GEXIV and GTKSPELL dependencies to LIB_DEPENDS and remove ${LOCALBASE}/lib/
• Convert post-stage step to post-install

[woodsb02]

Address comments from mat and adamw

[adamw]

Approved. If you want to rewrite the pkg-descr so it doesn't look so ridiculous you're approved for that too :-)

[mat]

In D6439#136555, @woodsb02 wrote:

• Add USES=localbase to replace common Makefile lines

Unless mistaken, this is not true.

[woodsb02]

In D6439#136564, @mat wrote:

In D6439#136555, @woodsb02 wrote:

• Add USES=localbase to replace common Makefile lines

Unless mistaken, this is not true.

My apologies - I mistakenly left it out. However, my test builds are working without any of these lines, so I suggest we leave it without them. Thoughts?

[woodsb02]

• Add new LIB_DEPENDS=x11-toolkits/osm-gps-map
• Fix pkg-descr

[woodsb02]

Note that portlint is giving me the following warning:
WARN: Makefile: "RUN_DEPENDS" has to appear earlier.

I have seen this before and been unable to determine what is causing that to trigger. Thoughts?

[mat]

In D6439#136603, @woodsb02 wrote:

In D6439#136564, @mat wrote:

In D6439#136555, @woodsb02 wrote:

• Add USES=localbase to replace common Makefile lines

Unless mistaken, this is not true.

My apologies - I mistakenly left it out. However, my test builds are working without any of these lines, so I suggest we leave it without them. Thoughts?

If it works without it on 9 and 10, sure, leave it out :-)

In D6439#136608, @woodsb02 wrote:

Note that portlint is giving me the following warning:
WARN: Makefile: "RUN_DEPENDS" has to appear earlier.

I have seen this before and been unable to determine what is causing that to trigger. Thoughts?

Yes, the *_DEPENDS should be a block, no empty lines between them, here, you have a blank line between LIB_DEPENDS and RUN_DEPENDS.

[woodsb02]

Put LIB_DEPENDS and RUN_DEPENDS together as one block