devel/glib20: Update to 2.66.7

[otis]

Summary

• Update to 2.66.7
• Reformat Makefile a bit according to portclippy/portfmt

Changelog: https://fossies.org/linux/glib/NEWS

[otis]

Do not silence commands in Makefile.

[otis]

Fix some nits:

• use python 3.6+
• USES should be "=", not "+="

[diizzy]

devel/glib20/Makefile

4–5 ↗

(On Diff #83823)

Use DISTVERSION
See Porter's Handbook - "Table 5.2. Package Naming Examples"

16 ↗

(On Diff #83823)

"Initial" variables should use = not +=

21 ↗

(On Diff #83823)

Same as previous comment

22 ↗

(On Diff #83823)

3.6 is the oldest supported version of the Python 3.X branch
https://svnweb.freebsd.org/ports/head/Mk/Uses/python.mk?revision=564032&view=markup#l22

56 ↗

(On Diff #83823)

Why is this defined separately?

70 ↗

(On Diff #83823)

What's the reasoning behind stripping @ ?

[otis]

Catch some stray +=

[otis]

Catch some more nits.

[otis]

devel/glib20/Makefile

4–5 ↗

(On Diff #83823)

Use DISTVERSION
See Porter's Handbook - "Table 5.2. Package Naming Examples"

I am indifferent on this. @tcberner your thought?

70 ↗

(On Diff #83823)

What's the reasoning behind stripping @ ?

Commands should not be silenced.

[tcberner]

devel/glib20/Makefile

4–5 ↗

(On Diff #83823)

I think DISTVERSION is fine here.

[tcberner]

devel/glib20/Makefile

42 ↗

(On Diff #83826)

^ you could move these three after the options block, see https://docs.freebsd.org/en_US.ISO8859-1/books/porters-handbook/porting-order.html 15.11

While you're here, the names makes those sound like flavor settings (due to the lovercase_UPPERCASE pattern). -- maybe rename those to
_GLIB_MAN and so on.

On second though, it looks like those variables are never used at all (see rP549741):

-post-install-MANPAGES-on:
-.for m in glib gio gobject
-       ${INSTALL_MAN} ${${m}_MAN:S|^|${WRKSRC}/docs/reference/${m}/|} \
-               ${STAGEDIR}${PREFIX}/man/man1

[otis]

• Remove leftovers
• Derive _LIBVERSION from DISTVERSION

[tcberner]

devel/glib20/Makefile

35 ↗

(On Diff #83839)

^ I think this is a bit too over complicated for what it does --- I would stick with the number.

[otis]

Catch another nits.

[tcberner]

[hs.freebsd_id.omnilan.de]

Couldn't find my old PR - ever since USES= ... python, I limit to build dependency only
Is there anything I miss which requires python for run-time?

Thanks

Index: devel/glib20/Makefile
===================================================================
--- devel/glib20/Makefile       (revision 565220)
+++ devel/glib20/Makefile       (working copy)
@@ -19,7 +19,7 @@
 # iconv:wchar_t - our iconv in base doesn't support utf-8 -> wchar_t (boooo)
 # (wchar_t is used by glibmm, rawtherapee triggered this)
 USES=          compiler:c11 gettext gnome iconv:wchar_t localbase:ldflags \
-               meson perl5 pkgconfig python:3.6+ tar:xz
+               meson perl5 pkgconfig python:3.6+,build tar:xz
 USE_LDCONFIG=  yes
 USE_PERL5=     build
 USE_PYTHON=    py3kplist

[tcberner]

In D28646#641507, @hs.freebsd_id.omnilan.de wrote:

Couldn't find my old PR - ever since USES= ... python, I limit to build dependency only
Is there anything I miss which requires python for run-time?

Thanks

You would have to properly investigate, and make sure, that nothing of glib actually requires python at runtime. And then make sure, that nothing else that got its python transitively through glib misses python when removing it here.
I'm not convinced, that it is adviceable, but certainly doable.