update to pymol 2.1.0 with PyQt interface support

[jwhowarth]

The attached Portfile.diff switches the build of pymol to use the new PyQt interface. The package retains the dependencies to build the legacy plugins against the tk interface so the apbs tools plugin is still usable. Note that pmw and pdb2pqr are still python 2.7-centric so that the apbs tools plugin is only usable for the default python27 variant of pymol.

[jwhowarth]

revised pmg_tk_platform.patch to adjust for upstream context changes

[jwhowarth]

Tested on 10.13 with MacPorts 2.4.3 for the default +python27 as well as the +python34, +python35 and a newly added +python36 variant.

[jwhowarth]

A few notes on how upstream has wired in the PyQt support. Pymol defaults to using PyQt if available and reverts to the Tk interface if it can't be found. However on darwin, the pymol display window only renders properly if pymol is built with the --osx-frameworks option passed to setup.py. So on darwin, a single build really can't render properly on both Tk and X11. If the packaging were enhanced to support a -x11 variant that would have to be built without the --osx-frameworks setup.py option. Also, since pymol now defaults to using PyQt if available, such a -x11 variant would have to explicity pass '-N pmg_tk' to the shell script that starts up pymol in order to suppress the use of PyQt.

[jwhowarth]

Note that the 'legacy plugins' in pymol 2.1.0 don't use PyQt. Hence the need to build with all of the previous tcl/tk dependencies in order to support running those legacy plugins.

[jwhowarth]

Note that you can test the apbstools plugin functionality for the default python27 variant of pymol by installing apbs and pdb2pqr and following the instructions at ​https://pymolwiki.org/index.php/Apbsplugin under the section 'Example 1'.

[jwhowarth]

Added notes that Pymol can be run in the classic Tk GUI by appending the '-N pmg_tk' runtime option.

[g5pw (Aljaž Srebrnič)]

Hello! from a quick glance the submission looks OK. in the file pdb2pqr.patch​, the prefix is hardcoded to /opt/local (and it shouldn't be). Use a placeholder like @@PREFIX@@ and replace it inside a reinplace block in the portfile.

[jwhowarth]

new pdb2pqr.patch to allow apbs-tools plugin to find required pdb2pqr.py from MacPorts pdb2pqr package

[jwhowarth]

Updated pdb2pqr.patch to remove hard-coded patch with placeholder and Portfile.diff to substitute it post-patch.

[jwhowarth]

Tweaked Portfile to make sure all the installed python scripts are compiled and also compile them optimized as well. Specifically, the current packaging misses compiling files in those directories like 'modules' which are manually installed.

[jwhowarth]

Fix pymol_shell.diff to properly set PYMOL_PATH=@@PYTHON_PKGDIR@@/pymol/pymol_path

[jwhowarth]

Properly set PYMOL_PATH=@@PYTHON_PKGDIR@@/pymol/pymol_path in pymol_shell.diff so that redundant installations of data, modules, examples, test and scripts subdirectories can be dropped.

[reneeotten (Renee Otten)]

thanks jwhowarth for keeping the pymol port up-to-date! The Portfile and patches in this ticket work perfectly for me - I like the PyQt interface. Anyway, while you're changing things perhaps you can also make the livecheck work - for me adding livecheck.regex v(\[0-9.\]+).tar.bz2 makes it behave as expected.

[jwhowarth]

Add lines for svn livecheck...

$ sudo port -d livecheck  pymol
DEBUG: Copying /Users/howarth/Library/Preferences/com.apple.dt.Xcode.plist to /opt/local/var/macports/home/Library/Preferences
DEBUG: Changing to port directory: /Users/howarth/ports/science/pymol
DEBUG: OS darwin/17.5.0 (Mac OS X 10.13) arch i386
DEBUG: Re-registering default for build.cmd
DEBUG: Re-registering default for destroot.cmd
DEBUG: Re-registering default for destroot.destdir
DEBUG: Sourcing PortGroup python 1.0 from /opt/local/var/macports/sources/rsync.macports.org/macports/release/tarballs/ports/_resources/port1.0/group/python-1.0.tcl
DEBUG: Sourcing PortGroup active_variants 1.1 from /opt/local/var/macports/sources/rsync.macports.org/macports/release/tarballs/ports/_resources/port1.0/group/active_variants-1.1.tcl
DEBUG: Sourcing PortGroup active_variants 1.1 from /opt/local/var/macports/sources/rsync.macports.org/macports/release/tarballs/ports/_resources/port1.0/group/active_variants-1.1.tcl
DEBUG: Sourcing PortGroup compilers 1.0 from /opt/local/var/macports/sources/rsync.macports.org/macports/release/tarballs/ports/_resources/port1.0/group/compilers-1.0.tcl
DEBUG: universal_variant is false, so not adding the default universal variant
DEBUG: Executing variant python27 provides python27
DEBUG: Running callback portconfigure::add_automatic_compiler_dependencies
DEBUG: Finished running callback portconfigure::add_automatic_compiler_dependencies
DEBUG: Running callback portbuild::add_automatic_buildsystem_dependencies
DEBUG: Finished running callback portbuild::add_automatic_buildsystem_dependencies
DEBUG: Starting logging for pymol
DEBUG: Executing org.macports.main (pymol)
DEBUG: livecheck phase started at Sat Apr 21 15:57:22 EDT 2018
DEBUG: Executing org.macports.livecheck (pymol)
DEBUG: Port (livecheck) version is 4187
DEBUG: Fetching http://svn.code.sf.net/p/pymol/code/trunk/pymol
DEBUG: The regex is "Revision (\d+):"
DEBUG: The regex matched "Revision 4187:", extracted "4187"
DEBUG: The regex matched "Revision 4187:", extracted "4187"
pymol seems to be up to date

[jwhowarth]

Note that release version doesn't mean a whole lot upstream as the Pymol open source version gets lax treatment with regard to tagging release branches and religiously creating release tar balls.

[jwhowarth]

Portfile diff for support of PyQt interface in pymol

[jwhowarth]

Add support for running the pymol testsuite (which produces the same results as seen for the debian linux pymol packaging).

[g5pw (Aljaž Srebrnič)]

Is the or in the second chunk of ​pmg_tk_platform.patch intentional?

[jwhowarth]

The rationale of the pmg_tk_platform.patch changes is that python always reports linux2 on modern linux releases and no longer uses the hack for pymol under Gnome and KDE. Since we ware using the same X11 server as a modern linux release, there is no reason to keep using that hack on Darwin.

[g5pw (Aljaž Srebrnič)]

Oh yeah, I get the reason, I was just wondering if there was a typo (since all the other instance use the "and" operator, but I didn't see the code so it may be correct, of course.

[jwhowarth]

The actual change goes back to 2008with the commit...

[​https://sourceforge.net/p/pymol/code/3451/

[jwhowarth]

FYI, I added the pmg_tk_platform.patch to track the behavior of the linux build on Darwin as closely as possible as upstream really only tests the open source builds on linux AFAIK. So are we okay for the commit now?

[g5pw (Aljaž Srebrnič)]

Commited in [00a4f52bcba6decfdb3ff9e0b5a69f81ccebf5fe/macports-ports] and [753d76539512c705a501d4934cd2cd74297aa9b4/macports-ports] (sorry for the double commit). Thanks for your contribution to MacPorts!