[MacPorts] #62580: swig fails test suite on Lion

MacPorts noreply at macports.org
Tue Mar 30 00:43:29 UTC 2021


#62580: swig fails test suite on Lion
-------------------------------+-------------------------
  Reporter:  RobK88            |      Owner:  michaelld
      Type:  defect            |     Status:  assigned
  Priority:  Normal            |  Milestone:
 Component:  ports             |    Version:  2.6.4
Resolution:                    |   Keywords:  Lion Hamlib
      Port:  swig swig-python  |
-------------------------------+-------------------------

Comment (by kencu):

 Most likely you have done some things to make your MacPorts installation
 "atypical".

 The usual thing people do to mess this up is to use the "port select"
 mechanism to force certain binaries to be found on the PATH that really
 shouldn't be found. This causes untold wreckage with configure scripts
 that aren't expecting this, and can force weird mismatches between
 components.

 So I see this in your configure log, and I can see that such binaries are
 being found.
 {{{
 :info:configure checking for /opt/local/bin/python version... 3.9
 638     :info:configure checking for /opt/local/bin/python platform...
 darwin
 639     :info:configure checking for /opt/local/bin/python script
 directory... ${prefix}/lib/python3.9/site-packages
 640     :info:configure checking for /opt/local/bin/python extension
 module directory... ${exec_prefix}/lib/python3.9/site-packages
 641     :info:configure checking Whether to build Tcl bindings... no
 642     :info:configure checking whether to build lua binding... no
 643     :info:configure checking for swig... /opt/local/bin/swig
 644     :info:configure checking SWIG version... 4.0.2
 645     :info:configure checking for SWIG library...
 /opt/local/share/swig/4.0.2
 646     :info:configure checking whether to build bindings... yes
 }}}

 So I would -- at least temporarily -- disable all the "port select'd"
 stuff you have selected -- certainly disable all the python ones (python,
 python2, python3) and all the swig ones, as these are the ones you are
 actively using right now and that are misconfiguring -- but while you are
 at it, disable every single thing you may have "port selected" -- clang,
 gcc, whatever you have done.

 None of these ports are written to be sufficiently robust to overcome such
 forces.

 Yes, sure -- ideally engineers would pore over the Portfile code and try
 to make sure that such forces are always considered, always dealt with,
 always managed.

 however, MacPorts is a bunch of volunteers, and no amount of wishing it
 would be different will make it so. You just can browbeat, pound, or
 otherwise cajole people into spending 10,000 hours fixing up Portfiles
 that work fine for the 99% of people to accomodate the 1% who actively go
 out of their way to mess themselves up...

-- 
Ticket URL: <https://trac.macports.org/ticket/62580#comment:9>
MacPorts <https://www.macports.org/>
Ports system for macOS


More information about the macports-tickets mailing list