[MacPorts] #71680: gnuregex fails on powerpc: symbol dyld_stub_binding_helper not defined for architecture ppc

MacPorts noreply at macports.org
Mon Dec 30 23:25:23 UTC 2024


#71680: gnuregex fails on powerpc: symbol dyld_stub_binding_helper not defined for
architecture ppc
---------------------------+--------------------
  Reporter:  barracuda156  |      Owner:  (none)
      Type:  defect        |     Status:  new
  Priority:  Normal        |  Milestone:
 Component:  ports         |    Version:  2.10.5
Resolution:                |   Keywords:
      Port:  gnuregex      |
---------------------------+--------------------

Comment (by kencu):

 Replying to [comment:5 ryandesign]:
 > Replying to [comment:1 kencu]:
 > > you probably need to set the deployment target to 10.5 due to 10,6-ppc
 idiosyncracies.
 >
 > What makes you suggest that? Can you reference a prior bug report in
 which that was the solution?

 I based that initial suggestion (done from bed, BTW, on my ipad  this
 morning)  on my experience two (or was it three) years ago when I
 originally did the work to get macports, the newer gcc toolchains,
 cctools, and ld64 working on 10.6-for-PPC and then donated that to the
 group working on it over at MacRumours with the suggestion they could take
 that, fork MacPorts, and run with a set of tools that might get them some
 software to run on that system.

 (Up until then, they had been using MacPorts-built ports on 10.5 PPC and
 copying them onto 10.6 PPC.)

 In so doing, I became quite familiar with the idiosyncracies fo the
 10.6-PPC system installation, which had a few features of 10.6, but was
 mostly 10.5 PPC with a different version number on it (and AFAIK, still is
 that).

 So the solution to a lot of problems was to tell the system it was
 building on 10.5, and it would then find the symbols it expected in the
 places they usually existed.

 Based on reading about  the trivial fix that made this port work on 10.5
 PPC (it was defaulting to 10.6, and telling it that it was really on 10.5
 fix the build) I expected to find the symbols for dyld_stub_binding_helper
 were not in libSystem.dylib on 10.5 (or on 10.6-PPC).

 But surprise me, they were in libSystem on 10.5 when I went looking for
 them on my 10.5 PPC system as above.

 I boiled and recylced my 10.6 PPC drive as i had much better things to do
 that keep wasting time on that (ie I installed Linux on it).

 So that is why.

 Deep, intimate knowledge of the guts of the system led to the suggestion,
 but actual experimentation led to a deeper question.

 Kind of like my undergrad physics program, with the theorists and the
 experimenters, come to think of it.

-- 
Ticket URL: <https://trac.macports.org/ticket/71680#comment:7>
MacPorts <https://www.macports.org/>
Ports system for macOS


More information about the macports-tickets mailing list