Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Macports development for PowerPC (10.4, 10.5, 10.6): [unofficial] invitation to cooperate

cellularmitosis said:
Ok, I was able to bootstrap an install using this script:

Code: 
#!/bin/bash

set -e
set -x

# grab a copy of curl with moderl SSL
if ! test -e /opt/tigersh-deps-0.1 ; then
    cd /opt
    url=https://leopard.sh/binpkgs/tigersh-deps-0.1.tiger.g3.tar.gz
    curl -f $url | gunzip | tar x
fi
export PATH="/opt/tigersh-deps-0.1/bin:$PATH"

# grab the ppcports src
url=https://github.com/macos-powerpc/macports-base/archive/refs/heads/master.tar.gz
curl -fL $url | gunzip | tar x

cd macports-base-master
CFLAGS="-mcpu=750 -Os" ./configure
make
sudo make install

I made a small change to macports.tcl:

Code: 
                set tar_version [exec tar --version | head -n1]
                if [regexp { 1\.14} $tar_version] {
                    # Tiger shipped with tar 1.14.
                    set striparg "--strip-path=1"
                } else {
                    set striparg "--strip-components=1"
                }

However, now I am encountering hundreds of PortGroup errors:

Code: 
$ sudo port -v sync
--->  Updating the ports tree
Synchronizing local ports tree from rsync://macos-powerpc.org/macos-powerpc/PPCPorts/powerpc-ports.tar
receiving file list ... done
./

sent 90 bytes  received 54 bytes  41.14 bytes/sec
total size is 0  speedup is 0.00
receiving file list ... done

sent 78 bytes  received 93 bytes  68.40 bytes/sec
total size is 17889536  speedup is 104617.17
receiving file list ... rsync: [sender] change_dir "/PPCPorts/PortIndex_darwin_8_powerpc" (in macos-powerpc) failed: No such file or directory (2)
done

sent 50 bytes  received 9 bytes  23.60 bytes/sec
total size is 0  speedup is 0.00
rsync error: some files could not be transferred (code 23) at /SourceCache/rsync/rsync-35.2/rsync/main.c(1400) [receiver=2.6.9]
Command failed: /usr/bin/rsync -rtzvl --delete-after --include=/PortIndex.rmd160 --include=/PortIndex --exclude=* rsync://macos-powerpc.org/macos-powerpc/PPCPorts/PortIndex_darwin_8_powerpc/ /opt/local/var/macports/sources/macos-powerpc.org/macos-powerpc/PPCPorts/remote
Exit code: 23
Creating port index in /opt/local/var/macports/sources/macos-powerpc.org/macos-powerpc/PPCPorts/ports
Error: LimeChat: PortGroup xcode 1.0 could not be located. xcode-1.0.tcl does not exist.
Failed to parse file aqua/LimeChat/Portfile: PortGroup not found
Error: MenuMeters: PortGroup github 1.0 could not be located. github-1.0.tcl does not exist.
Failed to parse file aqua/MenuMeters/Portfile: PortGroup not found
Error: fltk: PortGroup github 1.0 could not be located. github-1.0.tcl does not exist.
Failed to parse file aqua/fltk/Portfile: PortGroup not found
Error: iTerm2: PortGroup github 1.0 could not be located. github-1.0.tcl does not exist.
Failed to parse file aqua/iTerm2/Portfile: PortGroup not found
Error: qlipper: PortGroup cmake 1.0 could not be located. cmake-1.0.tcl does not exist.
Failed to parse file aqua/qlipper/Portfile: PortGroup not found
Warning: Using the mainstream/official Qt4 portgroup
Error: qt4-mac: PortGroup xcodeversion 1.0 could not be located. xcodeversion-1.0.tcl does not exist.
Failed to parse file aqua/qt4-mac/Portfile: PortGroup not found
Error: qt4-mac-devel: PortGroup xcodeversion 1.0 could not be located. xcodeversion-1.0.tcl does not exist.
Failed to parse file aqua/qt4-mac-devel/Portfile: PortGroup not found

(snip)

Code: 
Error: xorg-server-legacy: PortGroup compiler_blacklist_versions 1.0 could not be located. compiler_blacklist_versions-1.0.tcl does not exist.
Failed to parse file x11/xorg-server-legacy/Portfile: PortGroup not found

Total number of ports parsed:    598
Ports successfully parsed:    0
Ports failed:            598
Up-to-date ports skipped:    254

Synchronizing local ports tree from rsync://rsync.macports.org/macports/release/tarballs/ports.tar.gz


I have a bit more learning to do here, to get up to speed on how MacPorts works.
Click to expand...

Could you just run port sync again? All these are mainstream portgroups, they do not exist in my repos. So perhaps they are not found, because this is the first sync, and powerpc-power are indexed first, when upstream ports are not yet indexed.

This is ugly, I agree, and should be solved somehow in a better way. But probably another port sync will do for now.

P. S. Well, obvious and quick fix would be just to duplicate all portgroups, regardless of whether they are modified or not. Not sure it is the optimal solution, but perhaps it’s okay, and better than a failing 1st sync, that will freak users out LOL
 
  • Like
Reactions: cellularmitosis
cellularmitosis said:
However, now I am encountering hundreds of PortGroup errors:
Click to expand...

I have pushed a change to include all portgroups which are used in MacPorts and re-uploaded a new ports tarball.
Hopefully this fixes the issue for now.

P. S. You will see 4 ports of mainstream ones failing to parse, that is expected (they are irrelevant for us anyway, since they require Qt5).
 
  • Like
Reactions: cellularmitosis
barracuda156 said:
I have pushed a change to include all portgroups which are used in MacPorts and re-uploaded a new ports tarball.
Hopefully this fixes the issue for now.

P. S. You will see 4 ports of mainstream ones failing to parse, that is expected (they are irrelevant for us anyway, since they require Qt5).
Click to expand...

Thanks! Setting this up on a second machine now to test this change (the first machine is churning through an apple-gcc-4.2 build)
 
  • Like
Reactions: barracuda156
Great, it is syncing now!

There was one error, it looks like PortIndex_darwin_8_powerpc isn't being generated?

Code: 
receiving file list ... rsync: [sender] change_dir "/PPCPorts/PortIndex_darwin_8_powerpc" (in macos-powerpc) failed: No such file or directory (2)
done
client: nothing to do: perhaps you need to specify some filenames or the --recursive option?
rsync error: some files could not be transferred (code 23) at /SourceCache/rsync/rsync-24.1/rsync/main.c(738)
Command failed: /usr/bin/rsync -rtzvl --delete-after --include=/PortIndex.rmd160 --include=/PortIndex --exclude=* rsync://macos-powerpc.org/macos-powerpc/PPCPorts/PortIndex_darwin_8_powerpc/ /opt/local/var/macports/sources/macos-powerpc.org/macos-powerpc/PPCPorts/remote
Exit code: 23
 
Here is an initial PR with a proposed fix for tar 1.14: https://github.com/macos-powerpc/macports-base/pull/1

However, I am not sure how to trigger the code path to test this. When I built macports-base from source, I initially didn't include this patch (to reproduce the failure) and yet for some reason it didn't trigger the failure. I need to look closer at the source and figure out an invocation which will trigger this code path.
 
ok!

Code: 
Total number of ports parsed:    1147 
Ports successfully parsed:    1147 
Ports failed:            0 
Up-to-date ports skipped:    0
 
cellularmitosis said:
Great, it is syncing now!

There was one error, it looks like PortIndex_darwin_8_powerpc isn't being generated?

Code: 
receiving file list ... rsync: [sender] change_dir "/PPCPorts/PortIndex_darwin_8_powerpc" (in macos-powerpc) failed: No such file or directory (2)
done
client: nothing to do: perhaps you need to specify some filenames or the --recursive option?
rsync error: some files could not be transferred (code 23) at /SourceCache/rsync/rsync-24.1/rsync/main.c(738)
Command failed: /usr/bin/rsync -rtzvl --delete-after --include=/PortIndex.rmd160 --include=/PortIndex --exclude=* rsync://macos-powerpc.org/macos-powerpc/PPCPorts/PortIndex_darwin_8_powerpc/ /opt/local/var/macports/sources/macos-powerpc.org/macos-powerpc/PPCPorts/remote
Exit code: 23
Click to expand...

It is expected, I just did not upload indices :)
Don’t see much need for that for the primary overlay (it is just some 500 ports vs 40k in MacPorts, costs of local syncing are negligible). Once/if I add R ports repo by default, I will see how to make indices in a way rsync picks them.

P. S. Ah, for 10.4 I am not gonna make indices anyway. If you or someone does that, we can upload those.
 
A bit off-topic, but has anyone looked at building a replacement libSystem.dylib which directly incorporates a lot of the fixes from the MacPorts Legacy Support library? It occurs to me that it might be easier to modernize the base system rather than have a bunch of workarounds in the port files.
 
cellularmitosis said:
Ok, I can look into that. It sounds like they are optional, which is great.
Click to expand...

The point of distributing index files along is to avoid indexing locally. I did not find an easy way to export port sources from a git branch while preserving time-stamps. This results in a full reindex on every port sync. That is bad, but as long as there are just a handful of ports in the overlay repo, it is tolerable (I assume that nobody with a G3 gonna sync ports daily). If the index will be supplied, then port sync becomes faster (it still takes time to extract ports archive, but no need to re-index them).
 
cellularmitosis said:
A bit off-topic, but has anyone looked at building a replacement libSystem.dylib which directly incorporates a lot of the fixes from the MacPorts Legacy Support library? It occurs to me that it might be easier to modernize the base system rather than have a bunch of workarounds in the port files.
Click to expand...

What is the use-case of a “Frankenstein Tiger” though? From what I understand, someone will use 10.4 in two cases: 1) hardware is too old and slow or b) OS is used largely to run very old native software, which does not exist for 10.5+ (games, perhaps? those may be unique and have no open-source replacement). Messing with primary system libs is likely gonna make the system less stable and less compatible, especially with proprietary apps which cannot be rebuilt against updated libs and frameworks. Which seems to defy the purpose.
Or do you mean generally, not for 10.4 in particular?
 
  • Like
Reactions: GA204
>replacement libSystem.dylib which directly incorporates a lot of the fixes from the MacPorts Legacy Support library

Macports legacysupport has a version of libsystem.dylib that re-exports symbols. Couldn't you do a similar trick to replace the original libSystem.dylib entirely (but having it re-export your renamed tigerLibSystem.dylib instead)
 
f54da said:
>replacement libSystem.dylib which directly incorporates a lot of the fixes from the MacPorts Legacy Support library

Macports legacysupport has a version of libsystem.dylib that re-exports symbols. Couldn't you do a similar trick to replace the original libSystem.dylib entirely (but having it re-export your renamed tigerLibSystem.dylib instead)
Click to expand...

And replace system libgcc_s and libstdc++ with the ones from modern gcc (not completely an irony, in principle it could be made to work).
 
barracuda156 said:
What is the use-case of a “Frankenstein Tiger” though? From what I understand, someone will use 10.4 in two cases: 1) hardware is too old and slow or b) OS is used largely to run very old native software, which does not exist for 10.5+ (games, perhaps? those may be unique and have no open-source replacement). Messing with primary system libs is likely gonna make the system less stable and less compatible, especially with proprietary apps which cannot be rebuilt against updated libs and frameworks. Which seems to defy the purpose.
Or do you mean generally, not for 10.4 in particular?
Click to expand...
Well, we have very different perspectives here
 
cellularmitosis said:
Well, we have very different perspectives here
Click to expand...

I have nothing against anyone using 10.4, of course, whether modified or not, just been wondering.

Perhaps any experience with rebuilding system components with extra stuff added in will be helpful for all systems. I wanted, for example, to enable `pthread_threadid_np` in 10.6 for ppc, and that gonna require rebuilding Libc.
 
  • Like
Reactions: cellularmitosis
Ah no worries. From my perspective, Tiger is a lot less laggy than Leopard, so I don't use Leopard unless I have to.

I was prompted to build apple-gcc42 a second time (the first round was the bootstrap version), but it looks like I hit a circular dependency:

Code: 
--->  Building cmake-bootstrap                           
--->  Staging cmake-bootstrap into destroot              
--->  Installing cmake-bootstrap @3.9.6_0                
--->  Activating cmake-bootstrap @3.9.6_0
--->  Cleaning cmake-bootstrap                           
--->  Computing dependencies for gcc7-bootstrap
The following dependencies will be installed: 
 cctools
 darwin-xtools
 gcc10-bootstrap
 gcc7-bootstrap
Continue? [Y/n]: 
Error: The following dependencies were not installed because all of them have unmet dependencies (likely due to a dependency cycle): cctools darwin-xtools gcc10-bootstrap gcc7-bootstrap
Error: Problem while installing gcc7-bootstrap
Error: Unable to execute port apple-gcc42: upgrade apple-gcc42 failed
 
cellularmitosis said:
Ah no worries. From my perspective, Tiger is a lot less laggy than Leopard, so I don't use Leopard unless I have to.

I was prompted to build apple-gcc42 a second time (the first round was the bootstrap version), but it looks like I hit a circular dependency:

Code: 
--->  Building cmake-bootstrap                       
--->  Staging cmake-bootstrap into destroot          
--->  Installing cmake-bootstrap @3.9.6_0            
--->  Activating cmake-bootstrap @3.9.6_0
--->  Cleaning cmake-bootstrap                       
--->  Computing dependencies for gcc7-bootstrap
The following dependencies will be installed:
 cctools
 darwin-xtools
 gcc10-bootstrap
 gcc7-bootstrap
Continue? [Y/n]:
Error: The following dependencies were not installed because all of them have unmet dependencies (likely due to a dependency cycle): cctools darwin-xtools gcc10-bootstrap gcc7-bootstrap
Error: Problem while installing gcc7-bootstrap
Error: Unable to execute port apple-gcc42: upgrade apple-gcc42 failed
Click to expand...

Could you show output of `port -v rdeps cctools` and `port -v rdeps apple-gcc42 -bootstrap`?

`apple-gcc42 -bootstrap` needs `cctools`, but the current version of `darwin-xtools` does not use `gcc10-bootstrap`, and then `gcc7-bootstrap` should also not be needed, AFAIU.

However, as I mentioned, this has never been tested on 10.4, and I have no idea how it is even supposed to work (in a sense I never built anything on 10.4 after it was moved to gcc14 with accompanying bootstrap ports). I guess the setup should be adjusted.

As a quick check/fix, you could replace `cctools` and `ld64` with what mainstream MacPorts has, since that setup was presumably tested. (It should not matter whether you change the logic in portfiles or simply overwrite them.)

Upd. Set a local overlay above others in /opt/local/etc/macports/sources.conf, for example as file:///opt/local/tigerports (which assumes that you will have a port for cctools at /opt/local/tigerports/devel/cctools). Place mainstream cctools port there, as well as ld64 port (both inside devel), re-run port sync.
Then check if everything works correctly (perhaps it should, since only these two ports of relevant ones deviated substantially).

P. S. To avoid syncing and re-indexing all ports, a quick alternative is to run portindex inside needed directory. For example:
cd /opt/local/tigerports
sudo portindex
 
Last edited:
  • Like
Reactions: cellularmitosis
Thanks for all of the help, ironically these problems are giving me a good education on how to use MacPorts.

Here are the rdeps with both sources enabled:

Code: 
EDIT: see next post

Code: 
EDIT: see next post

And if I comment out the macos-powerpc.org source:

Code: 
EDIT: see next post

Code: 
EDIT: see next post
 
Last edited:
wait, I was ssh'ed into the wrong machine when I ran those rdeps commands.

here they are from the machine which experienced the failure:

with both sources enabled:

Code: 
$ port -v rdeps cctools
The following ports are dependencies of cctools @949.0.1_7+xtools:
  port:apple-gcc42
    port:ld64
      port:ld64-97
        path:include/mach-o/arm/reloc.h:libmacho-headers
        path:include/libunwind.h:libunwind-headers
          port:xz-bootstrap
        path:share/doc/libgcc/README:libgcc
          port:libgcc14
            bin:xz:xz
              port:gettext
                port:ncurses
                port:gettext-tools-libs
                  port:libiconv
                    port:gperf
                  port:libtextstyle
                  port:gettext-runtime
            port:texinfo
              port:help2man
                port:perl5.34
                  port:db48
                  port:gdbm
                    port:readline
                port:p5.34-locale-gettext
              port:libunistring
                port:gmake
                  bin:lzip:lzip
                port:autoconf
                  port:m4
                port:automake
                port:libtool
            port:gcc10-bootstrap
              port:gcc7-bootstrap
                port:cctools
                  path:libexec/darwin-xtools/bin/as:darwin-xtools
                    port:cmake-bootstrap
            port:gmp
            path:lib/pkgconfig/isl.pc:isl
            port:libmpc
              port:mpfr
            port:zstd
              port:lz4
              port:zlib
    port:gcc_select

Code: 
$ port -v rdeps apple-gcc42 -bootstrap
The following ports are dependencies of apple-gcc42 @5666.3_16:
  port:apple-gcc42
    port:ld64
      port:ld64-97
        path:include/mach-o/arm/reloc.h:libmacho-headers
        path:include/libunwind.h:libunwind-headers
          port:xz-bootstrap
        path:share/doc/libgcc/README:libgcc
          port:libgcc14
            bin:xz:xz
              port:gettext
                port:ncurses
                port:gettext-tools-libs
                  port:libiconv
                    port:gperf
                  port:libtextstyle
                  port:gettext-runtime
            port:texinfo
              port:help2man
                port:perl5.34
                  port:db48
                  port:gdbm
                    port:readline
                port:p5.34-locale-gettext
              port:libunistring
                port:gmake
                  bin:lzip:lzip
                port:autoconf
                  port:m4
                port:automake
                port:libtool
            port:gcc10-bootstrap
              port:gcc7-bootstrap
                port:cctools
                  path:libexec/darwin-xtools/bin/as:darwin-xtools
                    port:cmake-bootstrap
            port:gmp
            path:lib/pkgconfig/isl.pc:isl
            port:libmpc
              port:mpfr
            port:zstd
              port:lz4
              port:zlib
    port:gcc_select

And if I just use the macports.org source:

Code: 
$ port -v rdeps cctools
The following ports are dependencies of cctools @949.0.1_3:
  port:libunwind-headers
    port:xz-bootstrap
    port:apple-gcc42
      port:ld64
        port:ld64-97
          path:include/mach-o/arm/reloc.h:libmacho-headers
          path:share/doc/libgcc/README:libgcc
            port:libgcc14
              bin:xz:xz
                port:gettext
                  port:ncurses
                  port:gettext-tools-libs
                    port:libiconv
                      port:gperf
                    port:libtextstyle
                    port:gettext-runtime
              port:texinfo
                port:help2man
                  port:perl5.34
                    port:db48
                    port:gdbm
                      port:readline
                  port:p5.34-locale-gettext
                port:libunistring
                  port:gmake
                    bin:lzip:lzip
                  port:autoconf
                    port:m4
                  port:automake
                  port:libtool
              port:gcc10-bootstrap
                port:gcc7-bootstrap
                  port:cctools
              port:gmp
              path:lib/pkgconfig/isl.pc:isl
              port:libmpc
                port:mpfr
              port:zstd
                port:lz4
                port:zlib
      port:gcc_select

Code: 
$ port -v rdeps apple-gcc42 -bootstrap
The following ports are dependencies of apple-gcc42 @5666.3_16:
  port:apple-gcc42
    port:ld64
      port:ld64-97
        path:include/mach-o/arm/reloc.h:libmacho-headers
        path:include/libunwind.h:libunwind-headers
          port:xz-bootstrap
        path:share/doc/libgcc/README:libgcc
          port:libgcc14
            bin:xz:xz
              port:gettext
                port:ncurses
                port:gettext-tools-libs
                  port:libiconv
                    port:gperf
                  port:libtextstyle
                  port:gettext-runtime
            port:texinfo
              port:help2man
                port:perl5.34
                  port:db48
                  port:gdbm
                    port:readline
                port:p5.34-locale-gettext
              port:libunistring
                port:gmake
                  bin:lzip:lzip
                port:autoconf
                  port:m4
                port:automake
                port:libtool
            port:gcc10-bootstrap
              port:gcc7-bootstrap
                port:cctools
            port:gmp
            path:lib/pkgconfig/isl.pc:isl
            port:libmpc
              port:mpfr
            port:zstd
              port:lz4
              port:zlib
    port:gcc_select

edit: wait, do I need to run port sync after commenting out a source?
 
And here are the rdeps after adding a local overlay with vanilla cctools and ld64 (and running port sync)


Code: 
$ port -v rdeps cctools           
The following ports are dependencies of cctools @949.0.1_3:
  port:libunwind-headers
    port:xz-bootstrap
    port:apple-gcc42
      port:ld64
        port:ld64-97
          path:include/mach-o/arm/reloc.h:libmacho-headers
      port:cctools
      port:gcc_select


Code: 
$ port -v rdeps apple-gcc42 -bootstrap
The following ports are dependencies of apple-gcc42 @5666.3_16:
  port:apple-gcc42
    port:ld64
      port:ld64-97
        path:include/mach-o/arm/reloc.h:libmacho-headers
        path:include/libunwind.h:libunwind-headers
          port:xz-bootstrap
    port:cctools
    port:gcc_select

edit: yes, both machines in question are G3 iBooks running Tiger.

edit2: oh, I just realized I had the macos-powerpc.org commented out, so the above was for the macports.org source, together with the local overlay (which would do nothing, as it is identical to the macports.org ports)
 
cellularmitosis said:
And here are the rdeps after adding a local overlay with vanilla cctools and ld64 (and running port sync)


Code: 
$ port -v rdeps cctools             
The following ports are dependencies of cctools @949.0.1_3:
  port:libunwind-headers
    port:xz-bootstrap
    port:apple-gcc42
      port:ld64
        port:ld64-97
          path:include/mach-o/arm/reloc.h:libmacho-headers
      port:cctools
      port:gcc_select


Code: 
$ port -v rdeps apple-gcc42 -bootstrap
The following ports are dependencies of apple-gcc42 @5666.3_16:
  port:apple-gcc42
    port:ld64
      port:ld64-97
        path:include/mach-o/arm/reloc.h:libmacho-headers
        path:include/libunwind.h:libunwind-headers
          port:xz-bootstrap
    port:cctools
    port:gcc_select
Click to expand...

Yeah, okay, this is sensible.

I suggest avoiding darwin-xtool on Tiger, at least for now. I switched to them largely for the sake of supporting ppc64 on Leopard; they work fine on 10.6 ppc too, but not required there.
The priority, IMO, should be whatever minimizes bootstrap pathway to gcc14. There is no necessity – and possibly no benefit either – in a fancy toolchain on 10.4. Since bootstrapping gcc14 on Tiger has been tested in MacPorts, it makes sense just to use that, rather than reinvent a bicycle (on ppc64 a move to xtools was actually a necessity, since MacPorts setup was totally defunct).
 
  • Like
Reactions: cellularmitosis
cellularmitosis said:
And here are the rdeps after adding a local overlay with vanilla cctools and ld64 (and running port sync)


Code: 
$ port -v rdeps cctools          
The following ports are dependencies of cctools @949.0.1_3:
  port:libunwind-headers
    port:xz-bootstrap
    port:apple-gcc42
      port:ld64
        port:ld64-97
          path:include/mach-o/arm/reloc.h:libmacho-headers
      port:cctools
      port:gcc_select


Code: 
$ port -v rdeps apple-gcc42 -bootstrap
The following ports are dependencies of apple-gcc42 @5666.3_16:
  port:apple-gcc42
    port:ld64
      port:ld64-97
        path:include/mach-o/arm/reloc.h:libmacho-headers
        path:include/libunwind.h:libunwind-headers
          port:xz-bootstrap
    port:cctools
    port:gcc_select

edit: yes, both machines in question are G3 iBooks running Tiger.

edit2: oh, I just realized I had the macos-powerpc.org commented out, so the above was for the macports.org source, together with the local overlay (which would do nothing, as it is identical to the macports.org ports)
Click to expand...

If you can confirm that using mainstream ports of cctools and ld64 (without removing macos-powerpc ports, just overriding these two) works correctly, I can amend the code in my portfiles so that for 10.4 it is identical to mainstream settings.
 
  • Like
Reactions: cellularmitosis
ok, just for the record, I changed sources so that macos-powerpc.org and macports.org are enabled, but not my overlay:

Code: 
#file:///Users/macuser/local-overlay

# MacOS PowerPC ports:
rsync://macos-powerpc.org/macos-powerpc/PPCPorts/powerpc-ports.tar

# MacPorts upstream: 
rsync://rsync.macports.org/macports/release/tarballs/ports.tar.gz [default]

and ran port sync (which took 24 mintues), and these are the rdeps:

Code: 
macuser@ibookg32(tiger)$ port -v rdeps cctools
The following ports are dependencies of cctools @949.0.1_7+xtools:
  port:apple-gcc42
    port:ld64
      port:ld64-xtools
        path:libexec/darwin-xtools/bin/ld:darwin-xtools
          port:cmake-bootstrap
    port:cctools
    port:gcc_select

macuser@ibookg32(tiger)$ port -v rdeps apple-gcc42 -bootstrap
The following ports are dependencies of apple-gcc42 @5666.3_16:
  port:apple-gcc42
    port:ld64
      port:ld64-xtools
        path:libexec/darwin-xtools/bin/ld:darwin-xtools
          port:cmake-bootstrap
    port:cctools
    port:gcc_select

I'll switch back to using the overlay and resume the build.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back