Opened 19 years ago
Closed 18 years ago
#8752 closed defect (wontfix)
BUG: libtool-1.5.X / 1.9.X is broken on OS X 10.4.6 / MacBook Pro
Reported by: | armando@… | Owned by: | macports-tickets@… |
---|---|---|---|
Priority: | Normal | Milestone: | |
Component: | ports | Version: | 1.2 |
Keywords: | Cc: | markd@… | |
Port: |
Description
I've installed, cleaned out, reinstalled, retested libtool versions 1.5.X and libtool-devel (1.9.X) and both have similar behavior. Basically, the generated libtool script is completely broken, as in about 6 spots, what should be a new line is instead a "~".
The following is from libtool-devel, but the same exact thing happened in 1.5.X series. ~line 116:
# Commands used to install an old-style archive. RANLIB="ranlib" old_postinstall_cmds="\$RANLIB \$oldlib~chmod 644 \$oldlib" old_postuninstall_cmds=""
~line 234
# Commands used to build an old-style archive. old_archive_cmds="\$AR \$AR_FLAGS \$oldlib\$oldobjs\$old_deplibs~\$RANLIB \$oldlib"
~line 275
# Commands used to build a shared archive.
archive_cmds="\$CC -dynamiclib \$allow_undefined_flag -o \$lib \$libobjs \$deplibs \$compiler_flags -install_name \$rpath/\$soname \$verstring"
archive_expsym_cmds="sed -e
\"s,#.*
\" -e
\"s,[ ]*
\" -e
\"s,
(..*
),_&,
\" < \$export_symbols > \$output_objdir/\${libname}-symbols.expsym~\$CC -dynamiclib \$allow_undefined_flag -o \$lib \$libobjs \$deplibs \$compiler_flags -install_name \$rpath/\$soname \$verstring~nmedit -s \$output_objdir/\${libname}-symbols.expsym \${lib}"
# Commands used to build a loadable module if different from building
# a shared archive.
module_cmds="\$CC \$allow_undefined_flag -o \$lib -bundle \$libobjs \$deplibs\$compiler_flags"
module_expsym_cmds="sed -e
\"s,#.*
\" -e
\"s,[ ]*
\" -e
\"s,
(..*
),_&,
\" < \$export_symbols > \$output_objdir/\${libname}-symbols.expsym~\$CC \$allow_undefined_flag -o \$lib -bundle \$libobjs \$deplibs\$compiler_flags~nmedit -s \$output_objdir/\${libname}-symbols.expsym \${lib}"
... as you can see, there are problems where programs like chmod, $RANLIB, $CC, nmedit are trying to be run, seemingly, but there is a "~" instead of a new-line.
I will attach the entire file in the first comment if it lets me (this is my first time bug reporting here).
Attachments (1)
Change History (8)
Changed 19 years ago by armando@…
comment:1 Changed 19 years ago by armando@…
Just to note it somewhere, a suitable work around is:
- port install libtool (installs glibtool, glibtoolize)
- ln -s /opt/local/bin/glibtoolize /opt/local/bin/libtoolize
The system libtool works, and libtoolize is available ... so far, nothing has failed compiling with this setup (gstreamer 0.10.X, liboil, etc).
comment:2 Changed 19 years ago by armando@…
My assumption when I posted this bug is that it was "really bad." Considering it has had no updates from anyone else (who probably knows more than I do), I have to wonder if it is.
So, I have a question: is this not a major and really bad bug? If not, how are other people who use darwin ports and develop other software using libtool? Comment #2 below explains what I do to seemingly get around the issue, but I am not convinced that method is 100% correct.
Any information would be appreciated.
comment:3 Changed 18 years ago by rhwood@…
I am using the Apple provided libtool scripts instead of the DP provided scripts, which generally is a very good idea if Apple's XCode-bundled compilers are going to be used.
comment:4 Changed 18 years ago by pguyot (Paul Guyot)
Component: | base → dports |
---|
comment:5 Changed 18 years ago by markd@…
Summary: | libtool-1.5.X / 1.9.X is broken on OS X 10.4.6 / MacBook Pro → BUG: libtool-1.5.X / 1.9.X is broken on OS X 10.4.6 / MacBook Pro |
---|
comment:7 Changed 18 years ago by markd@…
Resolution: | → wontfix |
---|---|
Status: | new → closed |
So it sounds like using Apple's libtool with Apple gcc and possibly MacPorts libtool with MacPorts compilers is best.
The entire broken glibtool file from libtool-devel