This is the mail archive of the cygwin mailing list for the Cygwin project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Can't configure setup.exe from CVS [was RE: setup.exe 2.427 problem with installation from rsync mirrors]

Dave Korn wrote:
This calls for a little filtering of the mirrors.txt list in setup, doesn't

I was going to suggest this:

RCS file: /cvs/cygwin-apps/setup/,v
retrieving revision 2.34
diff -p -u -p -u -r2.34
--- 25 Oct 2004 17:06:08 -0000 2.34
+++ 17 Dec 2004 15:34:37 -0000
@@ -216,7 +216,7 @@ get_site_list (HINSTANCE h, HWND owner)
      while (eol > bol && eol[-1] == '\r')
      *eol = 0;
-      if (bol[0] != '#' && bol[0] > ' ')
+      if (bol[0] != '#' && bol[0] > ' ' && strnicmp (bol, "rsync", 5))
   char *semi = strchr (bol, ';');
   if (semi)

Thanks, I applied something base on it.

but something's borked in autotools-land, and after running aclocal, autoconf
and automake, my attempts to run configure using:

../apps/setup/configure -C --enable-dependencies --disable-shared
--host=i686-pc-mingw32 --build=i686-pc-cygwin 'CC=gcc -mno-cygwin' 'CXX=g++
-mno-cygwin' --enable-maintainer-mode

end in failure:

checking for i686-pc-mingw32-efc... no
checking for i686-pc-mingw32-pgf95... no
checking for i686-pc-mingw32-lf95... no
checking for i686-pc-mingw32-gfortran... no
checking for g77... g77
checking whether we are using the GNU Fortran 77 compiler... (cached) yes
checking whether g77 accepts -g... (cached) yes
../apps/setup/configure: line 1: test: =: unary operator expected
checking the maximum length of command line arguments... 8192
checking command to parse nm output from gcc -mno-cygwin object... dk@mace

Note that (despite whatever wrapping my mailer has done) that command prompt
appears on the same line as the "checking command to parse nm output" message
immediately after the dots; something has failed so badly it's crashed right out
of the configure script without so much as an error message.

That's extremely bizarre.

What if you use the ./ script in setup's directory?

Because I've been using that successfully on a perfectly standard cygwin install since I can remember.

Is there really an invocation of test on line 1 of the configure script?

Is it possible that test is somehow running some non-cygwin command?

Diagnosing auto-* problems is a bit out of my field, but could this output
from aclocal perhaps be what's gone wrong?

dk@mace /usr/build/apps/setup> aclocal
/usr/share/aclocal/pkg.m4:5: warning: underquoted definition of
 run info '(automake)Extending aclocal'
 or see
/usr/share/aclocal/libsmi.m4:8: warning: underquoted definition of
/usr/share/aclocal/freetype2.m4:7: warning: underquoted definition of
/usr/share/aclocal/cppunit.m4:4: warning: underquoted definition of

No, those are harmless (though annoying - I've asked the pkgconfig and freetype maintainers to fix but got no reply - I guess the cppunit and libsmi maintainers need a message too).


-- Unsubscribe info: Problem reports: Documentation: FAQ:

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]