Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
From: "Dave Korn" <dk AT artimi DOT com>
To: <cygwin AT cygwin DOT com>
Subject: RE: Rebuilding GDB
Date: Mon, 20 Sep 2004 18:14:42 +0100
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
In-Reply-To: <1095687843.15080.18.camel@localhost>
Message-ID: <NUTMEGFRaqbC8wSA1Xv000000c3@NUTMEG.CAM.ARTIMI.COM>

> -----Original Message-----
> From: cygwin-owner On Behalf Of Adrian Cox
> Sent: 20 September 2004 14:44

> I'm trying to build a Cygwin hosted GDB for debugging ARM and PowerPC
> boards.  After running into a lot of problems, I tried to rebuild the
> native GDB that Cygwin installed for me, and that didn't work either.
> 
> I'm up to date with Cygwin setup, and running on Win2k SP4. 
> Below is the
> end of the build where everything goes wrong:

  BTW, when discussing compile problems with gnu packages, you should always
quote the options you gave to configure (if any).

> make[3]: Entering directory `/tmp/inbuild/libgui/src'
> gcc -DHAVE_CONFIG_H -I. -I/usr/src/gdb-20030919-1/libgui/src 
> -I.. -DWIN32 -mwin3
> 2 -fwritable-strings -I/usr/include -I/usr/include 
> -I"/netrel/src/libtcltk/tk/xl
> ib" -DHAVE_NO_SEH=1 -DEXCEPTION_DISPOSITION=int   
> -I/usr/include/../unix -I/usr/
> include/../win -DTBL_VERSION=\"2.7\" -DTBL_COMMAND=\"table\" 
> -DTBL_RUNTIME=\"tkT
> able.tcl\" > -DTBL_RUNTIME_DIR=\"/usr/local/insight/share/redhat/gui\"
-DSTATIC_BU
> ILD    -g -O2 -c /usr/src/gdb-20030919-1/libgui/src/tclwinprint.c
> /usr/src/gdb-20030919-1/libgui/src/tclwinprint.c:26:22: 
> tkWinInt.h: No such file
>  or directory

  Hmm.  It should be in the source distribution itself, at
/usr/src/gdb-20030919-1/tk/win.

  Let me see compare with the command line I got when I built insight from
source last time....  I didn't use the cygwin sources, but they're currently
in fairly close sync with the repository.... I get the same warnings from
the compile as you did, but didn't get the fatal error from the missing
header.  So it should be instructive to compare the "-I" options in our
respective versions:

gcc -DHAVE_CONFIG_H -I. -I../../../../tools/gdb-6.1.1/libgui/src -I..
-DWIN32 -mwin32 -fwritable-strings -I/repository/tools/gdb-6.1.1/tcl/generic
-I/repository/tools/gdb-6.1.1/tk/generic
-I"/repository/tools/gdb-6.1.1/tk/xlib" -DHAVE_NO_SEH=1
-DEXCEPTION_DISPOSITION=int
-I/repository/tools/gdb-6.1.1/tk/generic/../unix
-I/repository/tools/gdb-6.1.1/tk/generic/../win -DTBL_VERSION=\"2.7\"
-DTBL_COMMAND=\"table\" -DTBL_RUNTIME=\"tkTable.tcl\"
-DTBL_RUNTIME_DIR=\"/usr/local/share/redhat/gui\" -DSTATIC_BUILD    -g -O2
-c ../../../../tools/gdb-6.1.1/libgui/src/tclwinprint.c

  Ok, so your one has this strange include path to
/netrel/src/libtcltk/tk/xlib which my one doesn't.  Looks like maybe you
have an older version of the tcl/tk headers in your installation and somehow
it got chosen at configure time over the ones included with the gdb source
distro?  Also, there really shouldn't be all those "-I /usr/include"s in
there.  And what on earth are "/usr/include/../win" and
"/usr/include/../unix"?  Maybe configure has somehow failed to find
something and is using /usr/include as a default, no-hope-last-chance
fallback; and then bogusly concatenating relative paths onto it that would
have been relevant if it had found the real headers but make no sense when
just hoping they're in the standard system includes dir.  Yech.  


    cheers, 
      DaveK
-- 
Can't think of a witty .sigline today....


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/