Date: Thu, 21 Oct 2004 11:31:59 -0700 (PDT) From: Don Lewis <truckman@FreeBSD.org> To: bt@ccgis.de Cc: freebsd-openoffice@FreeBSD.org Subject: Re: Install error OOo-1.1.3 on 4.10p3 Message-ID: <200410211832.i9LIVx8g004615@gw.catspoiler.org> In-Reply-To: <4177F54C.20508@ccgis.de>
next in thread | previous in thread | raw e-mail | index | archive | help
On 21 Oct, Benjamin Thelen wrote:
> Don Lewis wrote:
>> On 20 Oct, Benjamin Thelen wrote:
>>
>>>Hi lists,
>>>
>>>I am now really stumped and appreciate you for your assistance. I had
>>>this very error already with OOo-1.1.2 on three 4.10p2 FreeBSD boxes
>>>in the end of August, posted it to the list, but got no answer. I got
>>>along by installing a 1.1.0 package and hoped that this error will be
>>>solved in future ports.
>>>
>>>
>>>Compiling seems alright. This error-message occurs right after "make
>>>install".
>>>
>>>#
>>># UGLY hack to not have to specify a X-Display.
>>>#
>>>/usr/X11R6/bin/Xvfb :1001 -screen 0 800x600x24 > /dev/null 2>&1 & echo
>>>$! > /data/ports/editors/openoffice-1.1/work/.Xvfb.pid
>>>cd: can't cd to
>>>/data/ports/editors/openoffice-1.1/work/OOo_1.1.3_src/instsetoo/
>>>*.pro/01/normal/
>>>*** Error code 2
>>>Stop in /data/ports/editors/openoffice-1.1.
>>
>>
>> No problems here building and installing on 4.10-STABLE, though in my
>> case $DISPLAY is set so my build didn't use Xvfb.
>
> Hi Down,
> thanks for your reply.
>
> Well, I did not decide to use Xvfb. Did not know there is such a variable.
Some parts of the OpenOffice build want to access an X11 server. If you
are running the build in a terminal window under X11, $DISPLAY should be
automatically set. If $DISPLAY isn't set, the Makefile for the port
will start Xvfb, which is an X11 server for a virtual frame buffer. In
my case, $DISPLAY is set because I'm doing the build in an xterm window.
>> That shouldn't
>> matter, though because the directory passed to the cd command appears to
>> be the same in either case:
>> ${WRKSRC}/instsetoo/*.pro/${LANG_EXT}/normal/
>>
>> I wonder if you ran into some sort of build problem that didn't get
>> caught. The end of the build should look like:
>>
>> Replacing ${PRODUCTNAME} with OpenOffice.org
>> Replacing ${PRODUCTVERSION} with 1.1.3
>>
>> time needed: 0:1:0
>>
>>
>> WARNING! Project(s):
>> gtk
>>
>> not found and couldn't be built. Correct build.lsts.
>
>
>
> I now remember again, I saw that ending on a posting some time ago. No
> mine looks really different (that is how it allways looked):
>
>
> then mv -f ".deps/rulparse.Tpo" ".deps/rulparse.Po"; \
> else rm -f ".deps/rulparse.Tpo"; exit 1; \
> fi
> if gcc32 -DHAVE_CONFIG_H -I. -I. -I. -I. -I. -I./unix/linux -I./unix
> -I./unix/li nux/gnu -g -O2 -MT percent.o -MD -MP -MF
> ".deps/percent.Tpo" \ -c -o percent.o `test -f 'percent.c' || echo
> './'`percent.c; \ then mv -f ".deps/percent.Tpo" ".deps/percent.Po"; \
> else rm -f ".deps/percent.Tpo"; exit 1; \
> fi
> if gcc32 -DHAVE_CONFIG_H -I. -I. -I. -I. -I. -I./unix/linux -I./unix
> -I./unix/linux/gnu -g -O2 -MT function.o -MD -MP -MF
> ".deps/function.Tpo" \ -c -o function.o `test -f 'function.c' || echo
> './'`function.c; \ then mv -f ".deps/function.Tpo"
> ".deps/function.Po"; \ else rm -f ".deps/function.Tpo"; exit 1; \
> fi
> gcc32 -g -O2 -o dmake infer.o make.o stat.o expand.o dmstring.o
> hash.o dag.o
> dmake.o path.o imacs.o sysintf.o parse.o getinp.o quit.o state.o
> dmdump.o macpa
> rse.o rulparse.o percent.o function.o unix/libunix.a
> gmake[2]: Leaving directory
> `/data/ports/editors/openoffice-1.1/work/OOo_1.1.3
> rc/dmake'
> gmake[1]: Leaving directory
> `/data/ports/editors/openoffice-1.1/work/OOo_1.1.3_s
> rc/dmake'
> dmake has been successfully built
Your build appears to be terminating way too early. I see the matching
output about 21% of the way through my successful portupgrade run.
gcc32 -g -O2 -o dmake infer.o make.o stat.o expand.o dmstring.o hash.o dag.o
dmake.o path.o imacs.o sysintf.o parse.o getinp.o quit.o state.o dmdump.o macpa
rse.o rulparse.o percent.o function.o unix/libunix.a
gmake[2]: Leaving directory `/scratch/usr/ports/editors/openoffice-1.1/work/OOo_
1.1.3_src/dmake'
gmake[1]: Leaving directory `/scratch/usr/ports/editors/openoffice-1.1/work/OOo_
1.1.3_src/dmake'
dmake has been successfully built
build -- version: 1.86.18.2
Checking dmake...
=============
Building project x11_extensions
=============
deliver -- version: 1.50.8.8
COPY: build.lst -> /scratch/usr/ports/editors/openoffice-1.1/work/OOo_1.1.3_src/
solver/645/unxfbsd.pro/inc/x11_extensions/build.lst
>>
>> Do you still have the work directory around? If so, can you look under
>> /data/ports/editors/openoffice-1.1/work/OOo_1.1.3_src/instsetoo/ to see
>> what is missing?
>
> Yes. There are the following folders in .../instsettoo:
>
> CVS
> prj
> util
>
>
> No ".../instsettoo/*.pro/01/normal/".
This appears to get created near the end of the build.
Replacing ${PRODUCTVERSION} with 1.1.3
time needed: 0:0:59
mkdir -p ../unxfbsd.pro/01/normal
rm -f ../unxfbsd.pro/01/normal/*
rm: No match.
lzip -p . -e ../unxfbsd.pro/misc/lzip.log -l 01 -f openoffice.lst -d ../unxfbsd.
pro/01 -n OfficeOSL -e ../unxfbsd.pro/01/Logfile.txt -C ../unxfbsd.pro/01/checks
ums.txt
Local Script Particel Zip (c) 2000 Sun Microsystems
>>
>>
>> It looks like there is a problem on the next line of the Makefile,
>> though.
>>
>> do-install:
>> .if !defined(DISPLAY)
>> #
>> # UGLY hack to not have to specify a X-Display.
>> #
>> ${X11BASE}/bin/Xvfb :1001 -screen 0 800x600x24 > /dev/null 2>&1 & ${ECHO
>> } $$! > ${WRKDIR}/.Xvfb.pid
>> @sleep 5
>> @cd ${WRKSRC}/instsetoo/*.pro/${LANG_EXT}/normal/ ; \
>> ===> SAL_IGNOREXERRORS=1 TEMP=${WRKDIR} DISPLAY=${DISPLAY} ./install --prefix
>> =${PREFIX}
>> @-${CAT} ${WRKDIR}/.Xvfb.pid | ${XARGS} kill
>> @${RM} -f ${WRKDIR}/.Xvfb.pid
>>
>> I think "DISPLAY=${DISPLAY}" is incorrect because DISPLAY is not
>> defined. Judging by another place in the Makefile where Xvfb is used, I
>> think this should be "DISPLAY=${DISPLAYHACK}".
>>
>>
>
> I can see what you mean, but I'm absolutely not enough in such
> "Makefile-things". I've just changed it and building OOo again. But
> it'll take an hour. I'll tell (tomorrow).
That's way too fast ... It takes more than 13 hours on my 1.113 GHZ
Pentium III.
> But, what is obvious, that the end of OOo buildings looks so different!?
>
>
> Thanks very much,
> Ben
>
>
>
>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200410211832.i9LIVx8g004615>