Hi, I wasn't trying to do any particular analysis, all I was doing was to get the cygwin binaries built with least effort. I am not a cygwin guru, and I am definitely know almost nothing about libtool, configure etc.
So the stuff I said was simply what worked for me. The 3.0.3 release had copies of libtool in the rpm, but the 3.0.4 release they are not there. Good. What I will do is do a virgin install/build again (without make -i) and let you know if there are any problems appearing Richard Grevis Production Architecture Barclays Capital, Canary Wharf, London, E14 4BB *DDI : +44 (0) 20 7773 4915 * richard.grevis > -----Original Message----- > From: Carlo Marcelo Arenas Belon [mailto:[EMAIL PROTECTED] > Sent: 08 January 2007 03:22 > To: Grevis, Richard: IT (LDN) > Cc: [EMAIL PROTECTED]; ganglia-developers@lists.sourceforge.net > Subject: Re: [Ganglia-developers] Ganglia for Windows help > > > On Fri, Jan 05, 2007 at 04:20:37PM -0000, > [EMAIL PROTECTED] wrote: > > 1) Install a full cygwin system (specified on their downloader app), > > don't use the default install or any other subset. > > Why?, all you really should need is the basic stuff (going by > default) and add to it the compilers, make and the sun rpc package. > > following the `cygcheck -c` from my development workstation > which has way too many packages, because I am also using it > for development of new ganglia versions (hence why I have > autoconf, automake, m4 and libtool) and where I am testing a > native port (hence all the mingw) and some extra test cases > (hence why check). > > Cygwin Package Information > Package Version Status > _update-info-dir 00489-1 OK > alternatives 1.3.29a-1 OK > ash 20040127-3 OK > autoconf 4-1 OK > autoconf2.1 2.13-3 OK > autoconf2.5 2.60-2 OK > automake 2-1 OK > automake1.10 1.10-1 OK > automake1.4 1.4p6-4 OK > automake1.5 1.5-2 OK > automake1.6 1.6.3-3 OK > automake1.7 1.7.9-3 OK > automake1.8 1.8.5-3 OK > automake1.9 1.9.6-2 OK > base-files 3.7-1 OK > base-passwd 2.2-1 OK > bash 3.2.9-10 OK > binutils 20060817-1 OK > bzip2 1.0.3-2 OK > check 0.9.2-2 OK > coreutils 6.7-2 OK > crypt 1.1-1 OK > cygutils 1.3.1-1 OK > cygwin 1.5.23-2 OK > cygwin-doc 1.4-4 OK > diffutils 2.8.7-1 OK > e2fsprogs 1.35-3 OK > editrights 1.01-1 OK > expat 1.95.8-1 OK > file 4.19-1 OK > findutils 4.3.2-3 OK > gawk 3.1.5-4 OK > gcc 3.4.4-3 OK > gcc-core 3.4.4-3 OK > gcc-g++ 3.4.4-3 OK > gcc-mingw 20040810-1 OK > gcc-mingw-core 20050522-1 OK > gcc-mingw-g++ 20050522-1 OK > gdbm 1.8.3-8 OK > grep 2.5.1a-2 OK > groff 1.18.1-2 OK > gzip 1.3.9-1 OK > ioperm 0.4-1 OK > less 381-1 OK > libbz2_1 1.0.3-2 OK > libdb4.2 4.2.52-1 OK > libdb4.3 4.3.28-1 OK > libgdbm 1.8.0-5 OK > libgdbm-devel 1.8.3-8 OK > libgdbm3 1.8.3-3 OK > libgdbm4 1.8.3-8 OK > libiconv2 1.11-1 OK > libintl 0.10.38-3 OK > libintl1 0.10.40-1 OK > libintl2 0.12.1-3 OK > libintl3 0.14.5-1 OK > libintl8 0.15-1 OK > libltdl3 1.5.23a-2 OK > libncurses5 5.2-1 OK > libncurses6 5.2-8 OK > libncurses7 5.3-4 OK > libncurses8 5.5-3 OK > libpcre0 6.6-1 OK > libpopt0 1.6.4-4 OK > libreadline4 4.1-2 OK > libreadline5 4.3-5 OK > libreadline6 5.2.1-4 OK > libtool1.5 1.5.23a-2 OK > login 1.9-7 OK > m4 1.4.8-1 OK > make 3.81-1 OK > man 1.6d-2 OK > mingw-runtime 3.11-1 OK > mktemp 1.5-4 OK > ncurses 5.5-3 OK > perl 5.8.7-5 OK > run 1.1.10-1 OK > sed 4.1.5-2 OK > sunrpc 4.0-3 OK > tar 1.16.1-1 OK > termcap 20050421-1 OK > terminfo 5.5_20061104-1 OK > texinfo 4.8-3 OK > util-linux 2.12r-2 OK > vim 7.0.122-1 OK > w32api 3.8-1 OK > which 1.7-1 OK > zlib 1.2.3-2 OK > > indeed, installing "check" will most likely prevent you from > finishing a build, as there is a bug in libconfuse which will > be triggered if check is found at ./configure time and so we > are better NOT installing it until 3.0.5 is out (it includes > a fix for this bug). > > > 2) The ganglia sourse must come straight from the > downloaded tar file, > > and especially don't use a source tree where a ./configure has > > already > > been done on some other system. > > that applies to all platforms, indeed that is why `make > distclean` exists, but when in doubt it is better to nuke the > working directory and start fresh from pristine sources. > > > 3) find . '(' -name libtool -o -name config.cache -o -name > > config.status ')' -a -print -a -exec rm '{}' ';' > > (The pristine ganglia releases are sometimes not quite pristine > > enough). > > 3) ./configure make sure you do this within a cygwin cmd > window, not a > > DOS window.] > > be prepared to run ./configure more than once (if the > make does not > > generate .exe's. > > only one ./configure should be needed, and if the pristine > sources are not pristine enough then that is a bug that we > should be rather fixing then. I'd no problem or found any > leftovers from the packages I'd been using though (my own > snapshots from 3.0.5 and the last 3.0.4 release). > > indeed, I'd added a configure rule (which hopefully will be > out with 3.0.5 as > well) to abort the ./configure for libmetrics with an error > if "rpc/rpc.h" is not found, so that hopefully we can avoid > bogus errors because the sunrpc package wasn't installed and > fail in that case with a better error that the current one > (something cryptic in the middle of a make), so if configure > aborts you should rather close your cygwin session and get > that package installed with setup.exe (you shouldn't be > running cygwin when installing > any packages to it because of the way that windows manages open files) > > > 4) make -i, and be prepared to do this more than once. > > why? > > > 5) Only stop when you see gmond.exe exists. The make errors you see > > are sometmies for things that don't matter (hence the -i) > > i'd been unable to reproduce this problem, but again, i'd > been known by moving cygwin away and reinstall it from > scratch rather than try to work out with a damaged > installation (one that had some random scripts failing > because some cygwin application that had some cygwin file or > library open was running when setup.exe was executed). > > > 6) when you have the .exe files, remember that you need the > cygwin1.dll > > FROM the version of cygwin1.dll for the cygwin you used > to compile > > the agents (/bin/cygwin1.dll) > > ideally you should be able to use whatever cygwin1.dll that > there is in your system, and since there can be only 1 > version of it running at any time that should be better the > same one that you are using for other cygwin ported > applications you have, which should then match the one you > used to run gcc, make and the shell for compiling your > ganglia if you went that route. > > i'd been successful also installing the 3.0.0 package and > replacing the cygwin1.dll from it with the one I had > installed in my system without any obvious problems or > mismatched metrics. > > > Also note - it is only gmond and gmetad that will compile - gmetad > > will not compile under cygwin. > > gmond and gmetric, gmetad won't compile because there is no > port of rrdtool for cygwin, and even if we can make it run, > you should be insane if you think > that it is a good idea to run your master node in windows. > > Carlo > ------------------------------------------------------------------------ For more information about Barclays Capital, please visit our web site at http://www.barcap.com. Internet communications are not secure and therefore the Barclays Group does not accept legal responsibility for the contents of this message. Although the Barclays Group operates anti-virus programmes, it does not accept responsibility for any damage whatsoever that is caused by viruses being passed. Any views or opinions presented are solely those of the author and do not necessarily represent those of the Barclays Group. Replies to this email may be monitored by the Barclays Group for operational or business reasons. ------------------------------------------------------------------------
