When I install anything, I usually use sudo, at least in the final stages,
so that the end product can be installed in the right places (/etc, or /var or /usr/local as appropriate for whatever I am building). And usually, when I am building from source, only the binary products get moved to the final destination (and in fact the build directory tree is in a different location from the source directory tree), with make install moving the binaries (libraries and executables) to the appropriate places and then cleaning out intermediate build products and the source back in the original archive (actually, I would edit the makefile to ensure the last two steps, but then I can because I know how and because it is open source). On the machines on which I installed OFBiz, I had already installed the full LAMP stack, including the Perl, Java and C++ connectors, and I did so using YaST. But, at the end of the installation and setup/startup of OFBiz, I find that nothing has gone into the usual places, but it is all in root. Obviously we don't want it there, and we want a user specific to OFBiz that has only the rights of a mere mortal rather than those of root. I saw the page about running OFBiz as a server, and that talked about creating it's own user, but there was nothing about where the source and build products ought to be placed. Now, which I can do marvelous things using makefiles (including unit and integration testing), I do not know ant at all, other than that it exists. Thus, I do not know how to fix this. And, an ancillary question: Is there in the OFBiz distro that one gets as trunk using svn, the equivalent of my favourite and nearly ubiquitous makefile target: vis. 'check'. I always run 'make check' when the developers of a product (usually C++ libraries I use in my own development work) have been wise enough and considerate enough to provide that target. If they haven't, then my number one task after building the library is to begin building a proper test suite (but it is rare that I have to resort to that). One last question about where things ought to go: As I had mentioned, I already had he full LAMP stack, along with the Perl, C++ and Java connectors, installed in their default places using YaST. It isn't clear how to proceed, then, in setting up OFBiz to use MySQL. Yes, I saw the instructions to copy the java connector into the OFBiz directory tree, but the question is why, and why can it not build using the instance of the connector in the default location for it in a system that already has it installed? Thanks for your time. Cheers Ted |
In this type of project, it doesn't work the same as a C++ library or
similar. We have no idea what platform you plan to deploy on, and there are too many options to attempt a universal installation script. You can write a shell script to do an SVN checkout and build for the target server (pull the project to the server or push the project to the server). The same script could be used to remove unwanted files. If you want to remove the .svn folders from your deploy copy, you can use the SVN export command. -Adrian On 7/31/2012 9:11 PM, Ted Byers wrote: > When I install anything, I usually use sudo, at least in the final stages, > so that the end product can be installed in the right places (/etc, or /var > or /usr/local as appropriate for whatever I am building). And usually, when > I am building from source, only the binary products get moved to the final > destination (and in fact the build directory tree is in a different location > from the source directory tree), with make install moving the binaries > (libraries and executables) to the appropriate places and then cleaning out > intermediate build products and the source back in the original archive > (actually, I would edit the makefile to ensure the last two steps, but then > I can because I know how and because it is open source). > > > > On the machines on which I installed OFBiz, I had already installed the full > LAMP stack, including the Perl, Java and C++ connectors, and I did so using > YaST. > > > > But, at the end of the installation and setup/startup of OFBiz, I find that > nothing has gone into the usual places, but it is all in root. Obviously we > don't want it there, and we want a user specific to OFBiz that has only the > rights of a mere mortal rather than those of root. I saw the page about > running OFBiz as a server, and that talked about creating it's own user, but > there was nothing about where the source and build products ought to be > placed. > > > > Now, which I can do marvelous things using makefiles (including unit and > integration testing), I do not know ant at all, other than that it exists. > Thus, I do not know how to fix this. And, an ancillary question: Is there > in the OFBiz distro that one gets as trunk using svn, the equivalent of my > favourite and nearly ubiquitous makefile target: vis. 'check'. I always run > 'make check' when the developers of a product (usually C++ libraries I use > in my own development work) have been wise enough and considerate enough to > provide that target. If they haven't, then my number one task after > building the library is to begin building a proper test suite (but it is > rare that I have to resort to that). > > > > One last question about where things ought to go: As I had mentioned, I > already had he full LAMP stack, along with the Perl, C++ and Java > connectors, installed in their default places using YaST. It isn't clear > how to proceed, then, in setting up OFBiz to use MySQL. Yes, I saw the > instructions to copy the java connector into the OFBiz directory tree, but > the question is why, and why can it not build using the instance of the > connector in the default location for it in a system that already has it > installed? > > > > Thanks for your time. > > > > Cheers > > > > Ted > > > > |
> -----Original Message-----
> From: Adrian Crum [mailto:[hidden email]] > Sent: July-31-12 4:46 PM > To: [hidden email] > Subject: Re: Install location of OFBiz: still getting started ... > > In this type of project, it doesn't work the same as a C++ library or similar. We > have no idea what platform you plan to deploy on, and there are too many > options to attempt a universal installation script. > > You can write a shell script to do an SVN checkout and build for the target > server (pull the project to the server or push the project to the server). The > same script could be used to remove unwanted files. If you want to remove the > .svn folders from your deploy copy, you can use the SVN export command. > > -Adrian > Hi Adrian Thanks The same problem exists in the C++ world. There are many different configurations of a variety of distributions of Linux, not to mention the other variants of Unix and the versions of Windows. That is why, in the C++ world, they developed autoconf and automake, and a common file shipped with most libraries and Unix programs is 'configure.sh', which when finished makes various configuration files and, most importantly, the makefiles. That is a script that you use to define where everything is and where everything is to go. The fancier configure.sh scripts will actually specify which parts of the project to build depending on what resources are available. For example, the project may be able to use a png library, if it is present, but will build fine, and either doesn't build support for graphics files, or it will use another graphics file format if configure.sh finds some other graphics file it knows how to use. So, even though the project developers have no idea what your system is or how it is configured, they can make use of whatever you have installed that it makes sense for them to use, as long as you pass the appropriate info as arguments to 'configure.sh'. But that doesn't tell me where the ofbiz directory tree contents ought to go. I can see a number of candidate directories that logically ought to take at least some of the files; but if it is like some of the other servers I have installed, there will be some stuff put into (a subdirectory of) /etc, and other stuff put into /srv, and maybe /opt. And some of the files probably ought to go into /etc/init.d; but what ought to go where on a Suse Linux box? Where have others who use Suse Linux put OFBiz? Thanks Ted |
Traditionally in the unix world, optional software goes in /opt. So, I
have the following: /opt/ofbiz.1104/ /opt/ofbiz.1004/ Etc. It is up to you. Also, regarding your comparison to autoconf, etc, in the java world, it is "ant", which looks for "build.xml" files. In the ofbiz world, there are a lot of build.xml files, including a master one at the top, which is called when compiling ofbiz. On Tue, Jul 31, 2012 at 2:15 PM, Ted Byers <[hidden email]> wrote: > > -----Original Message----- > > From: Adrian Crum [mailto:[hidden email]] > > Sent: July-31-12 4:46 PM > > To: [hidden email] > > Subject: Re: Install location of OFBiz: still getting started ... > > > > In this type of project, it doesn't work the same as a C++ library or > similar. We > > have no idea what platform you plan to deploy on, and there are too many > > options to attempt a universal installation script. > > > > You can write a shell script to do an SVN checkout and build for the > target > > server (pull the project to the server or push the project to the > server). > The > > same script could be used to remove unwanted files. If you want to remove > the > > .svn folders from your deploy copy, you can use the SVN export command. > > > > -Adrian > > > Hi Adrian > > Thanks > > The same problem exists in the C++ world. There are many different > configurations of a variety of distributions of Linux, not to mention the > other variants of Unix and the versions of Windows. That is why, in the > C++ > world, they developed autoconf and automake, and a common file shipped with > most libraries and Unix programs is 'configure.sh', which when finished > makes various configuration files and, most importantly, the makefiles. > That is a script that you use to define where everything is and where > everything is to go. The fancier configure.sh scripts will actually > specify > which parts of the project to build depending on what resources are > available. For example, the project may be able to use a png library, if > it > is present, but will build fine, and either doesn't build support for > graphics files, or it will use another graphics file format if configure.sh > finds some other graphics file it knows how to use. So, even though the > project developers have no idea what your system is or how it is > configured, > they can make use of whatever you have installed that it makes sense for > them to use, as long as you pass the appropriate info as arguments to > 'configure.sh'. > > But that doesn't tell me where the ofbiz directory tree contents ought to > go. I can see a number of candidate directories that logically ought to > take at least some of the files; but if it is like some of the other > servers > I have installed, there will be some stuff put into (a subdirectory of) > /etc, and other stuff put into /srv, and maybe /opt. And some of the files > probably ought to go into /etc/init.d; but what ought to go where on a Suse > Linux box? Where have others who use Suse Linux put OFBiz? > > Thanks > > Ted > > |
> -----Original Message-----
Thanks Mike.
> From: Mike [mailto:[hidden email]] > Sent: July-31-12 5:53 PM > To: [hidden email] > Subject: Re: Install location of OFBiz: still getting started ... > > Traditionally in the unix world, optional software goes in /opt. So, I have the > following: > > /opt/ofbiz.1104/ > /opt/ofbiz.1004/ > > Etc. It is up to you. > This is useful. Did you put everything retrieved in these directories, source, and the files built from them, or did you identify and place just the binaries and configuration files there? > Also, regarding your comparison to autoconf, etc, in the java world, it is "ant", > which looks for "build.xml" files. In the ofbiz world, there are a lot of build.xml > files, including a master one at the top, which is called when compiling ofbiz. > I guess part of my problem is that I don't know ant, or what can be safely edited in the build.xml files you mention. Thanks again Ted > On Tue, Jul 31, 2012 at 2:15 PM, Ted Byers <[hidden email]> wrote: > > > > -----Original Message----- > > > From: Adrian Crum [mailto:[hidden email]] > > > Sent: July-31-12 4:46 PM > > > To: [hidden email] > > > Subject: Re: Install location of OFBiz: still getting started ... > > > > > > In this type of project, it doesn't work the same as a C++ library > > > or > > similar. We > > > have no idea what platform you plan to deploy on, and there are too > > > many options to attempt a universal installation script. > > > > > > You can write a shell script to do an SVN checkout and build for the > > target > > > server (pull the project to the server or push the project to the > > server). > > The > > > same script could be used to remove unwanted files. If you want to > > > remove > > the > > > .svn folders from your deploy copy, you can use the SVN export command. > > > > > > -Adrian > > > > > Hi Adrian > > > > Thanks > > > > The same problem exists in the C++ world. There are many different > > configurations of a variety of distributions of Linux, not to mention > > the other variants of Unix and the versions of Windows. That is why, > > in the > > C++ > > world, they developed autoconf and automake, and a common file shipped > > with most libraries and Unix programs is 'configure.sh', which when > > finished makes various configuration files and, most importantly, the > makefiles. > > That is a script that you use to define where everything is and where > > everything is to go. The fancier configure.sh scripts will actually > > specify which parts of the project to build depending on what > > resources are available. For example, the project may be able to use > > a png library, if it is present, but will build fine, and either > > doesn't build support for graphics files, or it will use another > > graphics file format if configure.sh finds some other graphics file it > > knows how to use. So, even though the project developers have no idea > > what your system is or how it is configured, they can make use of > > whatever you have installed that it makes sense for them to use, as > > long as you pass the appropriate info as arguments to 'configure.sh'. > > > > But that doesn't tell me where the ofbiz directory tree contents ought > > to go. I can see a number of candidate directories that logically > > ought to take at least some of the files; but if it is like some of > > the other servers I have installed, there will be some stuff put into > > (a subdirectory of) /etc, and other stuff put into /srv, and maybe > > /opt. And some of the files probably ought to go into /etc/init.d; > > but what ought to go where on a Suse Linux box? Where have others who > > use Suse Linux put OFBiz? > > > > Thanks > > > > Ted > > > > |
Everything is there, including the log files, source, java classes (i.e.
"binaries") and the multitude of configuration files (*.properties) and other data files (*.xml). It is quite a beast. On Tue, Jul 31, 2012 at 3:01 PM, Ted Byers <[hidden email]> wrote: > > -----Original Message----- > > From: Mike [mailto:[hidden email]] > > Sent: July-31-12 5:53 PM > > To: [hidden email] > > Subject: Re: Install location of OFBiz: still getting started ... > > > > Traditionally in the unix world, optional software goes in /opt. So, I > have the > > following: > > > > /opt/ofbiz.1104/ > > /opt/ofbiz.1004/ > > > > Etc. It is up to you. > > > Thanks Mike. > > This is useful. Did you put everything retrieved in these directories, > source, and the files built from them, or did you identify and place just > the binaries and configuration files there? > > > Also, regarding your comparison to autoconf, etc, in the java world, it > is "ant", > > which looks for "build.xml" files. In the ofbiz world, there are a lot > of build.xml > > files, including a master one at the top, which is called when compiling > ofbiz. > > > I guess part of my problem is that I don't know ant, or what can be safely > edited in the build.xml files you mention. > > Thanks again > > Ted > > > > On Tue, Jul 31, 2012 at 2:15 PM, Ted Byers <[hidden email]> > wrote: > > > > > > -----Original Message----- > > > > From: Adrian Crum [mailto:[hidden email]] > > > > Sent: July-31-12 4:46 PM > > > > To: [hidden email] > > > > Subject: Re: Install location of OFBiz: still getting started ... > > > > > > > > In this type of project, it doesn't work the same as a C++ library > > > > or > > > similar. We > > > > have no idea what platform you plan to deploy on, and there are too > > > > many options to attempt a universal installation script. > > > > > > > > You can write a shell script to do an SVN checkout and build for the > > > target > > > > server (pull the project to the server or push the project to the > > > server). > > > The > > > > same script could be used to remove unwanted files. If you want to > > > > remove > > > the > > > > .svn folders from your deploy copy, you can use the SVN export > command. > > > > > > > > -Adrian > > > > > > > Hi Adrian > > > > > > Thanks > > > > > > The same problem exists in the C++ world. There are many different > > > configurations of a variety of distributions of Linux, not to mention > > > the other variants of Unix and the versions of Windows. That is why, > > > in the > > > C++ > > > world, they developed autoconf and automake, and a common file shipped > > > with most libraries and Unix programs is 'configure.sh', which when > > > finished makes various configuration files and, most importantly, the > > makefiles. > > > That is a script that you use to define where everything is and where > > > everything is to go. The fancier configure.sh scripts will actually > > > specify which parts of the project to build depending on what > > > resources are available. For example, the project may be able to use > > > a png library, if it is present, but will build fine, and either > > > doesn't build support for graphics files, or it will use another > > > graphics file format if configure.sh finds some other graphics file it > > > knows how to use. So, even though the project developers have no idea > > > what your system is or how it is configured, they can make use of > > > whatever you have installed that it makes sense for them to use, as > > > long as you pass the appropriate info as arguments to 'configure.sh'. > > > > > > But that doesn't tell me where the ofbiz directory tree contents ought > > > to go. I can see a number of candidate directories that logically > > > ought to take at least some of the files; but if it is like some of > > > the other servers I have installed, there will be some stuff put into > > > (a subdirectory of) /etc, and other stuff put into /srv, and maybe > > > /opt. And some of the files probably ought to go into /etc/init.d; > > > but what ought to go where on a Suse Linux box? Where have others who > > > use Suse Linux put OFBiz? > > > > > > Thanks > > > > > > Ted > > > > > > > > |
Free forum by Nabble | Edit this page |