i have been busy getting accustomed to ssh and ftp and have a remote account that I am trying to comprehend.. my question is when I use ftp, ssh and remote accounts where do I download packages to.. which directory..
i have a cooledit package that is tarred and gziped which I ncftpd from metalab.unc.edu along with other packages some are docs some are man pages etc.. where do i target these files when I download them on my operating system or on my remote account.
Let's say you've opened an ftp session between your local machine and your remote server.
$ cd /some/dir
$ ftp my.remote.server.com
You then log in, and if your credentials are correct, receive some kind of prompt and can then start typing ftp commands. Your local directory will be initially set to wherever you started your session (in this case /some/dir), and the remote directory will be set to (usually) your home directory, or if the ftp server is chrooted it could be anywhere, wherever the administrator has specified.
You can find out where you are on the remote server by typing "pwd", and typing "cd" to change directory. You can type "lpwd" and "lcd" to do the same on the local filesystem.
You then use "get" to receive files, and "put" to transmit files. There are a plethora of commands, muchly dependent on the version of ftp you're using. Type "help" at the ftp prompt for details.
Make sure that you set the transfer type to binary when transfering tarballs around.
For example - to transfer /my/files/foo.tar.gz to the remote server my.remote.server.com you could do something like (this isn't real output BTW)...
$ cd /my/files
$ ftp my.remote.server.com
... login ...
ftp> cd /where/I/want/my/files/on/remote/server
ftp> put foo.tar.gz
Check out the ftp manpage (or the manpage for whatever client you use).
EDIT: It doesn't really matter where you are downloading the files to - usually /usr/local/src is where I put my tarballs - it's a matter of personal preference and site protocol. You can then unpack the tarball and read the README and INSTALL files.
I have a site on my other browser that is called www.pathname.com/fhs etc... this page is basically providing information on the fileheirarchy.. and it states that you should not place any binaries in etc.. (which I am not following to well.. it is saying to place packages in /ect/opt/package which i am not following either.. do I just place the install package in /etc/opt or the whole .tgz package. including all the readmes and so on.. if you could give a brief decription of how I should go about installing a package on my system in /etc/opt since I have been able to download them but they don't seem to work.
The tar.gz file is not what the FHS is referring to as a binary in this case. You can just copy/download the tar.gz file anywhere - /usr/local/src or /tmp will suffice. Then gunzip and untar the tarball and install per instructions (usually ./configure && make && su -c "make install" will do it- but read the README and INSTALL files contained within the tarball.)
Then, the make install step will actually copy the binaries over to /usr/local/bin (unless you specify otherwise during ./configure - manual pages etc will also be copied to the relevant locations) - which is in keeping with the FHS.
You can then delete the unpacked source if you want. This is all inline with the FHS.
i have a problem with packages that require make install installations..
i tried the syntax;
./configure && make && su -c "make install" as stated
is there a way to go about this installing the package manually like it is possible with some binary packages.
The "./configure && make && make install" three-card-trick is a generalisation - did you read the README and/or INSTALL files contained within the tarball?
Were you "cd'd" into the root of the source directory (i.e. the directory containing the configure script) when you issued the commands?
My usual procedure for installing a tarball under linux is...
$ cd /usr/local/src
$ zcat my-src-0.1.tar.gz | tar xvf -
$ cd my-src-0.1
$ more README
$ more INSTALL
$ ./configure && make && su -c "make install"
Of course, your install routine will no doubt differ. Make sure you read README and INSTALL for the correct installation procedure for your package.
If you can, try somewhere like rpmfind.net to download an rpm file for your distribution - then just use rpm -i to install it - could save you a lot of headaches.
the first time I tried i did't read the READ ME and the INSTALL, the second time I tried I had to manully load up the files and that worked great.. binary called ccrypt: put the binary itself in /usr/local and gzipped the lib docs for the binary and placed in the man1 and now it is fully operational in script form and has an info and man page. when I tried this the third time no success "error missing package". as for the samba ./configure && make && su -c "make install" this time it worked. I am still reading up on it.