One of the key advantages to Panda's package system is that any updates you make to a package are automatically downloaded and installed by users. You don't need to do anything special for this to work: simply rebuild a new package using the ppackage tool with the latest contents, using the -i parameter to point ppackage to your existing install directory. This will generate new package data and update the contents.xml file accordingly. When you then copy this install directory to your webserver host, p3d files that reference your package will automatically download and install the new contents for that package before they start to run.
This will mean all of your users will have to redownload the entire package file, which may be a large download. As a convenience to your users (and to relieve bandwidth burden on your servers), you may wish to provide patches instead. This works on the assumption that most of the package contents are not changing from one release to the next, and it allows your users who already have the previous version of a package to download only the changes necessary between the previous version and the current version, which is often a much smaller download and will save considerable time and bandwidth.
To enable this, you simply run "ppatcher -i /my/install/dir", where /my/install/dir is your install directory that you have already populated with ppackage.
The first time you run ppatcher on a particular install directory, it prepares that directory for future patches, and declares the current revision number "rev 1". You may then use ppackage to build new versions of the packages in that install directory.
The second and later times you run ppatcher, it builds patches from the last-patched version to the current version now visible in the install directory. Each time you run ppatcher, any packages in the install directory with changes since the last ppatcher will be patched and the current revision number will be incremented.
Thus, you should run ppatcher each time you are ready to make a formal release of your modified code. This will allow the users to download the minimal patches necessary to receive the latest code.
You may release intermediate versions for testing purposes without running ppatcher. If you do, users who visit your servers will have to download the entire file again, even if they already have a previous version installed. This way, these users can test your changes and report success or failure; you can continue to make new releases without running ppatcher and allow your users to continue to test them. Then, when you are satisfied that the release is ready for the world, you can run ppatcher to formalize it.
We don't recommend running ppatcher for each intermediate release, because each time you run ppatcher, a new patch file is added to the version chain. This patch file will then forever be a part of the version chain (until it is eventually aged out). Users who have an older version of the package will have to download and apply all patch files in the chain in order to update to the latest version. If there are many intermediate patch files in the chain, this process will be slower and involve more bandwidth than if there are only a few patch files.