I have an app that I currently distribute using AI installer. However, the application downloads and updates itself over the internet using custom technology we developed. A handful of customers need to run the software on servers not connected to the Internet. So I would like to enable the ability for new versions to upgrade older versions. This scared me in the past so I do not change the product code when I create new versions. This prevents people from running the installer to apply updates.
My installer sets up environment variables, registry entries and Windows services. It seems like if I allow the installer to upgrade, it is going to remove all of this stuff and then reinstall it. This seems like a recipe for disaster and does not account for any tweaks users might have made to their systems. Is that how the upgrade really works?
All I really want to do is copy in the updated files as we do in our own update process. It does not seem like Windows Installer gives you any ability to fine tune this.
Mark
Suggested Articles
-
GabrielBarbu
- Posts: 2146
- Joined: Thu Jul 09, 2009 11:24 am
- Contact: Website
Re: How does Upgrade work?
Hi Mark,
Windows Installer does indeed completely remove the previous version before installing the new.
There are two workarounds for this:
1) use patches instead of automatic updates.
Advantages:
a) patches have a small size
b) patches do not remove the already installed product
Disadvantages:
a) they are very hard to manage
b) full installer has to be installed on the system already, for the patch to be applied
c) patches need to conform to the patch rules
2) use an automatic update and use the Persistent User Data wizard from the top menu to select the resources that shouldn't get removed during an upgrade, thus keeping user data intact.
Advantages:
a) no previous install of the product is needed for the updated package to be installed
b) it's easier to create compared to a patch
Disadvantages:
a) When updating, the full installer has to be downloaded. Not nice when the product has a considerable size (a few hundred Mb or a few Gb)
b) During an update, previous user data will be uninstalled (unless using the Persistent User Data wizard)
You can select whichever method suits you best. We also have tutorials in our guide that could help you with the authoring and deployment process for updates.
Best regards,
Gabriel
Windows Installer does indeed completely remove the previous version before installing the new.
There are two workarounds for this:
1) use patches instead of automatic updates.
Advantages:
a) patches have a small size
b) patches do not remove the already installed product
Disadvantages:
a) they are very hard to manage
b) full installer has to be installed on the system already, for the patch to be applied
c) patches need to conform to the patch rules
2) use an automatic update and use the Persistent User Data wizard from the top menu to select the resources that shouldn't get removed during an upgrade, thus keeping user data intact.
Advantages:
a) no previous install of the product is needed for the updated package to be installed
b) it's easier to create compared to a patch
Disadvantages:
a) When updating, the full installer has to be downloaded. Not nice when the product has a considerable size (a few hundred Mb or a few Gb)
b) During an update, previous user data will be uninstalled (unless using the Persistent User Data wizard)
You can select whichever method suits you best. We also have tutorials in our guide that could help you with the authoring and deployment process for updates.
Best regards,
Gabriel