Hi all,

I was wondering: would it make sense to split the release schedule of the compiler and the included packages?

Compiler releases are not frequent (latest is May 2021). In the meantime, useful functionality, bug fixes and improvements have been added to the packages. To use these added and updated packages, AFAIK, there are two possibilities:

1. Use the trunk compiler (eg using fpcupdeluxe). Not recommended for production environments.
2. Use a stable compiler, manually download updated packages to a different location and make sure the compiler uses the updated packages instead of the ones included with the compiler distribution (bit of a hassle).

I think it would be very useful to have more frequent (stable) package releases, while the compiler itself can stick to the current frequency of (thoroughly tested) releases.

If I miss a smart way to use updated packages with a stable compiler, please let me know. Thanks!

Kind regards,
Peter

We are in agreement.

But it is not so easy as it may seem:
Some parts (e.g. rtti, namespaces etc.) require compiler support, a careful selection of what can be updated is needed.

If I miss a smart way to use updated packages with a stable compiler, please let me know. Thanks!

There is none better than what you described above.

We're currently attempting to get 3.2.4 out of the door, after that I plan a new debate in the FPC group about a more regular release schedule.

Several companies that use FPC have asked for a more regular release schedule, and I even received offers of funding to ensure a more regular release schedule. So you're not alone with your request.

Michael.

Having extra release burden on packages won't improve the situation. Main issue is simply that preparing a release is a lot of boring work and this does not change with the compiler excluded. If it was easy work, releases could be build more often. By excluding the compiler, not much of https://wiki.freepascal.org/Release_3.2.4 could be removed.