On Fri, Mar 8, 2019 at 12:55 PM Juha Manninen via lazarus <[email protected]> wrote: > Everybody wants to include their custom components in LCL. Why?
Most likely because it is much easier to have a single "batteries included" installer that provides everything you need out of the box now and much more likely to be there and working in the future (since it is part of the codebase that the Lazarus devs are compiling against after every modification), than chase after packages provided by others who may or may not keep them available and up to date (keeping your own copies can solve the "available" part, but then you take unofficial ownership of the package for the "up to date" part). FWIW i tend to avoid anything that isn't part of Lazarus itself personally. The only time i used an external package was with multithreadprocs and i was very happy when i saw it become part of Lazarus itself. Besides, Lazarus already comes out of the box with a ton of packages, it isn't weird to expect that you can add more to it. Which makes me wonder, what is the goal with the bundled packages? Are they going to transition to OPM with Lazarus only providing the "bare bones", are they going to stay frozen in time or are they going to be expanded and if so, what would be the criteria for expansion? FWIW i think Python's approach sounds good here: the installers come "batteries included" with stable APIs that you can generally rely on being there in the future with no (or very minimal) changes (ignoring the Python2-to-Python3 fiasco at the moment), but you can still get more stuff through PIP and over time useful packages "migrate" from PIP to the official distribution (as long as there aren't functionality duplicates, but that is mainly because of Python's "there should be only one way to do something" and Lazarus already has several duplicates already). -- _______________________________________________ lazarus mailing list [email protected] https://lists.lazarus-ide.org/listinfo/lazarus
