Free Pascal releases

[BildatBoffin]

Now that you are on git you could adopt a model of main/stable. The idea is to always commit in main and to "cherry-pick" from main to stable so that you can easily release "patchs" versions, e.g only critical fixes, but at the same time you have unreleased features already standing in main.

[Thaddy]

This is already the case,  cherry picking happens by back-porting from main to fixes. It is just a matter of the terminology, the mechanism is as you described.

[dbannon]

Why not make it an open source effort so the community can contribute to it?

Oh, like the existing documentation? How's that working out?

The Lazarus Documentation ?  Getting better ! Even becoming good !  (Thanks !)

A community effort for FPC ?  We already have the wiki. And its about as good as a wiki gets. I take every opportunity to contribute something but honestly, you never expect anything coherent from a wiki, its just not the nature of the beast.

Davo

[marcov]

This is already the case,  cherry picking happens by back-porting from main to fixes. It is just a matter of the terminology, the mechanism is as you described.

cherry-pick -x to be exact, with some scripting-fu from Florian to find eligible revisions to merge.  But this already worked fine with SVN from say the later FPC 2.2 days (when SVN implemented merge tracking). Some things went better in SVN, some are more convenient in git, in practice it is about equal nowadays, except that the GIT situation is more complicated to operate, and  needs more infrastructure to implement yourself ( extra scripting running on a job to maintain the list of eligible versions).

This list is somewhat categorized here: http://www.stack.nl/~marcov/mergelogs32/

[dsiders]

Why not make it an open source effort so the community can contribute to it?

Oh, like the existing documentation? How's that working out?

The Lazarus Documentation ?  Getting better ! Even becoming good !  (Thanks !)

My point was that just shouting "open source" does ensure participation by the larger "community". FPC seems to have it better in this regard... but in the last 7 years I can count the number of contributors to Lazarus docs on one hand.

A community effort for FPC ?  We already have the wiki. And its about as good as a wiki gets. I take every opportunity to contribute something but honestly, you never expect anything coherent from a wiki, its just not the nature of the beast.

It is my experience that people are so territorial abouit their wiki pages that I cannot contribute there. I don't try any more.

[Thaddy]

Tnx, Marco, forgot the details...

[PascalDragon]

The 3.2.4 release is in preparation phase.

I'm afraid to ask, does that mean we'll see v3.2.4 in, say, less than 6 months ?

That's the hope. We still have an open point with uploading the RC releases due to a change in the way the upload/download server is accessed and after a short RC phase we'll hopefully have the release.

Making main release ready compared to the fixes branch is a totally different beast. Don't expect a new major release for another year or two.

Is it correct to presume v3.2.4 is not a main release ?

What I'm suggesting/saying here is that "formalizing" mostly bug fix releases more frequently, say maybe every year or so, may not be a bad idea.  It shows that FPC is alive.

The problem is that releases are costly (regarding time). And for 3.2.4 an additional problem had been the move from SVN to Git.

I can just imagine what a task scanning over commits trying to pick out all the things that need to be listed ! I would be happy to help if I would be useful, I am sure many other forum members would be too. Hmm, there is, apparently, a maintained 'trunk' version, maybe that helps ?

New features are documented by the developer who implemented or merged them in the mentioned New Features Trunk and then they are simply moved to the corresponding New Features X.Y.Z page. So no need to scan through commits.

But overall, "doc-ing" must be a big task, the FPC devs should consider asking for help !

People are free to file bug reports and merge requests against the documentation.

Similarly, we have not seen any posts urging people to test their code with the 3.2.4 branch. I do find that strange.

You mean like this? 🤔

[440bx]

@PascalDragon,

Thank you for the additional info.

[dbannon]

Similarly, we have not seen any posts urging people to test their code with the 3.2.4 branch. I do find that strange.

You mean like this? 🤔

Yes, like that. But that one is three years old and probably does apply very well to the current proposed 3.2.4.

Maybe you were joking ? If so, I apologize for not getting the joke.

Davo

[PascalDragon]

Similarly, we have not seen any posts urging people to test their code with the 3.2.4 branch. I do find that strange.

You mean like this? 🤔

Yes, like that. But that one is three years old and probably does apply very well to the current proposed 3.2.4.

What other 3.2.4 would there be? That things got in the way of releasing it back then is not a problem of the thread. And the request to the users did not change after all.

Maybe you were joking ? If so, I apologize for not getting the joke.

I'm absolutly serious about that thread.