Rolling releases Lazarus[stable or main]+FPC[main]

[rasberryrabbit]

https://gitlab.com/freepascal.org/fpc/source/-/issues/41581
https://gitlab.com/freepascal.org/fpc/source/-/issues/41558
https://gitlab.com/freepascal.org/fpc/source/-/issues/41556
https://gitlab.com/freepascal.org/fpc/source/-/issues/41554
https://gitlab.com/freepascal.org/fpc/source/-/issues/41536
https://gitlab.com/freepascal.org/fpc/source/-/issues/41480
https://gitlab.com/freepascal.org/fpc/source/-/issues/41225
https://gitlab.com/freepascal.org/fpc/source/-/issues/41087
https://gitlab.com/freepascal.org/fpc/source/-/issues/40291
https://gitlab.com/freepascal.org/fpc/source/-/issues/41353

I hope these bugs are fixed in the next version. In particular, inline bugs are patches that were developed after more than a week of consideration together with artificial intelligence. Some things probably took about a month.

Add more bug fixes.

https://gitlab.com/freepascal.org/fpc/source/-/issues/41564  Record Helper bug
https://gitlab.com/freepascal.org/fpc/source/-/issues/41591  Error position bug
https://gitlab.com/freepascal.org/fpc/source/-/issues/41598  Circular reference detection(not work under linux)
https://gitlab.com/freepascal.org/fpc/source/-/issues/41599  Internal error 200501152
https://gitlab.com/freepascal.org/fpc/source/-/issues/40874  Fix for compile error position
https://gitlab.com/freepascal.org/fpc/source/-/issues/41591  Fix for compile error position "If ... else"

[PascalDragon]

OK then.
A suggestion: release FPC 3.2.4 now. It is much better already even with the remaining bugs. Then you can fix more bugs and release FPC 3.2.6 after 6 months.
That's how it should go.

That's not how it will go, because there are outstanding fixes that need to be merged that otherwise currently block productive use on macOS which is a Tier 1 platform.

I hope these bugs are fixed in the next version. In particular, inline bugs are patches that were developed after more than a week of consideration together with artificial intelligence. Some things probably took about a month.

These are not even fixed in main, they will for sure not be fixed in 3.2.4 which is the next version.

[dbannon]

OK then.
A suggestion: release FPC 3.2.4 now. It is much better already even with the remaining bugs. Then you can fix more bugs and release FPC 3.2.6 after 6 months.
That's how it should go.

That's not how it will go, because there are outstanding fixes that need to be merged that otherwise currently block productive use on macOS which is a Tier 1 platform.

And thats symbolic of the underlying problem. Apple in particular like to change the rules mid-game. Linux did the same thing with versioning of glibc. Long release cycles just don't work.

I understand that the MacOS problem is fixed in main ?  But not, somehow merge-able into 323 ? Has the community been asked for help ? Or is there a plan to do this ?

While we look at needed fixes, can those annoying gtk2 dependencies be removed ?  Getting FPC back into the next release of Debian would be nice.

Davo

Davo

[marcov]

While we look at needed fixes, can those annoying gtk2 dependencies be removed ?  Getting FPC back into the next release of Debian would be nice.

Debian themselves insist on adding that dependency, only because FPC contains GTK2 headers.

[dbannon]

Debian themselves insist on adding that (gtk2) dependency, only because FPC contains GTK2 headers.

FPC has a gtk2 package that provides some gtk2 bindings and some not so specificy gtk2 ones (Cairo, Pango..). And the .ppu and .o files need to be generated. In Debian's case, an automatic process that makes the binaries from a debian source package, it builds binary packages for all the platforms they support.

Its a very automatic process, they cannot in any way consider an alternative. They have over 120,000 packages !

And, obviously, to generate those gtk2 .ppu and .o files, gtk2 must be installed. And Debian have, sensibly, finally decided to do away with gtk2.

I mention a solution in https://forum.lazarus.freepascal.org/index.php/topic,73405.msg575564/topicseen.html#new  A better solution would be very welcome 

Davo

[marcov]

And, obviously, to generate those gtk2 .ppu and .o files, gtk2 must be installed. And Debian have, sensibly, finally decided to do away with gtk2.

You said that before, but I find that hard to believe, what error do you get ?  I don't see why this would be the case. It is all pascal code, and no binaries are linked that would require libraries available/setup.

I mention a solution in https://forum.lazarus.freepascal.org/index.php/topic,73405.msg575564/topicseen.html#new  A better solution would be very welcome 

I also mentioned a solution there.

[PascalDragon]

And, obviously, to generate those gtk2 .ppu and .o files, gtk2 must be installed. And Debian have, sensibly, finally decided to do away with gtk2.

No, GTK2 is not necessary to build those units. It's only necessary to link binaries that use them. It's Debian's decision to require the dependency, not a requirement by FPC.

[dbannon]

And, obviously, to generate those gtk2 .ppu and .o files, gtk2 must be installed. And Debian have, sensibly, finally decided to do away with gtk2.

No, GTK2 is not necessary to build those units. It's only necessary to link binaries that use them. It's Debian's decision to require the dependency, not a requirement by FPC.

Yes PD, thats been established in another thread. Abou is aware of the situation, and agrees gtk2 is not necessary but has been unusually quiet for the last two weeks.

I have built a current Forky/testing machine, it still has FPC/Lazarus in there, still dependent in gtk2. Probably removed from experimental but the debian package servers seem to be broken right now so I cannot easily check (would involve updating my Forky to Experimental).

Davo

edit: Nope, wrong again. Still unchanged in unstable. No idea what happening ....