Compiling Qt6 project on Linux Mint 20 (Ubuntu 20.04)

[Thaddy]

It is not obsolete, but Qt6 is a bit more stable, at least to me. It is an observation/opinion, not a fact.
This opinion is merely reached by all Wayland issues all over the forum.
(Half of which caused by not understanding Wayland)

[robert rozee]

I wish the Lazarus Project options Compile screen just had the choices (regardless of the Lazarus Project and Ubuntu versions):

Code: Text  [Select][+][-]

1. Make compatible with:
2.     [x] Glibc_2.31  [ ] Glibc_2.34  [ ] ....
3.     [x] libQt5pas 1.2.6  [ ] libQt5pas 1.2.8 .....
4.     [ ] libQt6pas ....

I suppose it all relies on too many system files for that to be possible though.

Yes, I agree its less than desirable !  With the GLib issue Its the same with other development platforms and, apparently, no one else has come up with a better solution. A small team here tried very hard to find a general solution but it seemed too complicated to me, search the forum.

since someone else has conveniently resurrected this thread... see here:
https://forum.lazarus.freepascal.org/index.php/topic,64731.0.html

"attached is a patch for FPC 3.2.2 [...] to enforce glibc BASE symbol versioning. this is 'yet another approach', where the compiler itself is patched to substitute BASE versioned glibc symbols for unversioned glibc symbols."

the patch is extremely simple, and there is no good reason why it should not be included in the .deb distributions of FPC for Linux. it completely solves the glibc symbol versioning problem. once applied to FPC, it can then be used to create a version of the Lazarus IDE's elf binary that is glibc symbol version agnostic.

i created this patch around 2 and a half years ago, and yet still the muppets in charge have not rolled it out   


cheers,
rob   :-)

[Thaddy]

i created this patch around 2 and a half years ago, and yet still the muppets in charge have not rolled it out   

Then create a patch that works and with tests. Those are usually accepted, even without discusson (I know, had to learn that too)
You did not do either, so don't complain.

Wallace and Gromit.

[Fred vS]

the patch is extremely simple, and there is no good reason why it should not be included in the .deb distributions of FPC for Linux. it completely solves the glibc symbol versioning problem. once applied to FPC, it can then be used to create a version of the Lazarus IDE's elf binary that is glibc symbol version agnostic.

i created this patch around 2 and a half years ago, and yet still the muppets in charge have not rolled it out   

And if people dont want to apply your patch, you may use fpc-ootb.

[Thaddy]

(The patch was bad Fred. Not your fault, nor mine)

[Fred vS]

(The patch was bad Fred. Not your fault, nor mine)

fpc-ootb does not use Rob's fix; it directly assigns the correct version of the glibc symbols used by fpc since its inception to each unit using glibc.

Furthermore, it has been thoroughly tested on both older and newer operating systems.

[robert rozee]

The patch was bad Fred. Not your fault, nor mine

Thaddy: the patch is just a few lines of code, nothing complicated. are you capable of coming up with a better solution?

yes - i am putting this forward as a challenge. show us how you would do it, if you indeed can.


rob.

[Thaddy]

@robert rozee

I have over 80% commit rate of my suggested contributions to compiler and rtl. I don't have to prove myself.
They are small, but quite a few over the years.

If you suggest a Qt6 patch, just ask, but I won't do any work on a deprecated GTK2.

[robert rozee]

I have over 80% commit rate of my suggested contributions to compiler and rtl. I don't have to prove myself.
[...]
If you suggest a Qt6 patch, just ask, but I won't do any work on a deprecated GTK2.

Thaddy: the patch, applied to FPC, is to enforce glibc BASE symbol versioning, the fact that this (or some variation thereof) is needed to get Qt6 over the line is secondary.

from your response i assume that you lack the ability to create the same functionality yourself; as i said before, the patch is quite short and straightforward. before you go claiming "the patch was bad", you need to be able to (a) provide evidence as to why it is 'bad', and (b) provide a 'non-bad' alternative.


rob.