Overview
Examples
Screenshots
Comparisons
Applications
Download
Documentation
Tutorials
Bazaar
Status & Roadmap
FAQ
Authors & License
Forums
Funding Ultimate++
Search on this site
Search in forums












SourceForge.net Logo
Home » Community » U++ community news and announcements » 2020.1 rc
2020.1 rc [message #53817] Tue, 05 May 2020 11:53 Go to next message
mirek is currently offline  mirek
Messages: 14255
Registered: November 2005
Ultimate Member
Well, I guess it never will be perfect, but current state is good enough....

https://sourceforge.net/projects/upp/files/upp/2020.1rc/

If there are no serious issues, these files will become 2020.1 release in a couple of days...

Report message to a moderator

Re: 2020.1 rc [message #53820 is a reply to message #53817] Tue, 05 May 2020 21:29 Go to previous messageGo to next message
Novo is currently offline  Novo
Messages: 1430
Registered: December 2006
Ultimate Contributor
This is, probably, not very important, but still ...
I checked compilation on Linux via "umk CLANG -bsu".
Problems:
benchmarks/stdlist - 'gtk/gtk.h' file not found (weird)
benchmarks/Sort - call to 'Sort__' is ambiguous
benchmarks/CoLoop - copying variable of type 'std::atomic<int>' invokes deleted constructor
benchmarks/BenchmarkTess - crashes Clang ...
benchmarks/AllocMT - use of undeclared identifier 'Parse'; did you mean 'Parser'?
examples/SDLSoundDemo - 'begin_code.h' file not found
reference/SDL2Uword - 'begin_code.h' file not found
reference/UWord_FB - use of undeclared identifier 'K_PLUS'

Regards,
Novo

Report message to a moderator

Re: 2020.1 rc [message #53821 is a reply to message #53820] Tue, 05 May 2020 22:11 Go to previous messageGo to next message
Novo is currently offline  Novo
Messages: 1430
Registered: December 2006
Ultimate Contributor
In addition to that in Release configuration via "umk CLANG -brus".
reference/FilterStream
reference/Tuple

Regards,
Novo

Report message to a moderator

Re: 2020.1 rc [message #53823 is a reply to message #53817] Wed, 06 May 2020 13:38 Go to previous messageGo to next message
Tom1
Messages: 1301
Registered: March 2007
Ultimate Contributor
Hi Mirek,

Everything I need appears to work here with 2020.1 rc!

Thanks and best regards,

Tom

Report message to a moderator

Re: 2020.1 rc [message #53824 is a reply to message #53820] Wed, 06 May 2020 13:42 Go to previous messageGo to next message
mirek is currently offline  mirek
Messages: 14255
Registered: November 2005
Ultimate Member
Novo wrote on Tue, 05 May 2020 21:29
This is, probably, not very important, but still ...
I checked compilation on Linux via "umk CLANG -bsu".
Problems:
benchmarks/stdlist - 'gtk/gtk.h' file not found (weird)
benchmarks/Sort - call to 'Sort__' is ambiguous
benchmarks/CoLoop - copying variable of type 'std::atomic<int>' invokes deleted constructor
benchmarks/BenchmarkTess - crashes Clang ...
benchmarks/AllocMT - use of undeclared identifier 'Parse'; did you mean 'Parser'?
examples/SDLSoundDemo - 'begin_code.h' file not found
reference/SDL2Uword - 'begin_code.h' file not found
reference/UWord_FB - use of undeclared identifier 'K_PLUS'


Thanks. Might help for the next release.

That said, benchmarks are just development, problems there are OK. SDLSoundDemo / SDL2Uword - it seems like there might be a problem with your SDL2 installation.

Report message to a moderator

Re: 2020.1 rc [message #53825 is a reply to message #53824] Wed, 06 May 2020 16:58 Go to previous messageGo to next message
Novo is currently offline  Novo
Messages: 1430
Registered: December 2006
Ultimate Contributor
mirek wrote on Wed, 06 May 2020 07:42

SDLSoundDemo / SDL2Uword - it seems like there might be a problem with your SDL2 installation.

Nope. My SDL2 installation installation is fine.
I have libsdl2-dev installed.
This is a bug with SDL2 itself.
Line #include "begin_code.h" in SDL_platform.h is supposed to look like
#include <SDL2/begin_code.h>

For the time being it can be fixed by adding "/usr/include/SDL2" as an include path for POSIX platforms to the Synth package ...

Regards,
Novo

Report message to a moderator

Re: 2020.1 rc [message #53826 is a reply to message #53825] Wed, 06 May 2020 18:45 Go to previous messageGo to next message
mirek is currently offline  mirek
Messages: 14255
Registered: November 2005
Ultimate Member
That is rather looking like compiler bug. #include " is supposed to look into the current directory first.

Really weird. I have got #include "begin_code.h" in SDL_platform.h and everything works just fine, no fixes nor include paths added anywhere....

[Updated on: Wed, 06 May 2020 18:46]

Report message to a moderator

Re: 2020.1 rc [message #53827 is a reply to message #53826] Wed, 06 May 2020 19:47 Go to previous messageGo to next message
Novo is currently offline  Novo
Messages: 1430
Registered: December 2006
Ultimate Contributor
mirek wrote on Wed, 06 May 2020 12:45
That is rather looking like compiler bug. #include " is supposed to look into the current directory first.

Really weird. I have got #include "begin_code.h" in SDL_platform.h and everything works just fine, no fixes nor include paths added anywhere....

Below is what I have on my machine (Ubuntu 20.04):

#include <SDL2/_real_SDL_config.h> (SDL_config.h)
_real_SDL_config.h is located in /usr/include/x86_64-linux-gnu/SDL2/
It contains #include "SDL_platform.h", which is supposed to be located in /usr/include/x86_64-linux-gnu/SDL2/
And it can be found there. It is just a link: SDL_platform.h -> ../../SDL2/SDL_platform.h
So, compiler is looking for "begin_code.h" in /usr/include/x86_64-linux-gnu/SDL2/, but it is supposed to be looking in /usr/include/SDL2/ ...
I doubt this is a bug with Clang ...

Regards,
Novo

Report message to a moderator

Re: 2020.1 rc [message #53828 is a reply to message #53827] Wed, 06 May 2020 19:54 Go to previous messageGo to next message
Novo is currently offline  Novo
Messages: 1430
Registered: December 2006
Ultimate Contributor
Well, I checked the same project with GCC, and it is fine.
Looks like a bug with either Clang or GCC ...

Regards,
Novo

Report message to a moderator

Re: 2020.1 rc [message #53829 is a reply to message #53828] Wed, 06 May 2020 20:12 Go to previous messageGo to next message
Novo is currently offline  Novo
Messages: 1430
Registered: December 2006
Ultimate Contributor
No matter whose fault is that, SDL2 cannot be used with Clang on Linux without adding of a system-wide include path ...

Regards,
Novo

Report message to a moderator

Re: 2020.1 rc [message #53830 is a reply to message #53829] Thu, 07 May 2020 00:05 Go to previous messageGo to next message
mirek is currently offline  mirek
Messages: 14255
Registered: November 2005
Ultimate Member
Novo wrote on Wed, 06 May 2020 20:12
No matter whose fault is that, SDL2 cannot be used with Clang on Linux without adding of a system-wide include path ...


I thing you generalize too much. It works out of box in ubuntu and rapsbian.. Either way, this is not our problem to fix.

Mirek

Report message to a moderator

Re: 2020.1 rc [message #53831 is a reply to message #53830] Thu, 07 May 2020 05:56 Go to previous messageGo to next message
Novo is currently offline  Novo
Messages: 1430
Registered: December 2006
Ultimate Contributor
mirek wrote on Wed, 06 May 2020 18:05
Novo wrote on Wed, 06 May 2020 20:12
No matter whose fault is that, SDL2 cannot be used with Clang on Linux without adding of a system-wide include path ...


I thing you generalize too much. It works out of box in ubuntu and rapsbian.. Either way, this is not our problem to fix.

Mirek

This is entirely up to you. Rolling Eyes
I'm just a messenger. Rolling Eyes
I just tested U++ and found several problems.

Regards,
Novo

Report message to a moderator

Re: 2020.1 rc [message #53832 is a reply to message #53817] Thu, 07 May 2020 06:02 Go to previous messageGo to next message
Novo is currently offline  Novo
Messages: 1430
Registered: December 2006
Ultimate Contributor
Another problem. I tried to build a shared object on Linux using flag DLL and got a bunch of "recompile with -fPIC" messages.
Code wasn't compiled with -fPIC.
Sorry for reporting problems ... Embarassed

Regards,
Novo

Report message to a moderator

Re: 2020.1 rc [message #53833 is a reply to message #53832] Thu, 07 May 2020 09:14 Go to previous messageGo to next message
mirek is currently offline  mirek
Messages: 14255
Registered: November 2005
Ultimate Member
Try to add "SO" too.

Report message to a moderator

Re: 2020.1 rc [message #53834 is a reply to message #53833] Thu, 07 May 2020 18:00 Go to previous messageGo to next message
Novo is currently offline  Novo
Messages: 1430
Registered: December 2006
Ultimate Contributor
mirek wrote on Thu, 07 May 2020 03:14
Try to add "SO" too.

Thanks.
"SO" has a different purpose. It turns each module into a shared object.
"SO" allows me to compile my project, but instead of one shared object I get a whole bunch of them.
Another small issue. My DLL is named "project.dll" on Linux ... Rolling Eyes

Regards,
Novo

Report message to a moderator

Re: 2020.1 rc [message #53835 is a reply to message #53834] Thu, 07 May 2020 18:20 Go to previous message
mirek is currently offline  mirek
Messages: 14255
Registered: November 2005
Ultimate Member
Novo wrote on Thu, 07 May 2020 18:00
mirek wrote on Thu, 07 May 2020 03:14
Try to add "SO" too.

Thanks.
"SO" has a different purpose. It turns each module into a shared object.
"SO" allows me to compile my project, but instead of one shared object I get a whole bunch of them.
Another small issue. My DLL is named "project.dll" on Linux ... Rolling Eyes


I am afraid that what you need is not yet supported.

We can do that for the next release.

Mirek

Report message to a moderator

Previous Topic: stl containers support in PDB debugger
Next Topic: Changes in hashing
Goto Forum:
  

[ Syndicate this forum (XML) ] [ RSS ] [ PDF ]

Current Time: Fri Apr 25 12:52:56 CEST 2025

Total time taken to generate the page: 0.01172 seconds
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.5.
Copyright ©2001-2013 FUDforum Bulletin Board Software