Page 5 of 22
Re: Blog post !
Posted: Sun Mar 14, 2021 5:38 pm
by useful
Axolotl wrote:
...
I think that too high expectations now are just as premature as concerns or fears.
...
I think everyone understands that. But everyone is having fun, too. Movement is always a good thing.
And trust is always a good thing, too. We trust Fred and the team.
We believe it will be fun, so we're all still on this forum, and some of us are also on SB
Re: Blog post !
Posted: Sun Mar 14, 2021 5:49 pm
by Keya
I also hope Fred is enjoying the REINVIGORATION of a new project
(personally I've found it difficult to work on a project year after year, I have to take my hat off to Fred for how long he's been developing Purebasic, but this really is a NEW PROJECT for Fred so hopefully he'll enjoy the fresh challenges)
Re: Blog post !
Posted: Sun Mar 14, 2021 6:12 pm
by infratec
Goodby small and fast code.
Welcome foreign CPUs.
Re: Blog post !
Posted: Sun Mar 14, 2021 6:41 pm
by Keya
infratec wrote:Goodby small and fast code.
On the contrary ... welcome a C compiler which can compile smaller and faster code than 99% of Purebasic coders can write, even in assembly. C compilers will even let you decide between small and fast code - PB can't. PB is not an optimising compiler ... C compilers are! This really will take PB code compilations up to a whole new level. Don't be afraid, this will be awesome, and give us state-of-the-art compilations.
Re: Blog post !
Posted: Sun Mar 14, 2021 7:00 pm
by Josh
Hoping, that the false positives messages from the virus scanners will be reduced too.
Re: Blog post !
Posted: Sun Mar 14, 2021 7:13 pm
by Mijikai
Josh wrote:Hoping, that the false positives messages from the virus scanners will be reduced too.
It wont fix broken AV software thats for sure.
Re: Blog post !
Posted: Sun Mar 14, 2021 7:14 pm
by StarBootics
Hello everyone,
I have a question, what will happen to code like that :
Code: Select all
Procedure RotateX(*This.Private_Members, Theta.f)
Protected Cos.f, Sin.f
!FLD dword [p.v_Theta]
!FSINCOS
!FSTP dword [p.v_Cos]
!FSTP dword [p.v_Sin]
SetLine1(*This, 1.0, 0.0, 0.0, 0.0)
SetLine2(*This, 0.0, Cos, -Sin, 0.0)
SetLine3(*This, 0.0, Sin, Cos, 0.0)
SetLine4(*This, 0.0, 0.0, 0.0, 1.0)
EndProcedure
Best regards
StarBootics
Re: Blog post !
Posted: Sun Mar 14, 2021 7:25 pm
by Keya
Also, consider this ... Apple had gone from PowerPC to Intel in just a few years, and now from Intel to ARM in just a few years ... do we really want Fred spending so much time developing a new asm platform every time Apple changes its mind on which architecture to use? (taking away time from all other development). With a C backend, ALL of those platforms can be/ARE supported. This means Fred will have more time to develop for _all of Purebasic_ rather than just Purebasic for one architecture.
Re: Blog post !
Posted: Sun Mar 14, 2021 7:31 pm
by Mijikai
Imho asm will most likely be supported in some way or the other looking at what is out there.
It aint as pretty but whatever i dont see a big issue.
And writing a lib with you fav assembler will always be possible.
Re: Blog post !
Posted: Sun Mar 14, 2021 7:35 pm
by Keya
yes, C supports inline asm. and yes, you'd still be able to compile object files and libraries in asm and include them in PB, just as you already can (C already has this capability)
Re: Blog post !
Posted: Sun Mar 14, 2021 7:43 pm
by StarBootics
Good to know about ASM supported by C compiler.
Best regards
StarBootics
Re: Blog post !
Posted: Sun Mar 14, 2021 7:47 pm
by infratec
The problem is ... C is not C.
Or, I know many codes from one C compiler which are not compilable with an other C compiler without manual tweaking.
So Fred will run into an other problem: find a C compiler which fits on the different OSs and CPUs
I think in the Linux world it will be gcc, not really the best compiler.
But that will be not a reason for me to give up PB coding
Re: Blog post !
Posted: Sun Mar 14, 2021 7:59 pm
by mk-soft
If you then programme in inline ASM, you also have to make sure that the processor knows the commands and the registers.
The instruction sets of the processors can be quite different.
Re: Blog post !
Posted: Sun Mar 14, 2021 9:19 pm
by the.weavster
infratec wrote:The problem is ... C is not C.
Or, I know many codes from one C compiler which are not compilable with an other C compiler without manual tweaking.
So Fred will run into an other problem: find a C compiler which fits on the different OSs and CPUs
As Plan A was LLVM wouldn't the most likely answer to this question be Clang?
That way you'd still indirectly be getting the advantages that made LLVM interesting in the first place.
Re: Blog post !
Posted: Sun Mar 14, 2021 9:34 pm
by Mijikai
the.weavster wrote:...
As Plan A was LLVM wouldn't the most likely answer to this question be Clang?
....
I dont see how dropping the soap and then asking for another one is helping PBs development.