Wiki improvements

Forum for discussion about the documentation project.
Post Reply
fxm
Moderator
Posts: 12082
Joined: Apr 22, 2009 12:46
Location: Paris suburbs, FRANCE

Re: Wiki improvements

Post by fxm »

@CoderJeff:

A last point for SCREENCONTROL (comparing commit text and fbc source)

In fbc 64-bit:
dim param as longint = &h80000000
initializes 'param' with +2147483648 but not -2147483648
What to choose as initializer value (rather as in the fbc source I guess)?
(but it is probably without consequences)
coderJeff
Site Admin
Posts: 4313
Joined: Nov 04, 2005 14:23
Location: Ontario, Canada
Contact:

Re: Wiki improvements

Post by coderJeff »

The only places the default value of &H80000000 is checked:
- SET_WINDOW_POS ... don't set X or Y position if parameter value is &H80000000
- SET_PEN_POS ... don't set X or Y position if parameter value is &H80000000
- SET_ALPHA_PRIMITIVES ... no change if parameter value is &H80000000

For all other SET_* functions, the parameter value is used as-is.

For all GET_* operations, parameter values (including &H8000000) have no effect.
fxm
Moderator
Posts: 12082
Joined: Apr 22, 2009 12:46
Location: Paris suburbs, FRANCE

Re: Wiki improvements

Post by fxm »

@Jeff,

changelog.txt:
Version 1.09.0
[changed]
- rtlib: freebsd: minimum thread stacksize 8192 KiB
Does this mean that the new default value of the '-t < stack size >' compiler option has been changed to 8192 KBytes for the freebsd platform? , for 32-bit? , for 64-bit?
MrSwiss
Posts: 3910
Joined: Jun 02, 2013 9:27
Location: Switzerland

Re: Wiki improvements

Post by MrSwiss »

It clearly states: thread stack which is different from application stack, which the -t switch addresses.
coderJeff
Site Admin
Posts: 4313
Joined: Nov 04, 2005 14:23
Location: Ontario, Canada
Contact:

Re: Wiki improvements

Post by coderJeff »

fxm wrote:changed to 8192 KBytes for the freebsd platform? , for 32-bit? , for 64-bit?
For the stack size in ThreadCreate

I just noticed my units are wrong - should be 8192 BYTES minimum for FreeBSD. Anything less was causing threading to crash.

On other platforms, the minimum is decided by the pthreads library at the time that fbrt library is compiled.
fxm
Moderator
Posts: 12082
Joined: Apr 22, 2009 12:46
Location: Paris suburbs, FRANCE

Re: Wiki improvements

Post by fxm »

Posts regarding the new '-entry' compiler option have been moved to a new topic in the Documentation forum:
Compiler Option: -entry
fxm
Moderator
Posts: 12082
Joined: Apr 22, 2009 12:46
Location: Paris suburbs, FRANCE

Re: Wiki improvements

Post by fxm »

Updated documentation page:
  • KeyPgOpProcptr → fxm [added syntax for PROCPTR( identifier, type )]
    • to allow getting procedure pointer for based on sub/function type (for fbc version >= 1.09).
fxm
Moderator
Posts: 12082
Joined: Apr 22, 2009 12:46
Location: Paris suburbs, FRANCE

Re: Wiki improvements

Post by fxm »

@Jeff,

I understood through the latest changes (fbc-1.09 08/15/2021) for PowerPC that the documentation needs to be updated as follows:
- On the Compiler Option: -arch page, two new values to add for PowerPC: 'powerpc' and 'powerpc64'.
- New __FB_PPC__ (KeyPgDdfbppc) page to create for PowerPC, as well as all the corresponding links to add to the various pages concerned.

Is it right (I am ready to do it) ?
coderJeff
Site Admin
Posts: 4313
Joined: Nov 04, 2005 14:23
Location: Ontario, Canada
Contact:

Re: Wiki improvements

Post by coderJeff »

fxm wrote:Is it right (I am ready to do it) ?
You got it right. Thank-you.
fxm
Moderator
Posts: 12082
Joined: Apr 22, 2009 12:46
Location: Paris suburbs, FRANCE

Re: Wiki improvements

Post by fxm »

fxm wrote: - On the Compiler Option: -arch page, two new values to add for PowerPC: 'powerpc' and 'powerpc64'.
- New __FB_PPC__ (KeyPgDdfbppc) page to create for PowerPC, as well as all the corresponding links to add to the various pages concerned.
Done:
- CompilerOptarch → fxm [added two new values for PowerPC]
- KeyPgDdfbppc → fxm [new page '__FB_PPC__']
- CatPgDddefines → fxm [added link to new page '__FB_PPC__']
- CatPgFullIndex → fxm [added link to new page '__FB_PPC__']
- CatPgFunctIndex → fxm [added link to new page '__FB_PPC__']
- PrintToc → fxm [added link to new page '__FB_PPC__']
coderJeff
Site Admin
Posts: 4313
Joined: Nov 04, 2005 14:23
Location: Ontario, Canada
Contact:

Re: Wiki improvements

Post by coderJeff »

fxm wrote: - CompilerOptarch → fxm [added two new values for PowerPC]
if you get to it before me (changelog.txt probably looks more complicated than it really is):
- '-arch powerpc' - big endian
- '-arch powerpc64' - big endian
- '-arch powerpc64le' - little endian
- '__FB_BIGENDIAN__' is defined for 'powerpc|powerpc64' (currently the only two arch's that define it)
fxm
Moderator
Posts: 12082
Joined: Apr 22, 2009 12:46
Location: Paris suburbs, FRANCE

Re: Wiki improvements

Post by fxm »

- CompilerOptarch → fxm [added new value 'powerpc64le']
- CompilerOpttarget → fxm [added new values and shortnames for PowerPC]

Apparently everything else is already up to date.
fxm
Moderator
Posts: 12082
Joined: Apr 22, 2009 12:46
Location: Paris suburbs, FRANCE

Re: Wiki improvements

Post by fxm »

fxm wrote:Proposed warning to add at description paragraph end of SUB and FUNCTION pages:
Warning for 64-bit compiler only: Regarding the choice of user procedure identifier names, an additional restriction to the normal Identifier Rules also excludes in the global namespace all register names or other symbols from the assembler format used, because causing assembler errors or runtime bugs (due to the 64-bit compiler not decorating procedure names).
Done:
- KeyPgSub → fxm [added warning on choosing user procedure identifier names for 64-bit compiler only]
- KeyPgFunction → fxm [added warning on choosing user procedure identifier names for 64-bit compiler only]
SARG
Posts: 1756
Joined: May 27, 2005 7:15
Location: FRANCE

Re: Wiki improvements

Post by SARG »

As not everyone knows the assembler I wonder if a list of the forbidden words should not be added.
fxm
Moderator
Posts: 12082
Joined: Apr 22, 2009 12:46
Location: Paris suburbs, FRANCE

Re: Wiki improvements

Post by fxm »

An exhaustive list of forbidden register and symbol names (for 'intel', 'att' assemblers) must be precise and exact, so everything should be tested !
Post Reply