search

MichaelChirico / r-bugs

A ⚠️read-only⚠️mirror of https://bugs.r-project.org/
20 stars 0 forks source link

[BUGZILLA #14998] Suggestion for two internal self-reference in "Writing R Extensions" manual #4586

Open MichaelChirico opened 4 years ago

MichaelChirico commented 4 years ago

Dear Brian,

Thanks for your email.

The links are internal links to your own documenation.

So your own documentation is evidence itself.

What more do you need?

The situation is certainly not unique to Solaris, which is a more robust version of Unix than is Linux.

Furthermore, the same gnu compiliers are used to build R in both Linux and Solaris. So the situation is common to all Unix installations of R.

Sincerely,

William Wheeler

--------------------------------------------------------------------------------

On Thu, 26 Jul 2012, r-bugs@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::>-project.org wrote:

https://bugs.r-project.org/bugzilla3/show_bug.cgi?id=14998

Brian Ripley <ripley@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::>> changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
Status|NEW                         |UNCONFIRMED
Ever Confirmed|1                           |0

--- Comment #1 from Brian Ripley <ripley@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::>> 2012-07-26 07:46:46
EDT ---
No evidence is provided for the claims, and usage in CRAN packages suggests
that the claims are false.  Perhaps they reflect quirks of a particular
linker
(more than one is on use on Solaris) with a non-standard build of R (a
standard
build does not have libR.so).

And in any case, we do not want to clutter up description of the API with how
to use it.

Please do study the section 'BUGS" in the R FAQ.

-- 
Configure bugmail:
https://bugs.r-project.org/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You reported the bug.

METADATA

MichaelChirico commented 4 years ago

May I please suggest inserting the following sentence into Sections 6.7 and 6.7.1 of the "Writing R Extensions" manual in order to make it easier to figure out how to link to the statistical subroutines.

See 6.16 Using these functions in your own C code and 8.1.1 Compiling against the R library for technical details.

===============

There appear to be some nontrivial differences about what is required to link to the various cumulative distribution function routines.

For instance, for pnormal, it is possible to link to libR.so using only the R.h and Rmath.h header files and get a working program.

But for pgamma, one needs to link to the libRmath library and use the line

define MATHLIB_STANDALONE line before the #include <R.h> and #include

lines in order to get a working program.

METADATA

MichaelChirico commented 4 years ago

No evidence is provided for the claims, and usage in CRAN packages suggests that the claims are false. Perhaps they reflect quirks of a particular linker (more than one is on use on Solaris) with a non-standard build of R (a standard build does not have libR.so).

And in any case, we do not want to clutter up description of the API with how to use it.

Please do study the section 'BUGS" in the R FAQ.

METADATA