search

MichaelChirico / r-bugs

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

[BUGZILLA #575] missing include in pnorm.c #695

Closed MichaelChirico closed 4 years ago

MichaelChirico commented 4 years ago

From: Prof Brian D Ripley <ripley@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::>> On Mon, 19 Jun 2000 faraway@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::> wrote:

Full_Name: Julian Faraway
Version: R-Release (June 15)
OS: Linux (redhat)
Submission from: (NULL) (141.211.66.172)

R fails to compile on the current released version. Some constants 
are undefined in pnorm.c. It appears that adding

#include "nmath.h"

Let us stress that pnorm.c does contain that line in `the current released version', which is R-1.1.0, and not R-release.tar.{gz,bz2}. You might have noticed traffic about this on R-help earlier today. Sorry to be picky, but quite a lot of time has been misdirected by initial reports on this not being precise enough (and this is the third report).

R-release is the current version with all committed patches applied.
(On release day there are not likely to be any patches.) Unfortunately, something has gone wrong with checking it out of the archives to the tar file. The file is correct in the archives, so the problem is in R-release.tar.*, not in R-release per se. With that sort of level of detail being important, we need to be as precise as possible.

I wondered it there was any room for misunderstanding in what was said on CRAN, but it looks clear enough to me. R-release.tar.* is described there as a `snapshot', and that is the intention. Those of us who download snapshots from other projects know that sometimes they are badly broken.

-- Brian D. Ripley, ripley@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::> Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ <CENSORING FROM DETECTED PHONE NUMBER ONWARDS; SEE BUGZILLA>

METADATA

MichaelChirico commented 4 years ago

From: faraway@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::> Full_Name: Julian Faraway Version: R-Release (June 15) OS: Linux (redhat) Submission from: (NULL) (141.211.66.172)

R fails to compile on the current released version. Some constants are undefined in pnorm.c. It appears that adding

include "nmath.h"

to pnorm.c solves this problem.

METADATA

MichaelChirico commented 4 years ago

From: Friedrich Leisch <Friedrich.Leisch@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::>>

>>>> On Mon, 19 Jun 2000 21:19:13 +0100 (GMT Daylight Time),
>>>> Prof Brian D Ripley (PBDR) wrote:

PBDR> On Mon, 19 Jun 2000 faraway@<::CENSORED -- SEE ORIGINAL ON BUGZILLA::> wrote:

> Full_Name: Julian Faraway
> Version: R-Release (June 15)
> OS: Linux (redhat)
> Submission from: (NULL) (141.211.66.172)
> 
> 
> R fails to compile on the current released version. Some constants 
> are undefined in pnorm.c. It appears that adding
> 
> #include "nmath.h"

PBDR> Let us stress that pnorm.c does contain that line in `the current released PBDR> version', which is R-1.1.0, and not R-release.tar.{gz,bz2}. You might PBDR> have noticed traffic about this on R-help earlier today. Sorry to be PBDR> picky, but quite a lot of time has been misdirected by initial reports on PBDR> this not being precise enough (and this is the third report).

PBDR> R-release is the current version with all committed patches applied.
PBDR> (On release day there are not likely to be any patches.) PBDR> Unfortunately, something has gone wrong with checking it out of the PBDR> archives to the tar file. The file is correct in the archives, so the PBDR> problem is in R-release.tar.*, not in R-release per se. With that sort of PBDR> level of detail being important, we need to be as precise as possible.

PBDR> I wondered it there was any room for misunderstanding in what was said on PBDR> CRAN, but it looks clear enough to me. R-release.tar.* is described there PBDR> as a `snapshot', and that is the intention. Those of us who download PBDR> snapshots from other projects know that sometimes they are badly broken.

I've added a much stronger statement to CRAN saying that ``These archives are created automatically every night from the CVS tree, hence might not even compile on your platform and can contain any number of bugs. They will probably work, but maybe not.''

Hope this helps, Fritz

PS: Of course ``night'' is only true from a European perspective :-)

METADATA

MichaelChirico commented 4 years ago

NOTES: Corrupt R-release file set

METADATA

MichaelChirico commented 4 years ago

Audit (from Jitterbug): Wed Jun 21 12:03:21 2000 ripley changed notes Wed Jun 21 12:03:21 2000 ripley moved from incoming to Installation-fixed

METADATA