[Rd] Bug#361968: Wrong name in manpage

[friedrich.leisch at stat.uni-muenchen.de]

I've committed a fix for 2.3 affecting only "R --version" (don't want to
play with real R output during feature freeze), and a cleaner solution
for 2.4 which gets the startup message and --version even more in
sync.

Passes all checks for me, and the man page looks OK again. Thanks for
spotting the bug.

Best,
Fritz


>>>>> On Wed, 12 Apr 2006 13:53:08 +0200,
>>>>> friedrich leisch (fl) wrote:

>>>>> On Tue, 11 Apr 2006 20:56:27 -0400,
>>>>> Duncan Murdoch (DM) wrote:

  >> On 4/11/2006 7:45 PM, Dirk Eddelbuettel wrote:
  >>> R-devel'ers, 
  >>> 
  >>> On 11 April 2006 at 10:24, Dirk Eddelbuettel wrote:
  >>> | 
  >>> | Kanru,
  >>> | 
  >>> | Thanks for the bugreport.
  >>> | 
  >>> | On 11 April 2006 at 22:03, Kanru Chen wrote:
  >>> | | Package: r-base-core
  >>> | | Version: 2.2.1.svn37668-1
  >>> | | Severity: minor
  >>> | | 
  >>> | | In manpage of /usr/bin/R, the first, fourth and last line shows `VERSION'
  >>> | | instead of `R'.
  >>> 
  >>> I haven't seen any follow-up yet -- here is what it looks like (cut and
  >>> pasted from Emacs man page viewer) and note the 'Version' in place of R:
  >>> 
  >>> VERSION(1)                            FSF                           VERSION(1)
  >>> 
  >>> NAME
  >>> Version - a language for data analysis and graphics
  >>> 
  >>> SYNOPSIS
  >>> R [options] [< infile] [> outfile]
  >>> R CMD command [arguments]
  >>> 
  >>> DESCRIPTION
  >>> Start  R,  a  system for statistical computation and graphics, with the
  >>> specified options, or invoke an R tool via the 'R CMD' interface.
  >>> [...]
  >>> 
  >>> 
  >>> | | I believe it is a typo.
  >>> | 
  >>> | More likely something is wrong with how R.1 is autogenerated using help2man.
  >>> | 
  >>> | Incidentally, that `R --version' now starts its ouput with 'Version' rather
  >>> | than R had bit us in the RPy builds where the version number was regexp'ed
  >>> | out of the result, and was still expecting the line to start with R just like
  >>> | help2man seems to expect the program name first. 
  >>> 
  >>> It seems to stem from src/main/version.c:
  >>> 
  >>> void attribute_hidden PrintVersionString(char *s)
  >>> {
  >>> if(strcmp(R_SVN_REVISION, "unknown")==0)
  >>> {
  >>> sprintf(s, "Version %s.%s %s (%s-%s-%s)",
  >>> R_MAJOR, R_MINOR, R_STATUS, R_YEAR, R_MONTH, R_DAY);
  >>> }
  >>> else{
  >>> if(strlen(R_STATUS)==0){
  >>> sprintf(s, "Version %s.%s (%s-%s-%s)",
  >>> R_MAJOR, R_MINOR, R_YEAR, R_MONTH, R_DAY);
  >>> }
  >>> else{
  >>> sprintf(s, "Version %s.%s %s (%s-%s-%s r%s)",
  >>> R_MAJOR, R_MINOR, R_STATUS, R_YEAR, R_MONTH, R_DAY,
  >>> R_SVN_REVISION);
  >>> }
  >>> }
  >>> }
  >>> 
  >>> Would replacing 'Version ...' with 'R (Version ...)' be an acceptable
  >>> alternative ?

  >> I think the problem is that PrintVersion (a few lines up from there) 
  >> used to put the R in front; PrintVersionString doesn't include the R. 
  >> PrintVersionString is called from other places where the R would not be 
  >> appropriate, but PrintVersion is only called when acting on `R 
  >> --version' or synonyms.

  >> Fritz, I think this was your change in r36923 a few months ago.  Do you 
  >> have time to deal with it?

  > Yes, will do later today.

  > Best,
  > Fritz

  > ______________________________________________
  > R-devel at r-project.org mailing list
  > https://stat.ethz.ch/mailman/listinfo/r-devel