ubuntu/debian package for latest ESS

[Martin Maechler]

>>>>> "DE" == Dirk Eddelbuettel <edd at debian.org>
>>>>>     on Mon, 19 May 2008 12:46:51 +0000 writes:

    DE> On Mon, May 19, 2008 at 02:42:22PM +0200, Martin Maechler wrote:
    >> >>>>> "DE" == Dirk Eddelbuettel <edd at debian.org>
    >> >>>>>     on Mon, 19 May 2008 11:47:42 +0000 writes:
    >> 
    DE> On Mon, May 19, 2008 at 09:30:06AM +0200, Martin Maechler wrote:
    >> >> >>>>> "DE" == Dirk Eddelbuettel <edd at debian.org>
    >> >> >>>>>     on Wed, 14 May 2008 15:48:37 +0000 writes:
    >> >> 
    DE> Well, I guess we need to triage them all.  
    >> >> 
    DE> On Wed, May 14, 2008 at 05:29:52PM +0200, Martin Maechler wrote:
    >> >> >> Hmm, I'm now looking at that and see --- the only bug marked as important
    >> >> >> >> Package: ess
    >> >> >> >> Version: 5.3.0-1
    >> >> >> >> Severity: important
    >> >> >> >> Justification: fails to build from source
    >> >> 
    DE> 'Fails to build from source' (or FTBFS for the iniated) typically
    DE> arises when the toolchain changes. It comes up all the time as some
    DE> people (like Lucas) have access to grids with >> 100 cpu and often
    DE> auto-rebuild the entire archive. 
    >> >> 
    DE> I think this was just a matter of changing (old) Depends on tetex to
    DE> (new) Depends on texlive. That one is clearly in my Debian court.
    >> >> 
    DE> Leaving bugs 
    >> >> 
    DE> * #440194: ess: Tries to load when package is removed
    DE> * #371058: ESS cannot detect help(foo, package=bar)
    DE> * #400382: ess: options(error=recover) in R confuses ESS
    DE> * #442159: ess: upgrade to fix complaints about .Last.value
    >> >> 
    DE> * #456001: Warning: make-variable-buffer-local' should be called at toplevel
    >> >> 
    >> >> I now have also committed a change to fix the above.
    >> 
    DE> I do not know what you refer to in "the above".  Could you use a bug number, please ?
    >> 
    >> #456001
    >> 
    >> 
    >> >> You might want to note that in  debian/changelog  
    >> >> before releasing the debian 5.3.8-svn.... package
    >> 
    DE> I tried. Could you please refer to debian/changelog, possibly with
    DE> patches.  As a certain MM wrote in another email, "that is we all have
    DE> shared SVN access..."
    >> 
    >> but I don't speak Debian-English .. even my regular English
    >> seems confusing enough ...
    >> 
    >> Since my commit added to lisp/ & lisp/ChangeLog I did not want to
    >> add to yet another changelog, particularly, since the change
    >> was *not* in the debian/ subdirectory but rather the lisp/ one.
    >> 
    >> The main reason for my e-mail was that you should note the svn
    >> commit ..

    DE> I see I was confused because 456001 had already been identified by Stefan as fixed.  

aha

    DE> So are we all-aboard then that

    DE> #400382: ess: options(error=recover) in R confuses ESS	

    DE> is the last open bug (and it is understood that ESS Core now knows it
    DE> and may get to it, or not, in a future release) 

No!  In my eyes, the last open "bug" is

       * #371058: ESS cannot detect help(foo, package=bar)

which I'd label a feature request rather than a bug,
and that's the only open one I think.

The    options(error=recover)
was dealt with here {file lisp/ChangeLog } :

   2007-01-22  Martin Maechler  <maechler at stat.math.ethz.ch>

	   * ess-cust.el (inferior-ess-safe-names-command): new
	   try(names(.)); this finally gets rid of "hanging" in object
	   completion, e.g., when options(error=recover) is active in R.

and I would not know of any other problem with it.

    DE> so that, as far as the Debian bug tracker is concerned,
    DE> released 5.3.8~svn3... would bring us down to one?

(or zero)

    DE> Dirk, now looking to upload the shiny new package this eve thanks to everybody's input

looking forward myself!
Martin

    DE> -- 
    DE> Three out of two people have difficulties with fractions.