Skip to content

sources of 2.4.0 patched

10 messages · apjaworski@mmm.com, Peter Dalgaard, Martin Maechler +2 more

#
Hi,

I have a question about the availability of tarballs for 2.4.0-patched.

I routinely compile and install fresh versions of R-patched as well as
R-devel every few days.  I do it mostly "for fun" but also to check for
possible build problems or problems with my development environment
(Win2000, cygwin plus all the necessary tools including MinGW compilers and
MiKTex).

I get the tarballs form ftp://ftp.stat.math.ethz.ch/Software/R and I
noticed that the only R-patched there is dated 2006-10-03, while R-devel is
up to 2006-10-25.  Did, by any chance, a depository for the R-patched
tarballs move someplace else, or is the 2006-10-03 version the latest one.

Thanks in advance,

Andy

__________________________________
Andy Jaworski
518-1-01
Process Laboratory
3M Corporate Research Laboratory
-----
E-mail: apjaworski at mmm.com
Tel:  (651) 733-6092
Fax:  (651) 736-3122
#
apjaworski at mmm.com writes:
To my knowledge, neither. Something has gone wrong. The build date on
the file is Oct 27... I suspect that the svn update is somehow failing
on the machine that does the builds (i.e. it just keeps on building
from the 2006-10-03 checkout).
#
On 28 October 2006 at 00:18, Peter Dalgaard wrote:
| apjaworski at mmm.com writes:
| > I get the tarballs form ftp://ftp.stat.math.ethz.ch/Software/R and I
| > noticed that the only R-patched there is dated 2006-10-03, while R-devel is
| > up to 2006-10-25.  Did, by any chance, a depository for the R-patched
| > tarballs move someplace else, or is the 2006-10-03 version the latest one.
| 
| To my knowledge, neither. Something has gone wrong. The build date on
| the file is Oct 27... I suspect that the svn update is somehow failing
| on the machine that does the builds (i.e. it just keeps on building
| from the 2006-10-03 checkout).

That seems to be the case again this morning.

Dirk
#
Dirk> On 28 October 2006 at 00:18, Peter Dalgaard wrote: |
    Dirk> apjaworski at mmm.com writes: | > I get the tarballs form
    Dirk> ftp://ftp.stat.math.ethz.ch/Software/R and I | >
    Dirk> noticed that the only R-patched there is dated
    Dirk> 2006-10-03, while R-devel is | > up to 2006-10-25.
    Dirk> Did, by any chance, a depository for the R-patched | >
    Dirk> tarballs move someplace else, or is the 2006-10-03
    Dirk> version the latest one.  | | To my knowledge,
    Dirk> neither. Something has gone wrong. The build date on |
    Dirk> the file is Oct 27... I suspect that the svn update is
    Dirk> somehow failing | on the machine that does the builds
    Dirk> (i.e. it just keeps on building | from the 2006-10-03
    Dirk> checkout).

    Dirk> That seems to be the case again this morning.

but no longer now.
[A  "svn cleanup"  was needed.
 I've amended the cron-job script to do an "svn cleanup" now 
 always before "svn update".  This of course is a waste of time in
 99.9% of cases; but it's just a CPU's time, not a person's ....]

Usually such problems would be solved more quickly if someone
e-mailed me directly...

Martin Maechler, ETH Zurich
#
On 28 October 2006 at 22:15, Martin Maechler wrote:
| >>>>> "Dirk" == Dirk Eddelbuettel <edd at debian.org>
| >>>>>     on Sat, 28 Oct 2006 08:30:38 -0500 writes:
| 
|     Dirk> On 28 October 2006 at 00:18, Peter Dalgaard wrote: |
|     Dirk> apjaworski at mmm.com writes: | > I get the tarballs form
|     Dirk> ftp://ftp.stat.math.ethz.ch/Software/R and I | >
|     Dirk> noticed that the only R-patched there is dated
|     Dirk> 2006-10-03, while R-devel is | > up to 2006-10-25.
|     Dirk> Did, by any chance, a depository for the R-patched | >
|     Dirk> tarballs move someplace else, or is the 2006-10-03
|     Dirk> version the latest one.  | | To my knowledge,
|     Dirk> neither. Something has gone wrong. The build date on |
|     Dirk> the file is Oct 27... I suspect that the svn update is
|     Dirk> somehow failing | on the machine that does the builds
|     Dirk> (i.e. it just keeps on building | from the 2006-10-03
|     Dirk> checkout).
| 
|     Dirk> That seems to be the case again this morning.
| 
| but no longer now.

Thanks!

| [A  "svn cleanup"  was needed.
|  I've amended the cron-job script to do an "svn cleanup" now 
|  always before "svn update".  This of course is a waste of time in
|  99.9% of cases; but it's just a CPU's time, not a person's ....]
| 
| Usually such problems would be solved more quickly if someone
| e-mailed me directly...

Put a README into the directory ?

[ BTW SVN 1.4.0, now in Debian testing testing apparently shrinks archive
size by a third but you need to 'dump' and 'restore'. ]

Dirk
#
On 28 October 2006 at 22:15, Martin Maechler wrote:
| but no longer now.
| [A  "svn cleanup"  was needed.
|  I've amended the cron-job script to do an "svn cleanup" now 
|  always before "svn update".  This of course is a waste of time in
|  99.9% of cases; but it's just a CPU's time, not a person's ....]
| 
| Usually such problems would be solved more quickly if someone
| e-mailed me directly...

Something still looks still fishy. Cut&pasted from the directory view:

Directory: R-patched.tar.bz2  	 	10/29/2006  	01:00:00 AM
Directory: R-patched.tar.gz 		10/29/2006 	01:00:00 AM
File: R-patched_2006-10-03.tar.bz2 	11798 KB 	10/28/2006 	12:58:00 AM
File: R-patched_2006-10-03.tar.gz 	14214 KB 	10/28/2006 	12:57:00 AM
File: R-patched_2006-10-27.tar.bz2 	11810 KB 	10/29/2006 	01:00:00 AM
File: R-patched_2006-10-27.tar.gz 	14221 KB 	10/29/2006 	12:59:00 AM
File: R-release.diff.gz 	68 KB 	10/29/2006 	01:17:00 AM

Or is the date in R-patched derived from the last SVN checkin which would
then happen to have been 2006-10-27?  Not according to the top of the svn log
from http://developer.r-project.org/R.svnlog.2006: (and cut down to dates
only)

r39739 | murdoch | 2006-10-28 21:13:45 -0400 (Sat, 28 Oct 2006) | 1 line
r39738 | murdoch | 2006-10-28 21:07:16 -0400 (Sat, 28 Oct 2006) | 1 line
r39737 | murdoch | 2006-10-28 19:56:27 -0400 (Sat, 28 Oct 2006) | 1 line
r39736 | murdoch | 2006-10-28 13:04:58 -0400 (Sat, 28 Oct 2006) | 1 line
r39735 | maechler | 2006-10-27 17:53:36 -0400 (Fri, 27 Oct 2006) | 1 line
r39734 | hornik | 2006-10-27 13:59:17 -0400 (Fri, 27 Oct 2006) | 1 line
r39733 | hornik | 2006-10-27 13:59:16 -0400 (Fri, 27 Oct 2006) | 1 line
[...]


Hth, Dirk
#
On 29 October 2006 at 10:29, Dirk Eddelbuettel wrote:
|
| On 28 October 2006 at 22:15, Martin Maechler wrote:
| | but no longer now.
| | [A  "svn cleanup"  was needed.
| |  I've amended the cron-job script to do an "svn cleanup" now 
| |  always before "svn update".  This of course is a waste of time in
| |  99.9% of cases; but it's just a CPU's time, not a person's ....]
| | 
| | Usually such problems would be solved more quickly if someone
| | e-mailed me directly...
| 
| Something still looks still fishy. Cut&pasted from the directory view:
| 
| Directory: R-patched.tar.bz2  	 	10/29/2006  	01:00:00 AM
| Directory: R-patched.tar.gz 		10/29/2006 	01:00:00 AM
| File: R-patched_2006-10-03.tar.bz2 	11798 KB 	10/28/2006 	12:58:00 AM
| File: R-patched_2006-10-03.tar.gz 	14214 KB 	10/28/2006 	12:57:00 AM
| File: R-patched_2006-10-27.tar.bz2 	11810 KB 	10/29/2006 	01:00:00 AM
| File: R-patched_2006-10-27.tar.gz 	14221 KB 	10/29/2006 	12:59:00 AM
| File: R-release.diff.gz 	68 KB 	10/29/2006 	01:17:00 AM
| 
| Or is the date in R-patched derived from the last SVN checkin which would
| then happen to have been 2006-10-27?  Not according to the top of the svn log
| from http://developer.r-project.org/R.svnlog.2006: (and cut down to dates
| only)
| 
| r39739 | murdoch | 2006-10-28 21:13:45 -0400 (Sat, 28 Oct 2006) | 1 line
| r39738 | murdoch | 2006-10-28 21:07:16 -0400 (Sat, 28 Oct 2006) | 1 line
| r39737 | murdoch | 2006-10-28 19:56:27 -0400 (Sat, 28 Oct 2006) | 1 line
| r39736 | murdoch | 2006-10-28 13:04:58 -0400 (Sat, 28 Oct 2006) | 1 line
| r39735 | maechler | 2006-10-27 17:53:36 -0400 (Fri, 27 Oct 2006) | 1 line
| r39734 | hornik | 2006-10-27 13:59:17 -0400 (Fri, 27 Oct 2006) | 1 line
| r39733 | hornik | 2006-10-27 13:59:16 -0400 (Fri, 27 Oct 2006) | 1 line
| [...]

... unless of course all those checkins went to trunk/ aka R-devel.

Dirk
#
On 10/29/2006 12:43 PM, Dirk Eddelbuettel wrote:
Tbey're not all to trunk, but 39735 is the last one to R-patched, so the 
date in the dir listing looks okay.

Duncan Murdoch
#
Dirk Eddelbuettel <edd at debian.org> writes:
... which "svn -r 39733:BASE" on the branch would have told you:

$ svn log -r 39733:BASE
------------------------------------------------------------------------
r39734 | hornik | 2006-10-27 19:59:17 +0200 (Fri, 27 Oct 2006) | 1
line

Updated.
------------------------------------------------------------------------
r39735 | maechler | 2006-10-27 23:53:36 +0200 (Fri, 27 Oct 2006) | 1
line

belongs to r39726 (dotchart par tweak)
------------------------------------------------------------------------
r39741 | murdoch | 2006-10-29 17:23:55 +0100 (Sun, 29 Oct 2006) | 1
line

Port r39740 (Rconsole fixes) from trunk
------------------------------------------------------------------------

i.e. there is one from the 29th after the snapshot at ~1 AM and none
from the 28th.
#
On 29 October 2006 at 19:29, Peter Dalgaard wrote:
| Dirk Eddelbuettel <edd at debian.org> writes:
| 
| 
| > ... unless of course all those checkins went to trunk/ aka R-devel.
| 
| ... which "svn -r 39733:BASE" on the branch would have told you:

At that point it "couldn't have" as I was quoting from the web-log as I wrote
(quoting from my email now)

| then happen to have been 2006-10-27?  Not according to the top of the svn log
| from http://developer.r-project.org/R.svnlog.2006: (and cut down to dates

I am now proud owner of (partial) svn checkrout (of 2.4-patched only) so I
take the hint with re the svn usage with nodding approval.

Thanks, Dirk