Martin,
Thanks for taking a quick look. Yes it looks like it must be something
local.
I've done the following to make sure I have a clean box:
reboot
sudo apt-get update
sudo apt-get upgrade
Rerun the "svn up" command and do a
make distclean
svn up
tools/rsync-recommended
./configure
make
The problem persists in 2017-09-01 r73179
I pulled the source code for 3.4-1 and compiled it, the problem persists.
I then did sudo apt-get install R, which installs version 3.2.3 on the machine.
The plots
work as expected there.
The new plots are the same size, same layout and margins, but the text and lines
are
definitely thinner. According to the package manager I have the latest version
of
texlive-fonts-extra.
I don't see how to get it to pull a newer compiled version than 3.2.
Anything else that would help?
Terry T.
On 09/01/2017 02:37 AM, Martin Maechler wrote:> > My on-screen plots with the latest build are acting
> > strange. If the y label is longer than some small value it is
not shown.
>
> [MM: moved the important part of the script up here:]
>
> plot(1:5, 1:5, ylab="abcde") # has a y label
>
> plot(1:5, 1:5, ylab="abcdefghi") # no label
>
> That is amazing.
> I don't see a problem here (Linux Fedora F24, or F26, in both
> case using GNOME as windowing system),
> also with this, both interactive or pdf :
>
> plot(1:5, 1:5, ylab="abcdefghi or even quite an order of magnitude
longer")
>
> pdf("long-ylab.pdf")
> plot(1:5, 1:5, ylab="abcdefghi or even quite an order of magnitude
longer")
> dev.off(); system("evince long-ylab.pdf &")
>
> I have no further idea but guess this must be specific to your platform.
>
> Did you try the same thing with R 3.4.1?
> Did you install both in the same way -- from the sources ??
On 1 September 2017 at 14:28, Therneau, Terry M., Ph.D. wrote:
| Martin,
| Thanks for taking a quick look. Yes it looks like it must be something
local.
| I've done the following to make sure I have a clean box:
| reboot
| sudo apt-get update
| sudo apt-get upgrade
|
| Rerun the "svn up" command and do a
| make distclean
| svn up
| tools/rsync-recommended
| ./configure
| make
|
| The problem persists in 2017-09-01 r73179
| I pulled the source code for 3.4-1 and compiled it, the problem persists.
|
| I then did sudo apt-get install R, which installs version 3.2.3 on the
machine. The plots
| work as expected there.
Read the fine README here:
https://cloud.r-project.org/bin/linux/ubuntu/README.html
Similar one for Debian at https://cloud.r-project.org/bin/linux/debian/
Many of us have been using R from CRAN on Ubuntu and Debian for many years,
and many update cycles. I try hard to get new R versions into Debian the
morning of the release [1] and Michael and Johannes usually produce binary
for the mirrors within a day or two too. On this box still running 16.10:
edd at bud:~$ apt-cache policy r-base-core
r-base-core:
Installed: 3.4.1-2yakkety0
Candidate: 3.4.1-2yakkety0
Version table:
*** 3.4.1-2yakkety0 500
500 https://cloud.r-project.org/bin/linux/ubuntu yakkety/ Packages
100 /var/lib/dpkg/status
3.4.1-1yakkety0 500
500 https://cloud.r-project.org/bin/linux/ubuntu yakkety/ Packages
3.4.0-1yakkety0 500
500 https://cloud.r-project.org/bin/linux/ubuntu yakkety/ Packages
3.3.3-1yakkety0 500
500 https://cloud.r-project.org/bin/linux/ubuntu yakkety/ Packages
3.3.2-1yakkety0 500
500 https://cloud.r-project.org/bin/linux/ubuntu yakkety/ Packages
3.3.1-1build1 500
500 http://us.archive.ubuntu.com/ubuntu yakkety/universe amd64 Packages
edd at bud:~$
Dirk
[1] And once every couple of years I get to do it on a high-speed train...
--
http://dirk.eddelbuettel.com | @eddelbuettel | edd at debian.org
I did read that fine file. I added the following line to my /etc/apt/sources.list file: tmt-local2023% tail -2 /etc/apt/sources.list # R sources deb https://mirror.las.iastate.edu/bin/linux/ubuntu xenial/ Here is what apt says on my machine tmt-local2024% apt-cache policy r-base-core r-base-core: Installed: 3.2.3-4 Candidate: 3.2.3-4 Version table: *** 3.2.3-4 500 500 http://us.archive.ubuntu.com/ubuntu xenial/universe amd64 Packages 100 /var/lib/dpkg/status My long term habit is to stick to LTS versions, which currently is 16.04: "still running 16.10" is a different outlook. The system admins here tend to be more conservative than me. Again, thanks for any input. Getting this particular box straightened out quickly isn't critical, but I would like to understand the root of the issue. Terry T. On 09/01/2017 02:58 PM, Dirk Eddelbuettel wrote:> Read the fine README here:https://cloud.r-project.org/bin/linux/ubuntu/README.html > Similar one for Debian athttps://cloud.r-project.org/bin/linux/debian/ > > Many of us have been using R from CRAN on Ubuntu and Debian for many years, > and many update cycles. I try hard to get new R versions into Debian the > morning of the release [1] and Michael and Johannes usually produce binary > for the mirrors within a day or two too. On this box still running 16.10: > > edd at bud:~$ apt-cache policy r-base-core > r-base-core: > Installed: 3.4.1-2yakkety0 > Candidate: 3.4.1-2yakkety0 > Version table: > *** 3.4.1-2yakkety0 500 >[[alternative HTML version deleted]]