Hi there,

I want to bring up an issue with the perl/PGPLOT package on Debian/Ubuntu. The 
relevant apt package is ‘libpgplot-perl’. (This came up when I was installing 
PDL on Ubuntu and getting an unexpected graphics library instead of pgplot.)
 
In short, in the latest version of Ubuntu (>=17)  the dependency on the pgplot5 
package has been replaced with one on libcpgplot0. See 
https://packages.ubuntu.com/search?keywords=pgplot 
<https://packages.ubuntu.com/search?keywords=pgplot>
 
libcpgplot0 replaces the true pgplot library with a supposedly API compatible 
library provided by the giza package (see 
http://giza.sourceforge.net/documentation/pgplot.shtml 
<http://giza.sourceforge.net/documentation/pgplot.shtml>). I can imagine that 
it was thought that this was a good idea as it is newer, C based, etc.
 
While it is great to see this active development I do not think this has been 
handled in quite the best way. Giza is not a 100% pgplot API replacement, as 
yet. I am actually quite familiar with giza and have played with it. The giza 
library does NOT implement all pgplot functions and some of them work 
differently to the true pgplot library. (I know as I have tested this and have 
also corresponded with the developers who are also astronomers just down the 
road from at Monash Uni).This is clearly documented on the above web page and 
is an important issue for those of us who rely on pgplot for legacy code.
 
Also I do not think it should be done this way in principle. Obviously giza 
provides a drop in libpgplot replacement for those who wants to use it, so  how 
it should work is to make a ‘libpgplotgiza-perl’ instead to provide the option 
FOR THOSE WHO WISH TOO.
 
(I wish someone had asked me before making this change, after all I am the 
owner still of the CPAN perl/PGPLOT package whose functionality this is 
providing.)
 
I would like to get this revered. It is quite unclear to me who is making these 
decisions, I am not familiar with the Debian packaging universe. I was told by 
emailing here I could reopen this issue

thank you,
 
Karl Glazebrook

Attachment: signature.asc
Description: Digital Signature

Reply via email to