2

I've installed program qcad to /opt/qcad-dir directory. Binaries are in /opt/qcad-dir at all in one place. I had to change native qcad.sh script to use program as alternative (update-alternatives). I added DIR=/opt/qcad-dir to run the program with its libs, otherwise, qcad that run from another directory isn't saw its' libs.

So. I moved all program files into /opt/qcad-dir, then I change qcad.sh to make it possible to run from another directory. I created symlink to alternative. And now it is available to run qcad from anywhere.

Is it more correct way to install program (which stores all binaries in the same dir) system widely?

Juriy
  • 63
  • 1
  • 7

1 Answers1

1

The Linux Standard Base and the Filesystem Hierarchy Standard are arguably the standards of where and how you should install software on a Linux system and would suggest placing software that isn't included in your distribution either in /opt or /usr/local/ or rather subdirectories therein (/opt/<package> /opt/<provider> /usr/local/bin).

Best practice is to convert software you download to a installable software package native to your distribution, for instance dpkg or rpm formats. That will facilitate reporting, upgrading and cleanly removing the software.

XsiSecOfficial
  • 954
  • 8
  • 20
  • If there is no building scripts, should I do the package for program by myself? – Juriy Sep 05 '16 at 11:37
  • In addition, to precise my question. In whole story I didn't like the way I changed run script. This is dirty way, I think, to modify maintainer-supplied stuff. – Juriy Sep 05 '16 at 11:39