Re: Yum output and term size

pirla <[email protected]> writes:

> Well... I would like to have a list of available packages for every
> repos, so search should be done even if offline.
> Comparison can be done with a simple command (ie sdiff)
> Reinstallation can be done with a simple script and so on.

 From a cmd line you can do:

% repoquery -a --qf '%{nevra} %{repo}'

...and that's probably always going to do the "right" thing for
you. However, the snippets page also deals with this in a couple of ways.

> I think, and with me some other people on the fedora mailing list, that
> if output is sent to non terminal devices or to files or to pipe, it
> should not be fixed to a number of char (80 char in this case).

 If the output is going to a non-terminal then yum can't find out how
big it is. So the only way to make "yum blah" and "yum blah | cat"
look the same is to guess at what the terminal size is.
 You might be arguing that a default size of say "1000" is more often
correct than the current default of 80. I disagree, do you have any
data? You might argue that piping the output to anything means that
people want the output to look different ... and I would disagree very
strongly that people would expect that.

> We think that the optimal should be an option to have output for each
> package on one row, divided by a special char (ie tab).
>
> And, if possible, to have the first lines (control lines) not printed,
> or printed to stderr, so one can divide output from control.

 The snippets will show you how to do that.

-- 
James Antill -- [email protected]