Re: Bio-Linux
Hi Andreas,
On 08/09/2018 18:14, Andreas Tille wrote:
Hi William,
On Sat, Sep 08, 2018 at 05:05:05PM +0100, William McCaffery wrote:
As I understood it, the issue was that the virtual packages would install
one of a group of packages that all have similar functionality, but none of
these packages were available.
No. What you describe are metapackages. A virtual package is for instance
www-browser. This is provided by for instance firefox or chromium - but the
package www-browser itself does not exist (thus the name "virtual").
I think my phrasing was wrong -- no one seemed to be providing for the
virtual package if that's possible. It could just as easily be a
misleading error message or I may have attributed an error to the wrong
aspect of what I was doing.
I may have misunderstood what was going
wrong. I have not tested how they will affect the task, they may not cause
any issues.
Virtual packages have nothing to do with our metapackages since
metapackages are really existing and are defining a set of dependencies.
Looking at the server task, I also had a quick question, although this may
be obvious I just need a bit of a sanity check -- is the server task going
to contain all of the desktop packages with the additional server packages,
or is it only going to contain server packages?
Hmmm, may be I was misleading you with the server package. I was lazy enough
to copy a "random" task of Debian Med. You possibly need to remove all those
recommends from that task and replace it by something that makes sense for
Bio-Linux. I personally can not tell what makes sense since I need to admit
that I do not know Bio-Linux sufficiently enough. Just feel free to change
*everything* and do not assume any sensible content in my lazy example.
I understand that I need to replace it; this may have been more of a
subjective question -- would we want the server package to contain
exclusively server packages, or would it contain normal bio-linux
packages as well as server packages?
Since I've based my tasks off of Tim's list, it doesn't have any why, remark
or section entries, although I would like to carry some of these over in a
future commit.
There is no need for Remark entries. That was just used for Debian Med
internal use. Feel free to remove all remarks. I have no idea what you
mean by section entries.
It still seems like a good practice to keep, although I know that it has
no effect on the functionality of the package. I also meant category,
rather than section. My mistake.
I would still like to do a check over before pushing to the repo.
Please note that you can not break anything. The repo is not used by
anything. The tasks pages are not rendered at all. So your chances to
break anything are absolutely zero. However, I could do a look at your
work and fix bugs if I see some. So I keep on recommending to push as
soon and frequent as possible.
I admit I've developed something of a bad habit in pushing quite
infrequently, mostly since in earlier years of university we were
somewhat inept at using version control and as a result tried to push
very carefully to avoid messing up the repo (in fact we often resorted
to sending files over Facebook by the end of a practical). Things are a
bit hectic at the moment since I'm about to head back to university but
I'll make sure to get a push in soon.
To
clarify, when I said that wanted to produce a usable result I more meant
that I wanted to avoid pushing an obvious mistake.
Any mistake would not have a bad effect. It would have the good effect
that I volunteer to fix it and help you learning the stuff by doing so.
Thanks,
You are welcome
Andreas.
PS: Thanks for switching to the list.
Sorry I completely forgot to add the list in the past few emails, thanks
again for all your help,
William
Reply to: