[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: gstreamer1.0-libav - necessary for browsers to play videos?



On 06/18/2017 07:54 AM, Brian wrote:
On Sun 18 Jun 2017 at 00:27:29 -0400, Jape Person wrote:

Apropos of nothing but wishing to supply an explanation to anyone
else who might run into the same issue.

It is my habit to perform apt update followed by apt full-upgrade
every day on my testing systems. I get the impression that this may
not be a common practice, but I've been doing this (apt
full-upgrade or, earlier on, apt-get dist-upgrade) on a daily basis
for years with only rare resulting problems, all of which have been
fixed easily.

I also routinely run apt --purge autoremove and debfoster to clear
out packages that are no longer needed.

All sensible procedures.


Good to know. Some of the messages I've read on this list have made me wonder if dist-upgrade / full-upgrade is almost dreaded by some users.

The recent firefox-esr upgrade resulted in the following output in
/var/log/apt/history.log:

Start-Date: 2017-06-16  10:15:49 Commandline: apt full-upgrade
Install: libjsoncpp1:amd64 (1.7.4-3, automatic) Upgrade:
firefox-esr:amd64 (45.9.0esr-1, 52.2.0esr-1~deb9u1) End-Date:
2017-06-16  10:15:54

I ran debfoster, and it asked me if I wanted to keep
gstreamer1.0-libav. I ran aptitude why gstreamer1.0-libav and got
this result:

# aptitude why gstreamer1.0-libav i   task-xfce-desktop Recommends
libreoffice i A libreoffice       Suggests   gstreamer1.0-libav

Hmmm. Looks like there's no reason to keep gstreamer1.0-libav, so I
let debfoster remove it.

debfoster (which I do not use) queries whether you should keep a
package which firefox-esr recommends? deborphan doesn't do this.


On my systems:

$ deborphan -Ps --ignore-suggests
main/libs                 gstreamer1.0-libav       optional

Having had to use at least some Windows systems up through the release of Vista, I abhor having anything hang around that isn't truly necessary. I may be guilty of overdoing it sometimes.

Following this, no browser on the three testing systems I have
(firefox, epiphany, or qupzilla) would play any kind of video at
youtube.com or at any other location.

My main Jessie machine does not install recommended packages; it
plays youtube clips within firefox-esr.


So you don't even install recommends normally? I would have supposed (from reading various descriptions of recommends) that this would result in significant functional compromise in most packages. Not usually so?

I think it's odd that I always install Recommends but not Suggests, and that my browsers won't play video without this particular Suggested package.

Following re-installation of gstreamer1.0-libav all browsers were
once again able to play videos.

I would have thought that aptitude why might have given me a hint
about the browsers requiring this package. I've looked to be sure
the browsers do, indeed, have all of their depends and recommends
installed, and they do. (I do not install suggests as a rule, and I
don't use any kind of proprietary codecs or player software. So I
am dependent upon the DFSG-compliant software available in the
Debian repositories to play any video or audio I'm going to use on
these systems.)

This is, obviously, not a very serious problem, but it's an
interesting one that might bite others as unwary as I. Maybe it's
implicated somehow in some of the odd reports we see from
time-to-time of someone who can't get a browser to play videos.

Worthy of a bug report?


As an additional note, I actually run both deborphan (with the -Ps and --ignore-suggests options) and debfoster after upgrades / installations / removals. A little OCD, perhaps, but I've seen deborphan and debfoster behave slightly differently on these systems -- depending, I'm guessing, on how packages and their Depends / Recommends have been installed in the first place.

Thanks for your notes, Brian.

JP


Reply to: