Solved Chromium disappeared

I'm on latest:

sh:
jb@bsd ~> doas pkg update -f
Updating FreeBSD repository catalogue...
Fetching meta.conf: 100%    163 B   0.2kB/s    00:01   
Fetching packagesite.pkg: 100%    7 MiB   7.4MB/s    00:01   
Processing entries: 100%
FreeBSD repository update completed. 34439 packages processed.
All repositories are up to date.
jb@bsd ~> doas pkg search chromium
chromium-bsu-0.9.16.1_2        Arcade-style, top-scrolling space shooter
jb@bsd ~> chrome --version
ld-elf.so.1: Shared object "libicui18n.so.73" not found, required by "chrome"
jb@bsd ~ [1]>
 
Well, I note this too. But that's probably just a problem of port compilation. It will reappear later.

The error you get with chrome --version is something not related as I have no error with this command.

So the question is: what did you do concerning pkg upgrade? Have you changed your repos from quarterly to latest recently? What version of FreeBSD ?
 
Yes, that's correct. I have switched between latest and quarterly. But from the user perspective there shouldn't be any "reappear later", I guess. I'm trying to show the greatness of the new FreeBSD 14 to friends and silly issues like this one simply prevent more people come to BSD. Really, it's that simple. Single try => yes/no.
 
You can try building it yourself. Really, it's that simple.
I don't think it's necessary to be ironic. I just pointed out that people do try FreeBSD (I'm not speaking about myself). But if after installing it some random package as important as Chromium is missing (and it was there few weeks before and will be again & they did follow the Handbook) they simply conclude that it's still not worth trying on desktop. Just a silly feedback.

Anyway, putting irony aside, why there's no Chromium as pkg until a newer package is built?
 
You can try building it yourself. Really, it's that simple.

And yet it can take so much time...

However,

jb82, what i do in such a case is using the same repository but from an other FreeBSD release so that i do not run in an library mismatch as you.

In case like chromium its not an issue to pick it from an other release, but wouldn't do it with kernel modules etc..

At this time of writing:
 
Guys, this really should be a five minute thread :) I really do know how to overcome this stupid issue. I'm just humbly asking what might've happened server-side when a standard package is not readily available as now. And my two cents here are that silly issues like this one annoy newcomers. And I really do think it's relevant to point that out.
 
Chromium and friends like iridium are ports which doesn't build easily on the FreeBSD build clusters and do fail to build or timeout very often.

I'm just humbly asking what might've happened

1704917831130.png


=>> Killing timed out build after 172800 seconds
 
Why there's no previous version readily available until a new package is successfully built.

The version of the port got bumped so a new build is triggered. It makes sense that it gets removed if it fails, imagine the dependency hell of package groups that need to match like GNOME or Mesa.
 
Guys, this really should be a five minute thread :) I really do know how to overcome this stupid issue. I'm just humbly asking what might've happened server-side when a standard package is not readily available as now. And my two cents here are that silly issues like this one annoy newcomers. And I really do think it's relevant to point that out.
Newcomers use quarterly ports - less flux in there than latest.

Temporary solution is pkg install firefox :)

Earlier post is sarcasm, by the way. Not irony.
 
I'd say they do not easily compile anywhere. But my question was, why there's no previous version readily available until a new package is successfully built.
Building chromium locally forces me for 6 to 9 hours every time. So even on build cluster, it would take much shorter, but still long time even on successful build.
If I understand correctly, the resulting repo by build cluster is master repo itself and build cluster deletes previous build of each port on starting up the new build, so the previous one dissapears.
Then why not changing the behaviour to keep previous one, or double the repo by creating independent master from builders' one? I'm not a insider of FreeBSD project, but maybe because in many cases new version includes fixes for vulnerability, meaning older version is VULNERABLE.
This time, commit message includes "Security: https://vuxml.freebsd.org/freebsd/ec8e4040-afcd-11ee-86bb-a8a1599412c6.html".
The link doesn't seem to work yet for now, but it would come later, hopefully.
 
Building chromium locally forces me for 6 to 9 hours every time. So even on build cluster, it would take much shorter, but still long time even on successful build.
If I understand correctly, the resulting repo by build cluster is master repo itself and build cluster deletes previous build of each port on starting up the new build, so the previous one dissapears.
Then why not changing the behaviour to keep previous one, or double the repo by creating independent master from builders' one? I'm not a insider of FreeBSD project, but maybe because in many cases new version includes fixes for vulnerability, meaning older version is VULNERABLE.
This time, commit message includes "Security: https://vuxml.freebsd.org/freebsd/ec8e4040-afcd-11ee-86bb-a8a1599412c6.html".
The link doesn't seem to work yet for now, but it would come later, hopefully.
Finally some motivation behind this behavior. Thanks!
 
My understanding is that there's a small group of mostly unpaid volunteers doing the work so if you have the time, skills or dollars to help out I'm sure they would be welcomed.

There's no-one behind the scenes going "oh, today's a nice day to remove an important package" - just a bunch of people doing their best with the resources available. And those resources often don't include any wages or payments to those people doing the work.

In an ideal world there would be backup repos, repos of historic releases, endless platform support, perfect Wifi and GPU support etc. But that world takes a lot more time and money and people.
 
My understanding is that there's a small group of mostly unpaid volunteers doing the work so if you have the time, skills or dollars to help out I'm sure they would be welcomed.
Have you ever used Arch Linux? A project created by mostly unpaid volunteers, yet people really often suggest improvements and complain about wrong/missing/faulty packages because they want to improve things/processes.
 
Have you ever used Arch Linux? A project created by mostly unpaid volunteers, yet people really often suggest improvements and complain about wrong/missing/faulty packages because they want to improve things/processes.
Fair comment.

But do they also put unhelpful wording in their suggestions e,g. “silly” and “it’s that simple” or are they more constructive?

Edit: but thinking about it things are a bit easier for Arch Linux - the kernel is written by others. Wifi and GPU drivers won’t be their concern and may even be supplied by the manufacturers. The bulk of the software is developed and tested on Linux - there’s not such a need for porting effort - Linuxisms aren’t a problem on Linux. Google support Linux but not FreeBSD. They (Arch Linux) are packaging a kernel and software into their distribution (a lot more than I could manage so definitely not belittling their work).

There‘s a lot more for the FreeBSD developers to do and porting is a considerably bigger task. So more to go wrong. Something like Alpine Linux where they’ve got the complication of the musl library might be a closer example of the porting effort.
 
Have you ever used Arch Linux? A project created by mostly unpaid volunteers, yet people really often suggest improvements and complain about wrong/missing/faulty packages because they want to improve things/processes.
Arch Linux is a distribution built upon GNU/Linux which has had a lot more volunteers and a lot more corporate involvement too, particularly in the kernel.

If you used the current/develop branch of Arch Linux, you'd have the same issue. Or pacman would just obliterate your entire install because of dependency hell, and uninstall itself without possibility of recovery. I have been there a lot.

Most Linux distros use the snap store (containerisation) to deliver Chromium because it's such a giant pain in the arse to maintain and keep up to date.

If I'm using Linux, I use a crippled variant of Xubuntu. When it goes wrong, it's easy to get back to the same place. So basically reinstall every three or four weeks.
 
Back
Top