3 ms·

systemd really is fine until you have a hosed system or it does something wrong or a scary bug is deemed a feature by the totalitarian maintainer and then it's

by setquk 9y ago

systemd really is fine until you have a hosed system or it does something wrong or a scary bug is deemed a feature by the totalitarian maintainer and then it's like fighting your way through a dungeons and dragons adventure book written by someone who has been on the crystal meth.

ninkendo 9y ago
Could you try to make that point again but without the hyperbole and with some specifics? It's possible you're trying to engage in a real discussion here but it reads like you're just venting.
- setquk 9y ago
  I believe it deserves as much respect as the systemd maintainers treat their users with. 1. Hosed system down. All systemctl commands fail with "Failed to get D-Bus connection: Operation not permitted". Only option was to reinstall it from scratch, like a Windows machine 2. Does something wrong. So many random weird bugs like "failed to start reboot.service" ffs. Answer is always "upgrade systemd" which is a problem when you're running CentOS. 3. Totalitarian maintainer. Pwnies award for lamest vendor response 2017 covers my complaints there. Edit: just a point I wish to add. The idea that systemd embodies is the right direction, but the architecture and implementation sucks as do the maintainers.
  - chimeracoder 9y ago
    > 3. Totalitarian maintainer. Pwnies award for lamest vendor response 2017 covers my complaints there. It really doesn't. One of the pull requests they link to as the "lamest vendor response" literally has only three comments from Poettering, one which says "Looks great otherwise", one which says "excellent! thanks a ton for fixing this!", and one which answers a question about why a local variable is named MTU. Seriously, this is the thread itself: https://github.com/systemd/systemd/pull/6214 https://github.com/systemd/systemd/pull/6214 There's another linked thread where he admits that they misjudged a commit as irrelevant, but explains that they can't rewrite git history, which is unfortunate but hard to take serious issue with. And a third where he says that systemd will defer to validating input according to POSIX and other common cross-distribution requirements, instead of enforcing its own, stricter set of requirements that would actually break other systems. It's not even hard to find places where Poettering has said something that makes him look less sympathetic than this stuff. But referring to these as the "lamest vendor response" really just makes it look like they're searching for anything to take issue with for Poettering himself, regardless of the topic at hand.
    - the_why_of_y 9y ago
      Just to have some context, the very same Pwnie "lamest vendor response" award was given to the OpenBSD project in 2007.
    - jancsika 9y ago
      > It's not even hard to find places where Poettering has said something that makes him look less sympathetic than this stuff. Right, like this: https://www.youtube.com/watch?v=skPOeOPGpwM https://www.youtube.com/watch?v=skPOeOPGpwM That is a CCC talk where Poettering takes the presenter's invitation for a single response as a cue to interrupt the rest of his presentation slides. Responses like, "Do you hate handicapped people?" give you a taste of the general decorum with which Poettering handled himself. That's not to say that Poettering isn't entitled to some response, nor that all his responses in that video are unfounded. But interrupting a presenter's presentation is extremely aggressive behavior. Even the CCC presenter who trashed Perl 5 two years in a row in a fairly distasteful fashion was met with polite responses from the audience. What's amazing is that Poettering consequently wrote a piece blaming Linus Torvalds for the sorry state of communication in open source[1]. I'm certainly not a fan of Linus' outbursts. But an audience member making a hostile takeover of a tech presentation is the stuff of nightmares. Anyone capable of doing that has no business opining on the health of open source communication. [1] https://plus.google.com/+LennartPoetteringTheOneAndOnly/posts/J2TZrTvu7vd https://plus.google.com/+LennartPoetteringTheOneAndOnly/post...

by setquk 9y ago

ninkendo 9y ago

Could you try to make that point again but without the hyperbole and with some specifics? It's possible you're trying to engage in a real discussion here but it reads like you're just venting.

setquk 9y ago
I believe it deserves as much respect as the systemd maintainers treat their users with. 1. Hosed system down. All systemctl commands fail with "Failed to get D-Bus connection: Operation not permitted". Only option was to reinstall it from scratch, like a Windows machine 2. Does something wrong. So many random weird bugs like "failed to start reboot.service" ffs. Answer is always "upgrade systemd" which is a problem when you're running CentOS. 3. Totalitarian maintainer. Pwnies award for lamest vendor response 2017 covers my complaints there. Edit: just a point I wish to add. The idea that systemd embodies is the right direction, but the architecture and implementation sucks as do the maintainers.
- chimeracoder 9y ago
  > 3. Totalitarian maintainer. Pwnies award for lamest vendor response 2017 covers my complaints there. It really doesn't. One of the pull requests they link to as the "lamest vendor response" literally has only three comments from Poettering, one which says "Looks great otherwise", one which says "excellent! thanks a ton for fixing this!", and one which answers a question about why a local variable is named MTU. Seriously, this is the thread itself: https://github.com/systemd/systemd/pull/6214 https://github.com/systemd/systemd/pull/6214 There's another linked thread where he admits that they misjudged a commit as irrelevant, but explains that they can't rewrite git history, which is unfortunate but hard to take serious issue with. And a third where he says that systemd will defer to validating input according to POSIX and other common cross-distribution requirements, instead of enforcing its own, stricter set of requirements that would actually break other systems. It's not even hard to find places where Poettering has said something that makes him look less sympathetic than this stuff. But referring to these as the "lamest vendor response" really just makes it look like they're searching for anything to take issue with for Poettering himself, regardless of the topic at hand.
  - the_why_of_y 9y ago
    Just to have some context, the very same Pwnie "lamest vendor response" award was given to the OpenBSD project in 2007.
  - jancsika 9y ago
    > It's not even hard to find places where Poettering has said something that makes him look less sympathetic than this stuff. Right, like this: https://www.youtube.com/watch?v=skPOeOPGpwM https://www.youtube.com/watch?v=skPOeOPGpwM That is a CCC talk where Poettering takes the presenter's invitation for a single response as a cue to interrupt the rest of his presentation slides. Responses like, "Do you hate handicapped people?" give you a taste of the general decorum with which Poettering handled himself. That's not to say that Poettering isn't entitled to some response, nor that all his responses in that video are unfounded. But interrupting a presenter's presentation is extremely aggressive behavior. Even the CCC presenter who trashed Perl 5 two years in a row in a fairly distasteful fashion was met with polite responses from the audience. What's amazing is that Poettering consequently wrote a piece blaming Linus Torvalds for the sorry state of communication in open source[1]. I'm certainly not a fan of Linus' outbursts. But an audience member making a hostile takeover of a tech presentation is the stuff of nightmares. Anyone capable of doing that has no business opining on the health of open source communication. [1] https://plus.google.com/+LennartPoetteringTheOneAndOnly/posts/J2TZrTvu7vd https://plus.google.com/+LennartPoetteringTheOneAndOnly/post...