@karel I don't think software-installation applies here.

OT-Mint

seems dupe

dupe

@Natty report askubuntu.com/a/1247670/22949

[ Natty | Sentinel ] Link to Post Low Length; No Code Block; Low Rep; Unregistered User; 2.5;

no longer needed - the post is not really a rant, it's just a recommendation that's far outside of what the OP is likely to want and therefore NAA

That comment expresses a statement of policy that is at best ambiguous. I was a little reluctant to even vote to delete the post while that comment is on it -- though I did cast my delete vote.

not a link only answer. It has a wrong canned comment too.

dupe

@Kulfy Ordinarily I'd flag the wrong review comment as no longer needed, but do you think I should wait to do so, so that the reviewer can see your reply?

[ SmokeDetector | MS ] Link at end of answer (60): Install homebrew in Ubuntu 14.04 by maccevedor on askubuntu.com

@EliahKagan guntbert do act on their comments. So, I flagged the comment and pinged guntbert as well.

I'm not sure if the notification would still be displayed in the user's inbox after the comment being replied to is deleted, so perhaps I should hold off on adding my flag.

Doesn't matter if the comment get deleted meanwhile. guntbert would still be notified, I believe.

Okay. I'll go ahead and add my flag then.

[ Natty | Sentinel ] 18.04 PlayOnLinux asks for wine 4.2 By: Joney Caldeira 5.0;

@Natty tp

no longer needed. Typo was corrected.

@pomsky It has the one more downvote required to be deleteable now.

[ Natty | Sentinel ] chromium Ubuntu 20.04 - Google Chromium is slow after wake from sleep By: Petr Olexa 4.0;

@Natty tp

@Natty report askubuntu.com/a/1247751/816190

[ Natty | Sentinel ] Link to Post Contains Blacklisted Word - Can someone help; Contains ?; Low Length; Contains Salutation - Thanks; IntelliBL - 0.5; Low Rep; 4.5;

[ Natty | Sentinel ] anbox Anbox on 19.10 - claims kernel modules are missing By: s-k-y-e---c-a-p-t-a-i-n 4.5;

@Natty tp

@Natty tp

All comments are no longer needed

no longer needed

This answer and this answer seems to suggest same thing. Should one of the two be deleted?

@Natty report askubuntu.com/a/1247753/816190

[ Natty | Sentinel ] Link to Post Low Length; No Code Block; One Line only; Low Rep; 3.0;

OT Linux Lite

@Kulfy well, one of them says, in addition to all these packages, you need this one, and the other one says, now you only need this one. So, they are not the same I think

As per package details of linuxbrew-wrapper, those packages are dependencies of the linuxbrew-wrapper. So, it doesn't matter which one you follow, you'll still get same packages. I believe.

@pomsky @Zanna That is (and always has been) an actual answer to the question. It's even a correct answer. Killing sshfs will fix a significant fraction of hangs where nautilus cannot be killed even with SIGKILL(and even by root), i.e., the precise situation described in the question. Furthermore, of the answers that exist on that post that recommend specific actions to fix (rather than prevent) the problem, this answer is actually the most likely to work.

I've edited to clarify that it is intended as a solution to the specific problem described in the question, and I've also expanded it quite a lot with a detailed explanation of why it works (when it does). My explanation should make clear why sshfs can cause nautilus to hang unkillably while nonetheless itself remaining killable, and why killing it makes nautilus killable again. I very much recommend that this answer be undeleted. I've voted to undelete it and it needs two more votes.

Also, I hope my edit is clear. Believe it or not, I actually had to leave quite a bit out in order to avoid it being even longer. But I wanted to give enough information to elucidate why this really does work, as well as the principle behind it that is applicable to other situations and relates to other answers. If my edit is unclear in places or otherwise can itself be improved, of course I very much welcome feedback on that--or, of course, just submit another edit.

I think the relationship between FUSE driver malfunctions and seemingly hard-to-resolve IO-related hangs (especially in file browsers) is one of those things that may seem hard to believe, until explained in detail. (In this way, it reminds me of how .basrhc is sourced by some noninteractive shells and how Unicode actually smiles, rather than frowning, on curly apostrophes.)

Still, I anticipate that not everyone is always thrilled when I enormously expand their answer. Yet what I added truly is an explanation of the recommendation in that answer, and not a new answer. Also, the author of that answer deserves credit for the recommendation. I have personally run into this kind of problem with sshfs and file browsers.

[ Natty | Sentinel ] partitioning Ubuntu 19.04 | Hibernate not working - swap missing in /etc/fstab By: stephweb 5.5;

@EliahKagan thanks for the edit, it's undeleted now.

@pomsky No problem. Thanks for the help undeleting it. I'm not sure if I should flag all the comments as no longer needed, or reply to some of them, or what.

@Natty tp

@EliahKagan I feel all of them are obsolete now as your edit clarifies everything

dupe

@Natty report askubuntu.com/a/1247801

[ Natty | Sentinel ] Link to Post Low Length; Contains Whitelisted Word - You can; Low Rep; -0.5;

[ SmokeDetector | MS ] Link at end of body (99): Granite ME :- Numerous individuals need to improve by HiensGrieg on askubuntu.com

@EliahKagan awesome

I feel a bit worried about wrongly voting to delete that post. I wonder why I did that. It looks very much like something I would Skip

@Zanna I am concerned we are overly willing to vote to delete posts for quality reasons. (I do not intend this phrasing euphemistically; I am not excluding myself from this critique.) I don't disagree with the especially broad view Ask Ubuntu takes of NAA--that an answer that is entirely based on a total misunderstanding of the question is NAA--but I don't think we should be delete posts merely because they they appear to be wrong or because it is not clear why they would be right.

Even though not intended as such, the effect of flagging or voting to delete posts for those reasons is that flags and delete votes come to be used as super-downvotes, which is not what they're for.

(Though if a post is deleted for quality reasons or because it is so entirely wrong as to be useless, presumably it should have at least as many downvotes as delete votes... and on Ask Ubuntu this is rarely the case. I don't know if that reflects that deletion is being used where there's inadequate confidence to support it, or if it's just part of Ask Ubuntu's culture of not voting enough on posts, up or down.)

In the case of that particular post, I am not sure what happened there. At least four experienced users expressed that the post should be deleted, one by saying so in a comment and three others by casting delete votes on it. When a post says something that seems to have no connection to the question and thus seems like it can't be a real solution to the problem being asked about, the first thing to do is to look at the surrounding context.

In the case of that post, searching the page for sshfs should have turned up blkpws's comment about sshfs being the culprit, which is itself on an answer that describes doing something outside of Nautilus that fixed the problem--though it could be read as saying that it prevented the problem from recurring rather than that it immediately fixed it.

wrong/misplaced answer

@Kulfy It's just been self-deleted.

@karel I'm not sure if it's ok to report duplicate answers to Natty. There's a difference in NAA and a duplicate answer I believe.

@EliahKagan Okay.

@EliahKagan Also, one of the answers never works to solve the problem described in the question, though it is relevant to the general problem of Nautilus not responding. (If sudo killall -9 nautilus doesn't work, how would killall nautilus work?) I'm not advocating that it be deleted, and I haven't even downvoted it at this point.

Rather, I bring this up because it suggests that people have been looking at the answers one by one, rather than looking at them in the overall context of the question and each other.

For clear cases of NAA that may be fine (though still not ideal), but for subjective quality assessments that require a nuanced view of the technical issues... well, I just don't think deletion can be used reliably for that. But the closer to that one gets, the more important it is to look in detail at the surrounding context, I think.

@Zanna However, you (and @pomsky, and people not currently pingable here) are even further from being alone in this than I've been letting on. A moderator single-vote deleted a different answer on that question, years ago, which was (and is) an answer and should also be undeleted. (I presume it was in response to flags, and thus a mistake by multiple people, not just one.)

I've encountered weird stuff with sshfs a couple times, but I've ended up getting processes shown as being in the D state by ps (i.e., in uninterruptible sleep) to terminate by unplugging malfunctioning external disks quite a few times over the years, and I don't think I'm the only one. That's a real solution, because when you force the operation to fail, the system call returns. It shouldn't be use cavalierly, but it definitely qualifies as an answer.

The post should be edited, though, at least for formatting but probably to clarify it a bit. It would probably be good to do that before flagging it for moderator attention to request undeletion. I have not flagged it.

@Kulfy If it was intended as a success report or other commentary on another answer, then it could be reported to Natty. But if it's just an answer that covers the same ground as earlier answers, it's not reportable to Natty, even though such posts are sometimes appropriate to delete.

[ Natty | Sentinel ] software-center Problem with Ubuntu Software v 20.04 LTS By: Mohammad Shahjahan 5.5;

@EliahKagan Likewise, when a post is considered NAA here because it misunderstands the question and thus effectively answers something other than what was asked (which isn't considered NAA on most SE sites, I think AU may be the only one for that) but it doesn't have any other problems that would make it reportable to Natty, then it's appropriate to delete it and the NAA flag is the flag to use... but it's still not reportable to Natty.

I believe both of these situations are covered by Bhargav Rao's warning about scope creep. Natty is (and is supposed to be) taught to predict the likelihood that a post is NAA from the contents of the post, not from the contents of other posts.

@Natty tp

@Kulfy I was told to report duplicate answers as NAA by an Ask Ubuntu moderator. This is when I started doing it. Since then I have discovered that reporting duplicate answers as NAA is common on every Stack Exchange network site at which I currently review, so even changing this policy at Ask Ubuntu wouldn't effect the general policy on other Stack Exchange network sites.

@EliahKagan that would only confuse Natty :)

@EliahKagan Yeah but this particular answer doesn't look like a success report.

@karel I'm not asking to change policies or anything like that. I was just concerned with the impact of reporting such posts to Natty.

@Kulfy Natty originally started at Stack Overflow which is the most mature Stack Exchange network site due to its large size and large number of gold badge tag users who are also active reviewers. If Stack Overflow likes it, it's probably OK for Natty to like it too.

@EliahKagan should we flag that post to have it undeleted?

@Zanna Yes but it would be good to edit it first.

@karel Separate from the question of whether or not duplicate answers should be reported to Natty, I don't think this is actually a duplicate answer. If it is, then the current comment saying so is misleading and links the wrong original.

The answer we're talking about gives a different procedure to achieve the same goal, consisting of different commands, and cites it to a different source. These instructions also look a lot more pleasant to follow than the instructions in the answer linked to in that comment.

[ Natty | Sentinel ] command-line service command not found By: user1092071 3.5;

@Natty ne

^^ Considering systemctl is an alternative to service nowadays.

The same thing happens at Stack Overflow every day. User A posts an answer that includes A = C. A year later user B posts an identical code block answer to the same question except that A is replaced by B so now B = C. If answer this isn't deleted then that same question will fill up with more and more duplicate answers of the same type, so there's no logical way to control this obsessive duplication other than to delete disguised duplicate answers that masquerade as non-duplicates generally

Regarding the question of whether duplicate answers should be reported to Natty, though, I'm thinking that, since duplicate answers aren't treated differently here from the way they are treated on Stack Overflow, and Natty has been active longer and more intensively on Stack Overflow than here, it shouldn't be necessary to guess. But, searching SOBotics and SOCVR, I'm not finding anything either way, so I'm not sure what to think.

Part of the problem is that the phrase "duplicate answer" is used to mean more than one thing. Often it's used to mean that someone posts the same answer of their own on multiple questions, which is not what we mean by it.

It might be good to ask Bhargav Rao about this for Natty. The problem I'm anticipating is that there is nothing about the contents of a duplicate answers that correlates with it being one. That it is a duplicate answer is entirely an emergent property arising from the combination of it and at least one other answer.

I think Natty is deliberately designed to consider only the contents of individual posts (plus a small contribution by the rep of the user). Since Natty doesn't read questions and doesn't read separate answers together, it fundamentally, even at a theoretical level, can only be misled by having duplicate answers with no other problems be reported to it.

However, perhaps I am erring in some respect here. I've looked at Natty's code a good bit but not thoroughly read it, and even if I had, it's possible I could be misunderstanding something at an abstract level. I will hold off on giving conflicting feedback on that item (until either I understand why you think it's a duplicate answer or a clear answer emerges as to whether or not duplicate answers should be reported to Natty).

@EliahKagan sorry, you actually said so earlier. I should have read all the messages before replying to any of them. I have edited it (though it may well benefit from further editing...) and flagged it

@Zanna Thanks. It might benefit from further editing, but I think that edit is enough for starters and that flagging after it was fine. Thanks for editing and flagging!

@Natty ne

:) thanks for catching it

No problem. :)

Can I have an upvote on this CW answer? OP originally posted the answer in question itself. I have converted it into a CW answer and left a comment for them.

This answer could be a good comment.

@Kulfy Done.

@Kulfy Or it could be expanded with material from the linked Q&A.

Right now it's a link-only answer (it describes the linked resource, but does not include even a partial answer from it or otherwise) and thus NAA.

The linked question contains multiple answers. I'm not sure which answer to include.

@Natty report askubuntu.com/a/800175/22949

[ Natty | Sentinel ] Link to Post Low Length; No Code Block; One Line only; Low Rep; 3.0;

I think it's better to have a link to question in comments only.

Yes, probably so.

@Kulfy May be I should delete the answer and the comment. The problem was originally solved by Serg in comments and they have already posted an answer.

Yeah, actually that cites the comment even (though doesn't link to it).

@Zanna Your flag was handled fast! :)

@EliahKagan I have deleted the answer and replaced the comment.

@EliahKagan :D

@Kulfy Looks good, though the OP hasn't visited the site in over two years, so they probably won't come back to mark the answer as accepted. (So you may want to remove that comment after a couple days.)

They last visited SO on May 2. I hope they visit either of the site soon. :)

Ah. They may well notice the comment notification then.

Yup.

(unrelated) dupe

NAA

Query to find abandoned questions. Worth having a poke.

*When I mean abandoned, I mean when question owner never visited the site after asking the question.

@jokerdino I have modified query a little bit. I have taken only date into account considering there could be some difference in login timestamp and post timestamp.

Previous query returned 92 results and the modifies query returned 4308 results.

Nice!

we should give it a year before we assume they have been abandoned :D

If I consider 6 months, about 2.9k rows are returned and for 30 days ~3.6k

Should view count be taken into account?

I have it for checking if roomba needs any help.

But higher views means other people are possibly having similar troubles.

[ Natty | Sentinel ] themes Change a certain program GTK theme By: Antonio Micari 9.0;

Some Error Occurred on this post

@Natty tp

This return 2k rows

Why is this post getting caught?

I have been using query for helping roomba.

@jokerdino Which query?

@Kulfy data.stackexchange.com/askubuntu/revision/1248327/1535355/…

They never visited the site after posting that question

A few questions and then disappeared. :/

@Kulfy Older than one year? Sounds fair.

365 days, to be precise

and view count < 1000

That's close enough

Assumption: More comments might indicate that the problem was solved in comments only.

We can make CW answer out of that.

Yes, if there are any answers to be found there.

(unrelated) @jokerdino Would you mind converting this answer into a comment?

Thanks.

Done.

Don't the links not autoexpand anymore?

Or does that only happen inside the post?

like "POST TITLE" instead of https://postlink ?

yeah, I think it never does in comments ever

I suspect I had a script before.

Links are resolved in main and child meta. Post links aren't resolved to Title in comments. It won't be resolved in posts too if the post is on some other SE site.

I got the script I was using earlier.

I wouldn't limit the views as they might just indicate a misleading title

If only I could write a query to find all questions with misleading titles

Is there a regex for that?

@EliahKagan looks great!

This week's data dump to SEDE has taken place

the query to find posts with code fence issues is returning 1775 items

[ Natty | Sentinel ] boot grub version 2.04 not supporting USB keyboard By: Volkodav 5.5;

[ Natty | Sentinel ] usb My Lifecam HD-300 is not recognized By: moep 5.0;

[ Natty | Sentinel ] ssd How to securely wipe my SSD before reinstalling an encrypted Ubuntu? By: Bash Stack 4.0;

@Natty fp

no longer needed - even before the edit, that wasn't a link-only answer

causes harm - We don't only support 20.04, and that answer removes information for previous releases.

* and that edit removes

@Zanna Did you regenerate the contents of etherpad.snopyta.org/p/fence?

It seems to still have some of my old notes (and a note from you, I'm not sure how old), so I'm thinking not. Should its contents be replaced with the result of filtering the query with that shell script? Or should a new Etherpad be created?

I think it would be best to use an even less picky query to generate input to the script.

If I could, I'd give the script every post that appears to have a code fence in it anywhere. I think that may be too many for SEDE to return, though.

[ Natty | Sentinel ] email How to set up mail server (postfix or anything else) on Ubuntu? By: b1nch0 4.0;

@EliahKagan no, because I was not sure what you wanted to do next

@EliahKagan was that not what the query returned before you started adding extra LIKEs to it?

@Zanna No, the LIKE expressions are separated by OR operators, so with fewer LIKEs that query is more restrictive.

Oh. Sorry. Do you mean the version of the query you wrote?

yes...

data.stackexchange.com/askubuntu/revision/1246925/1535435/… <-- 13027 results

Yes, I think so. You'd have to ask you about that. :)

Sorry.

:D

@EliahKagan oh I see

Yes, let's use that, if we can.

Also, if it works, I'd be interested to look for ~~~ as well.

Accommodating ~~~ will require a modification to the script, too, but that should be no problem.

I get 13427 and I'm suspicious, like how can there be exactly 400 more? Did I not just misremember the number?

[ Natty | Sentinel ] server Is it possible to perform speed test with any Ubuntu server? By: Benji Bear 3.5;

ok it actually seems to work

data.stackexchange.com/askubuntu/revision/1246925/1535442/…

@EliahKagan I think we can just replace the contents of the Etherpad with what we currently want to work on as we can always use the old query or write a new query to find whatever stuff we are interested in

Sounds good.

:)

@Zanna Are there really fewer than 14,000 posts with code fences?

I think we have about 800,000 posts in total.

I guess it makes sense that code fences are rare, since for most of the history of the site they haven't been supported.

[ Natty | Sentinel ] sound Internal microphone unplugged on HP Zbook G6 on Kubuntu 19.10 By: Ankit Bhargava 5.0;

@Natty tp

@Zanna Sorry, I didn't pay close enough attention and didn't notice that message.

If you mean there are exactly 400 more posts with ~~~ but not ```, I don't find it all that surprising. We've been dealing with a lot of numbers, it's not that unlikely that one of them happens to be congruent to 0 modulo 100. :)

Also, I just managed to write "```" in chat, which is a very proud moment for me.

not NAA - this is the only answer that mentions nvidia-detect, and even without details of how to use, it's still informative

similarly not NAA, though this won't work on newer releases

attempt to reply

[ SmokeDetector | MS ] Link at end of answer (60): Voice control over desktop environment ✏️ by Paul Schneider on askubuntu.com

Trello board | guidelines for it | feedback query | off-topic | historical

dupe

unclear

unclear

dupe

I find it rather alarming that this on-topic question has four close votes as "Not about Ubuntu." It's asking how to make a video file smaller in Ubuntu. It shows specifically what they tried. It explains their goal to avoid being an XY problem; that's the significance of the mention of Yahoo! Mail.

People can have actual lives with goals that are external to Ubuntu. Narrowly scoped, answerable, technical questions about using Ubuntu to achieve those goals are still on-topic.

[ Natty | Sentinel ] chroot GUI applications and chroot By: Darin 4.0;

@Natty ne

[ Natty | Sentinel ] multiple-monitors How do I get the Ubuntu Dock off the secondary monitor? By: Steve 5.5;

@Natty tp

[ Natty | Sentinel ] networking activation of the network connection failed. login screen disappeares By: frankmcpeeters 4.0;

@EliahKagan That question was wrongly closed and I've voted to reopen it.

@Zanna Contents replaced.

Here's the updated script that I used to process the query output:

#!/bin/sh
# filter-fences - Uses csvgrep to find posts that may have broken code fences.

# shellcheck disable=SC2016  # The operand to csvgrep -r is meant literally.

csvgrep -c Text -r '(?mx)  # Match ^ on each line. ("." does not match "\n".)
            ^[ \t]*(([`~])\2{2,})(?!`)(?!.*\1)  # Opening line of a code fence.
            [a-z]*                              # Lowercase letters, maybe.
            (?:[^a-z\s]     # Either (a) non-whitespace non-lowercase, OR
              |[ \t]+\S)    #        (b) blanks followed by non-whitespace.

There seems to have been a change in the way a PostId is converted to CSV. It happens for old queries too, so perhaps it's an intentional change by SE. I used to get stuff like:

"{
  ""id"": 20036,
  ""title"": ""Differences between doublequotes \"" \"", singlequotes ' ' and backticks ` ` on commandline?""
}"

Now, I just get:

"1247774"

This broke the old script. The new one, shown immediately above, is simplified to accommodate that, and separately made more complicated to also support ~~~ code fences, since your new query returns them.

Regarding this script, I've changed it to this (which produces the same output):

#!/bin/sh
# filter-fences - Uses csvgrep to find posts that may have broken code fences.

# shellcheck disable=SC2016  # The operand to csvgrep -r is meant literally.

csvgrep -c Text -r '(?mx)  # Match ^ on each line. ("." does not match "\n".)
            ^[ \t]*(([`~])\2{2,})(?!`)(?!.*\1)  # Opening line of a code fence.
            [a-z]*                              # Lowercase letters, maybe.
            (?:[^a-z\s]    # Either (a) non-whitespace non-lowercase, OR
              |[ \t]+\S)   #        (b) blanks followed by non-whitespace.

@Natty report askubuntu.com/a/1247997/22949

[ Natty | Sentinel ] Link to Post Low Length; No Code Block; Contains Salutation - thx; Low Rep; 2.5;

@Natty fp

no improvement

[ SmokeDetector | MS ] Link at end of answer (60): Bind9 - openssl problem ✏️ by Darxus on askubuntu.com