We had an intermittent outage, but now the site is back up.

The outage was caused by an upgrade to the MySQL client libraries, which made them not want to talk to the server anymore. The error code was ERROR 2027 (HY000): Malformed packet. The solution is described on StackOverflow.

The resolution was to upgrade the DB password on the server to a new algorithm, which was a fix done in three minutes.

The hard parts were getting home, looking through all the logs, and finding the error :)

I got this message in "Chatterbox":

root says You've got a reply to Re: Perl Multi Processing. from Anonymous Monk at [no such node, ID 11139417]

Any ideas?

Just noticed, Selected Best Nodes currently has dem bones in 12^th place with a Rep of 137. This caught my eye because zombified accounts are supposed to have no write-ups. And no, this one doesn’t have any write-ups — the high-ranking node is the user page itself! I’m thinking something may be amiss here... ??

Thanks for looking,

ddg: schwartzian transform perlmonks

Google doesn't seem to have this problem (anymore), any chance to adjust the robot text?

google: schwartzian transform perlmonks

This node in

• bare
• mobile

update

Hmm ... these results with my FF searchbar are very different from the monastery markup [ddg://...]

https://duckduckgo.com/?t=ffsb&q=schwartzian+transform++perlmonks&ia=web

G'day All,

Would it be possible to truncate individual "Nodes To Consider" items that exceed a certain size.

This request was prompted by the fact that, for about the last day or two, the first three entries take up about 20 screenfuls (on my 27'' monitor with my browser extended to its full height). I have encountered single items like this in the past, but never this many: it makes navigation very difficult.

I'm not fussed about what actual size is chosen, perhaps 2,000 characters. I wouldn't want the number of lines to be the criterion: the second item, at the moment, contains a single line with 4,263 characters.

I'd envisage some simple line, following the truncation, which says something like: "Truncated at 2,000 characters, follow the link above to see the full post.".

Normal posts which exceed a maximum number of characters — I'm not sure of the limit; 64K? — have something similar. Perhaps the code used for that could be reused for this.

I notice that the monks appear to have succumbed to the popularity gods and implemented a similar system.

In all seriousness: there are only 2 kinds of readers:

1. . Folk who know their stuff, understand the situation, and properly cast a ++ vote.
2. . Ev everyone else - especially including trolls, angry fools, idiots who don't-know-shit, folk who do not like the subject itself or happen to have a different opinion to it, haters who dislike the poster or their alias, and so forth.

I propose to delete the -- option entirely.

Think really carefully about this:
a) what kind of person do you need to be in order to be qualified to cast a "++" vote?
b) what kind of person do you need to be in order that your casting of a "--" vote is on-topic and useful to the conversation and deserving to decrement the vote of person (a) ?

See my point? There's no possible way to ensure that only skilled down-voters are counted, so not counting any downvotes at all actually sovles the problem. Garbage still ends up with "zero" reputation as it properly deserves, and all the gems are all but guaranteed to float to the top - without risk of being scuttled by fools.

Just stumbled across this today...

When I click on any of the reviews in the Module Reviews section, I see, not the contents of the review, but metadata for the node. For example, the entry for the review of Acme::Apache::Werewolf by rob_au displays:

• node_iip: none
• package: 0
• type: modulereview
• core: 0
• nodeupdated: 2005-08-13 16:55:42
• reputation: 69
• locktime: 0000-00-00 00:00:00
• node_id: 307626
• author_user: rob_au
• hits: 2110
• title: Acme::Apache::Werewolf
• ucreatetime: 1069065933
• createtime: 2003-11-17 05:45:33
• postbonus: 1
• votescast: 69
• lockedby_user: none

The reviews in the Book Reviews section are displaying correctly.

Cheers,

id://11136579, doc://pack and mod://Text::CSV_XS no longer work in the search box. I just get redirected from https://www.perlmonks.org/ to https://perlmonks.org/.

id://11136579, doc://pack and mod://Text::CSV_XS still work in posts, though.

Perhaps broken by the Open redirects are disabled change?

I recently graduated to novice(?) and have been allocated some reputation points to bestow on other people. I read that there would be some +'s and -'s next to the posts in my threads. I've been unable to find them though but whether that's due to my mediocrity or your appallingly wonderful UI I'm not sure. I've tried reading around but either I'm too lazy or the explanations are too opaque.

Any chance of a link to a dumbed down version?

Regards, sbrothy.

Is there a way how to get the target link from a shortcut? I.e. I don't want to redirect to it, I just want to get the link back so I can open it myself.

I got this message yesterday

NodeReaper says I've reaped a node you considered: Re^3: Why does it seem as though Perl has the only community of friendly, non patronizing or demeaning, programmers? What is with every one else? -- because: REAP: No need for anonymous insults

Alas, the post was unreaped and there is no entry in janitation history by whom and why...

Voting inside longer sub-threads with many posts is easy, you select ++ or -- and hit return.

But the request will return at the very beginning of the page and you need to scroll again to were you started.

actually there are anchors allowing to return from where you voted, like

?node_id=11135115#11135141

one only needs to isolate the ID from the radio's name <input type="radio" name="vote__11135141" value="-1">

Minor complication is that one can vote on multiple nodes at the same time, identifying the one where the request was triggered might require some JS trickery.

update

In hindsight, implementing to vote with an AJAX request might be the better approach tho. It wouldn't require to reload the whole page and fiddling with the current vote mechanism.

Can someone with the proper access and authority please look into this? Thank you.