We're not perfect. There are a lot of items that need work, and a lot of different things that are broken in minute ways. They are the growing pains of a publicly developed website. We are in a constant state of beta, but hey, it keeps things exciting. We'd like you to report any problems that you're having, because we, the caretakers and floorwashers, like this to be a generally pleasant experience for you.

Although we're working on a ticketing system, for the moment it is not complete, so please for now just tell us here where we messed up, and hopefully it will work out.

Things that probably are NOT bugs:

• "Error: Timeout/Server Error/Unknown Error" which appears after a vote or message takes a while to send. These indicate either the server took too long to finish a request or you lost the network connection midprocess.
• "Sumbit", "stumbit" and other such creative buttons are one of our little in-jokes. They're intentional.

Please report the following things whether or not they seem relevant:

• Page the error occured on
• A screenshot if it's hard to describe. Post them in Picpaste.
• Your theme
• Your browser
• The time it occured, down to about a half an hour block. We don't need a stopwatch and second, just about when. We'll do the rest.

Hopefully with all of your eyes and ears out there, we'll keep this place running smooth as um, an active website.

Thanks

The umlauts in Prügelknabe and Käsewähe make it impossible to get to the node via normal links. They also don't appear in searches if you try to find the node by using the non-accented character.

This is a known issue relating to double-encoding of non-Latin-1 characters. Anything outside of the lower 128-bits, for the time being, should have its title encoded using HTML entities. For instance, using ä (ä) instead of ä. Once the issue with generating links to exotic characters is fixed, should it be desired, we can programmatically search for and replace these entities. Doing the reverse is not nearly so easy.
-om
There was an issue with our version of Perl and our version of CGI.pm not playing well together. Fixed.
-om 2010 April 25

Awesome! Out of curiousity, is there a way to get the search function to recognize "a" as being similar to "ä" and etc? In case someone can't/won't/doesn't know how to type the extended characters? - 4/25/10

It's OK for the Notifications nodelet to appear in Chatterlight, but the Notification Settings popup, with all of the attendant checkboxes, appears by default. It should appear only if someone checks the settings link, but it appears without asking, covering up the natterbox content (whatever that's worth). You can dismiss the settings popupby clicking on the "x" but it always comes back on the next reload.

Update, Feb. 23, 2011: No, not fixed. To clarify, when:

1. Notifications nodelet is enabled
2. No notification categories are enabled

the notifications nodelet's settings popup appears whenever chatterlight (or any other node for that matter) is opened for the first time in a browser window or tab, covering up whatever is below it. It does seem silly to have the nodelet enabled if you don't have any notification categories enabled, but it's still annoying. A message to the effect of "You have no notifications enabled" would be less annoying.

We don't feel inclined to change this: if you don't want notifications, you don't want the nodelet. If you don't want the nodelet, you can remove it immediately with the link at the bottom of its settings. If you do want it, you can immediately decide which notifications you want. If you want to be able to add and remove the nodelet without going to Nodelet Settings you'll have to wait a little. --DonJaime

Update, December 29 2011: DonJaime was talking past me. This is the sort of high-handedness that drives people from this site.

I am not, repeat NOT, talking about the nodelet itself. I'm talking about the stupid Notification Settings popup dialog. The Notifications nodelet contains a link that you click on to expand the dialgo to full extent. But for some reason, it's expanded ALL THE TIME, by default. This covers up areas outside the nodelet, sometimes submit buttons and links in other nodelets.

I don't want to see Notification Settings cover up parts of the screen it shouldn't be covering up. I don't want to have to dismiss the dialog every time I load a page. FIX IT.

Categories nodelet styling is either nonexistent or broken. I get bulleted categories and an inline scrollbar. Yuck.

kthejoker is using zen with Kernel Blue and no style defacer settings. He probably means categories, which shows the behavior he describes for me under different conditions (SharQ's Sharktank, a bunch of style defacer junk, Firefox 3.x) which, as far as I'm concerned, is correct behavior.
-om 2011 Feb 16

He's right, that's fugly. Why is it set (with inline styling!) to height:15em; overflow:auto? That's gruesome. Also, should probably have the class linklist (which in some stylesheets, at least, will remove the bullet points). - Oolong 2011-02-06

It wasnae me, but I should think the height was set to avoid it taking over the sidebar if the user can add to a lot of categories. It is indeed fugly, though. But hasn't it been superseded by that jolly little category add widget with the menu in it? - DonJaime 2011-02-11

New issue:

Arrows for collapsing nodelets are hard to see (at least in Kernel Blue.) Can their color be made to match the nodelete title text color?

Also, when I search for this node . I got a 404.(The link is just to the period, nothing else. Oh wait that isn;t really a bug is it? More to do with putting a period in the url and that not working. Hmm... maybe code it so it redirects to the html code like other symbols seem to?)

Also, users homenodes don't adjust the most recent writeup written for self-nukes.

Also, new user mnlqrzflrs has a bookmark in heir homenode keep holding that doesn;t lead anywhere.

Writeups by users hidden using the Pit of Abomination are still showing up in Cream of the Cool on the front page.

Cream of the Cool, along with most of the things on the front page, is generated cached once every few minutes, and displays the same for everybody. It is also a particularly expensive query that it runs, so changing it to running dynamically wouldn't be a great move either. Some solution similar to the feeder/filter method used right now for New Writeups might work, but I'd be concerned about the database load. -om

It might be a bit mad, but we could do it reasonably cheaply with CSS by giving titles and things 'by-user' classes, which would also have the inestimable advantage of allowing people to automatically embellish entries by their crushes with a purple background and little hearts and flowers. Just if anybody didn't have anything better to code... - Oo

Reports of chatbox fail

Some users have reported problems with the chatbox after trying to chat and getting an error:

1. User enters pearls into the box, hits return.
2. A neat error message appears.
3. User hits return again.
4. Error message reappears.
5. Repeat.

Loading a new page and starting again is the only solution.

Also: sometimes an error is reported although the message has been sent. At times the error appears in the chat-box quite a long time later.

DonJaime's speculations

I can't reproduce these bugs. I suspect but am not sure that some of the time it's a matter of the interaction of user expectations and server errors:

1. User sends a message when the server is unhappy.
2. Occasionally, the message gets through, but even then the server drops its response (updated chatbox) on the floor
3. User tries again while server is still acting up.
4. Meanwhile, the chatterbox is not updating because the server isn't providing the info.
5. Server gets happy again. Meanwhile, user has given up and loads a new page. Or if the user is still on the same page and their message had got through, it now appears in the chatbox.

As I say, this is only speculation, but it's consistent with the information I've got so far. If you can provide further information that indicates this is not what's happening, that would be welcome, too.

... and indeed such information has arrived. Sometimes after an error on Safari on Mac, trying to send a message again reproduces the same error so quickly that there hasn't plausibly been time for the server to have been involved. Apple has a history of brain-dead caching, but caching errors would be a new low. I've asked for I asked the user to try a minor change to the message next time the problem arose, that didn't help, so it doesn't seem to be brain-dead caching.

We also have a one-off unreproduced report of: 1) Server was being slow/wonky 2) Submitted a message which came back with an error 3) Cleared the message, (still on the same chatterlight page) and typed something else 4) Server came back, the something else posted. 5) I let automated refreshes update the nodelet for a while. 6) I submitted the form with the message box blank and then my ooold message when the wonkiness was happening went through.

Okay, I think this may be a homenode bug only.

On my homenode, I've got a section of stuff that I've written, with a header of "I've written some stuff I think is pretty neat:" This section has suddenly gotten vastly truncated.

When I go to edit my homenode, all the text is still there, but once I click "Sumbit," large chunk of that text vanish.

For example, the "Blasphemy" section now goes from "Everything King James Bible" to "Tarantula" -- which is actually a not-particularly-blasphemous movie that was listed under "Film." Everything between those two items, which includes the sections for "Comedy," "Comics," and "Fiction," are not visible.

Likewise, under "Inside Everything," everything between "Favorite Everything Quotes" and "The Voice of Thunder" under "Personal Stuff" vanishes.

I think I know what's where the problem lies -- the first missing section runs from "getting to know you in the Biblical sense" nodes to Tales of "The Rocky Horror Picture Show", and the second runs from the second listings for both of those writeups. In other words, if { functionName:arg1,arg2,arg3 }%there's a square bracket followed by quotation marks, everything after that disappears, until it hits quotation marks followed by a square bracket.

I don't know why that's happening, nor do I know how to fix it.

This one is my bad. I recently added a call to screen homenode text through stripCode() so that homenodes which had code in them, but then had code evaluation turned off, will display normally without any editing. Most of the time when we embed code on E2, it's in blocks like this:

[%
 ... code ...
%]

Or like this:

[{ functionName:arg1,arg2,arg3 }]

However, we also support Perl-interpreted strings like this:

["
 ... interpreted string ...
"]

Since this last feature is rarely used, I might just want to remove it from the stripping process by default.

-om 2011 Apr 19

This is a somewhat niche thing, but when I'm using my iPad and a Bluetooth keyboard in the catbox, it registers key presses as doubled ALL THE TIME. It also then double-registers the backspaces so correcting each can take like 5 cycles. I think this is because there's a lot of JScript stuff going on with the catbox and iPad Safari just can't keep up. It also loses the insert point when the 'box autoupdates and scrolls. The double-type doesn't happen when editing nodes, just in the catbox (with autoupdate on).

Today, I used Internet Explorer 9, (because it is the default...for updating purposes), and it was loaded, so I clicked my favorites, and E2, all was well until after sending a thank you message to NanceMuse.  I then noticed it said 'you know if you registered you could talk'. 

Well, I logged in again, and sent another thank you to Montecarlo, and the same thing happened.

Logged in again, and asked anyone in the catbox if they had trouble with ie9...same thing, logged off.

So, I came back on with Firefox 4, all is well.  None of the messages, or chat had gotten through from the attempts while using I.E.9.

So, there it is, hope it helps.

The last seen stamps on private messages are updating incorrectly. They're jumping from 7 minutes to 3 minutes while the user is obviously still there, and sometimes not updating at all, even on refreshes. Might only be a chatterlight thing. I'm in Firefox.

Here's an explanation for why you're seeing what you're seeing:

Timestamps don't exactly change. The way the message list in the chatterbox (chatterlight is more or less a fancy wrapping around chatterbox) is that it gets the list of current messages on first load. The last seen times there reflect whatever the last seen times were at the time they were grabbed. If no new messages come in and you delete no messages, those times will stay the same, regardless of whether that person logs in/out after that point. The only thing that causes that list to update is if there is a new message to display.

If you receive another message from the same person, that new message will scroll in (potentially scrolling out an older message), and it will have a last seen time on it which reflects the last seen time when that message was loaded. Therefore, you can have two messages from the same person showing different last seen times.

I understand this creates a non-intuitive display. The simple solution of just replacing the list every time, however, can cause distracting flicker. We'd have to write some special code just to update the last seen display.

-om

I use Chatterlight, Chatterlighter and My Chatterlight. I don't remember all the features for the three, but this evening it seems none of them can bring up both chat and private messages. I'm sure at least two of the three provide that service, but am pulling out grey hairs to remember all the options. But my instincts say something isn't quite working.

Private messages are provided by default in the chatterbox nodelet, which both chatterlight and chatterlighter use. However, if you have the Messages nodelet disabled, the chatterbox nodelet does not display private messages, so you don't get them listed twice. However, in the case of chatterlight and chatterlighter, this means you have to go to another page to view your private messages.

This is known behavior. My thought on the matter are that (1) chatterlight should be retired and (2) chatterlighter should have a tab for the Messages nodelet if it's enabled.

-om 2011 June 29

I have a reward up on Everything's Most Wanted... For reasons unexplained it's up there twice.

I have a suspicion about this - the second, lower reward used to be my ransom for Lake Baikal, but when a user cashed that in it didn't leave the page... Now, since I've added a new reward, it's entry on the page has become a duplicate of the new one... There's a fuxed indexing table somewhere, methinks.

Not quite what Auduster suspected. I was on this one and forgot about it. It's actually pretty simple, and due to the fragile design of Most Wanted. Whether a user has a bounty is tracked in their VARS. If this value somehow gets lost from somebody's VARS, when they award a bounty, the bounty won't be removed. Most Wanted just needs to be made a little more robust in this fashion.

-om

User Clockmaker reports that the achievement "Printing Press", which is supposed to be awarded for writing more writeups than any other noder in a month, does not appear to function. The code in question is located in writeupsMonthMost and simply returns 0. It was likely disabled after the switch to InnoDB, as aggregate statements, like the GROUP BY that was inevitably used, are far more expensive with row level locks.

This should be a monthly cron job (or implemented, equivalently, in a nodelet) which queries all writeups published in a month grouped by publisher_id, finds if there's a clear winner (SELECT COUNT(node_id) 'num_published' ... ORDER BY num_published LIMIT 2, if the two num_published don't match, there's a winner; or do something more complex to award to all people who tie for most that month), and awards the achievement appropriately. This is a good candidate to use the slave MySQL DB since stale data is unlikely to have an effect.

---

2011 Nov 28

Users IWhoSawTheFace and Tem42 both reported difficulty viewing their own homenodes.  This is most likely due to timeouts, being exacerbated by the presently overtaxed servers.

Diagnosis is difficult since neither user has provided further details in an E2 Bugs entry. If you have a problem like this, please report it here and provide time you were accessing your homenode (including timezone) and browser(s) you were using.

In general, a large number of slow pageloads right now are for user homenodes, whether accessed by the user erself or by others.  Most likely, accessing it by erself tips things over the edge because the counting of messages to oneself is fairly expensive.  Profiling homenode display would be wise.  The issue may simply be that there are many htmlcodes used only in homenode display, so a bunch of code has to be loaded from disk, and homenodes are relatively rarely displayed, so we might want to trim them down.

---

2011 Nov 28

Now that htmlcodes are being cached, the voting footers (maybe still in voteit ...?) accumulate text from each previous run.  This causes a confusing display esp. on E2 Bugs because the "(unvoteable)" text repeats a dozen times by the final writeup in the node.

Fixed!

---

2011 Nov 28

Using the ABC button to strike out text in the WYSIWIG editor causes the text to appear struck out in the editor but only in the editor. If you click on the "HTML" button, it shows no tag has been added, and submitting effectively erases the formatting.

---

2011 Nov 29

The indent button also appears to do nothing, reflecting only upon first application, but not actually changing any HTML.  It should create a blockquote.

It seems that nuking a recently-C!ed writeup causes it to remain in the Cool User Picks table, marked as having no chings. (C.f. at the time of writing »The Silence«, by Glowing Fish.) I expect you've noticed this already, but have a formal bug report.

I know someone has pointed this out somewhere, but pressing "publish draft" entails losing all the changes you've made since you last pressed "update draft". This is kind of unintuitive, as you'd think publishing would perform both functions and it's easy to forget to update if you're just making one last small change or working inside the WYSIWYG.

2011-12-07 OldMiner: riverrun's writeup somehow got posted to an e2node other than the one in which Webster's writeup already resided. I merged the two e2nodes, but that shouldn't be able to happen. riverrun, is there any chance you could submit an E2 Bugs report.. 2011-12-07 OldMiner: ...about how you posted this?

SOITenly. Unaccustomed as I am to all the NEW ways we have to fling stuff on the wall, instead of searching for a TITLE first, the way I used to post, in which case--in the olden days--if a writeup didn't exist you could CREATE one, I published my Draft under its original title Hooked. This gave me a dialog box--did I want to add it to the node called "Hooked?" Being fond of my titles, and not wanting to live at the bottom of what I imagined was a whole LIST of dreadful experiences with drugs and women, and being plagued with HORRIFIC timeouts, etc. (which was beginning to get to me), I decided to post (sorry, PUBLISH) the thing as "Everything's Different Now" (I believe it was).

I liked that title for about five minutes and decided to go back to Hooked after I realized Hooked was just a Webster Definition. It LOOKED to me like all the rigamarole we used to have to go through to change a title had been simplified, and I was on my way out the door, so I just followed my nose/the prompts, so to speak.

I should mention that I had no less than 20 timeouts, 502s, whatEVER these things are during this process (forGET about checking links, adding links, etc.), so time had become of the essence if I was gonna get the thing up before I left for the rest of the day.

So:

1. Lazy author, not checking to see if node exists with CONTENT the important part of existing writeup.
2. Title change.
3. Interface annoyance.
4. Lazy admin not double-checking new method of title changing.

I was curious if the very act of POSTING/PUBLISHING/GETTING SHIT DONE here is any easier than it used to be.

I guess, for the moment, the answer is both yes and no.

n.b. And OF COURSE, posting THIS has given me that dumb hamster riff, which gets old.

I'm having trouble posting to the everyone account. I clicked "advanced options" (after finishing my draft) and checked the box for publishing under as everyone.

then nothing happened. the writeups listing did not include the new node and the node i created had "("Code Year" is also a: writeup.)" on the top right corner, above the add to categories thing. on the top left corner it had "!!! bad WU type: 0 !!!".

help?

Newer problem: 4/6/2012 report date: when visiting E2 (before logging in) there is a reduced version of the cooled items on the front page. The top one, for the past few months, has been Naughty Meter Maids. There's some rather profane language in the first few sentences (ie where onlookers at work would first see). Can this prank be corrected?
Mr. Bonci,

submitting an error watch here because you asked me nicely to.

1/5/2012 at 10:03 pm a user failed at linking to an outside source. User used an "a href" tag which had a "less than bracket" at the end, right before the "less than bracket" a "greater than bracket".

below the "talk" button for the chatterbox, and above the "chatterbox settings" topic section, I noticed an error. Upon mousing over any active link in the chatbox, lines of white space were appearing in the nodelet. The whitespace appeared to correlate to how long I mouse-overed a link. If i refreshed the page, or talked, the whitespace reset itself. Error did not perpetuate after the broken a href link had cleared the immediate chat area.

I am using Internet Explorer 9. Specifically 9.0.8112.16421

Good hunting,

Muse

This is such a small thing that it seems churlish to complain, and I don't even remember if browsers provide an interface to scroll textareas. But it's bothered me for nearly a decade now. If they do, it would be very nice if, upon updating a draft, the textarea in which the draft is composed were scrolled to the bottom automatically, rather than the top, given that that's where you're likely to be concentrated.

This is a particularly odd one - searching for buffalo buffalo buffalo buffalo buffalo buffalo buffalo brings up a Nothing Found page - though searching for buffalo or buffalo buffalo buffalo buffalo buffalo buffalo buffalo buffalo correctly take you straight to those nodes, while buffalo buffalo buffalo buffalo buffalo buffalo goes to a Findings page.

OldMiner explains that this is because when E2 finds a draft you can't see, it returns Nothing Found rather than Findings. Which should be easily fixed.

Also - it turns out that if someone includes broken comments in the form <!--h5> ... <--/h5> that totally breaks the page, rather than just commenting out the rest of the writeup as it probably should, in principle.

Gigantic Code Lister is broken.

As I'm returning nodes, I'm noticing that the metadata (votes, XP, etc) is missing, meaning they're getting voted and C!ed up again. Should I be concerned about this as gaming the system or shrug and accept it?

alex sez: Known flaw in procedure design which may be addressed some day.

So... back in the day when dinosaurs roamed the Earth, my username did not have underscores except for /msg purposes, which was annoying, so the username/alias fab was created to bypass all that typing. Upon my return to somewhat more regular E2 usage, I couldn't log in until I noticed the underscores in the username on my homenode. No idea when that happened, not a fan. This feels like an incredibly ridiculous and petty thing to post a bug report about, but it is making my brain itch and I'm hoping it's an easy fix maybe?

Muchas smoochas,
-the not-as-fled noder once /msged via fab

P.S. Update/Clarification: A friendly E2god tried to help on April 9, removing the underscores from the username, which also restored the fab alias to usefulness... but this morning, the underscores were back!?! Why are the underscores back? I blame gremlins. Please help!

P.P.S. A data point from my much-neglected Message Inbox suggests that maybe I sprouted underscores sometime between my birthdays in 2006 and 2007. Lometa's greeting in '06 lacks underscores; in 2007 she included them. No idea whatall went on in that time that could've caused the nodegel to magically regenerate underscores five years later, but there ya go.

This has to be a bug - according to the options for Notifications there is supposed to be an alert for a /borging, which I have had switched on, but never received such a notification, unlike for /drag, which does work. - patched

The little drop-down for removing to draft does not carry the deletion message on to the next step. Will provide screenshots if desired. Maybe a Chrome/Vista thing.

Also, I am still having problems with softlink creation, and given that this is one of the features that drew me to E2 in the first place, would love to talk with someone about it. It seems to be Chrome under Vista, if that throws any light. Is it some sort of really bloody weeird cookie thing?

While we're at it, it seems users using Internet Explorer 7 can't create nodeshells after searching for a node then hitting "new node". When I try it, I keep getting the "Hmm, that's strange, there's nothing there" message or it returns me to the front page. Same thing happens when I try and create a "new draft" after doing a search.

I don't know if this is related but I'm using Explorer 7 and can't seem to delete nodeshells either.

Just an FYI: Lately I've frequently been getting "that weird screen" when I click on my PC's link to E2 (see attch image, if needed):

http://i.imgur.com/UwLgJ.jpg

Don't know if it is E2 programming or my computer ...

*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~

And on May 27, 2250 PDT I got the "weird screen" again AND the Everything2 and Ascorbic links on my PC started having an icon of a red square with BF in it ... that is probably something on my computer, but I thought I'd share the info anyway (I thought perhaps maybe Jay Bonci had coded that as a reverse Facebook thing, but some of the guys in the catbox said No.) Here is the link if you want to see it:

http://i.imgur.com/oc5OP.jpg

I traced this problem to buzzfeed.com (a site I visited thru a link someone sent me to look at some stupid awkward cat sleeping positions!) and "fixed" it by deleting the E2 icon from my favorites bar, re-opening and adding it back. This was not an E2 bug problem, but since I'm not sure if anyone has read /will read this bug report node, I'll probably leave this here for a while and then I'll delete it.

I just updated a draft, and it did not load. This happens. BUT! usually I just hit back, and it pops up with all the changes ready to post in the text edit box. This time it did not do that; it asked for me to confirm that I wanted to resubmit data, and to do so by hitting the reload button. It did not give me another option. Hitting the reload button lost all of my changes. I am using Chrome; I would very much like to know if this is something that Chrome has decided to do, or if it is something to do with changes to E2. Either way, it stinks.

---

Usergroup leaders do not appear to have a way to edit the user group descriptions.

---

The vote count in the header ("you have 2 votes left today") runs one page load behind. So if you spend two votes in a node, it will not record this after you vote, but only after you move onto the next page. Not a major bug, but I thought I'd mention it. I do have AJAX turned on.

---

If bug reports that can not be worked on quickly (or are not worth working on) could be deleted (and recorded elsewhere as necessary), that would make posting a second bug report easier.