akw_kovu said:
Safari 8 on Mac OS X still isn't working. Still getting the NSPOSIXErrorDomain:100 when trying to access the site.

I can only guess it has something to do with the force forwarding to https. No other site is giving me an error message. Only e621.net.

http://forums.macrumors.com/showthread.php?t=1759189

This forum post is... so upsetting. It's... it's like I just chased my own tails. No furry joke intended.

akw_kovu said:
Safari 8 on Mac OS X still isn't working. Still getting the NSPOSIXErrorDomain:100 when trying to access the site.

I can only guess it has something to do with the force forwarding to https. No other site is giving me an error message. Only e621.net.

Does this happen on e926 to you as well? (Make sure you go to an https page)

If anyone experience this problem could get in touch with me to test this out, (either through dmail or on our IRC), I'd be ever so happy. I believe I can resolve this by poking around in our ssl configuration, but I don't have any way to test. :)

spight said:
If anyone experience this problem could get in touch with me to test this out, (either through dmail or on our IRC), I'd be ever so happy. I believe I can resolve this by poking around in our ssl configuration, but I don't have any way to test. :)

Probably related to that but I've been getting this on Opera (both PC and Classic - v12.16) since the morning...

http://i.imgur.com/o4MuYcB.png

Edit: It also took a lot to post this

spight said:
http://forums.macrumors.com/showthread.php?t=1759189

This forum post doesn't help at all.

spight said:
Does this happen on e926 to you as well? (Make sure you go to an https page)

No. E926 works fine.

And I can add that I am not using any firewall or other security software which could interfere.

I know it have been reported few times that something is wrong with tag history but I don't remember anyone going into details with this.

Lance_Armstrong said:

Tagging histories appear broken.

Bug:

Tagging histories appear broken.

Actual behavior:

Tags that were on the post before I tagged are shown in my edit.

https://e621.net/post_tag_history/index?post_id=535104

I didn't add multi_limb or change the pokemon spelling. I'm not 100% sure they were there when I hit edit.

Steps to duplicate:

Write down current tag list before you make a tag edit to compare.

I think what you're seeing is that implicated/aliased tags aren't showing up in the initial tag history, but are showing up once there's a new entry in the tag history. The tags on the post itself SHOULD be showing up correctly (all aliases/implications applied) on initial upload, but the first entry in the tag history itself will show only the tags that the uploader entered in.

Yeah it's kind of weird and not really ideal behavior, but at the same time I don't think it's anything too terrible. Still something that needs to be addressed at some point though.

akw_kovu said:
No. E926 works fine.

And I can add that I am not using any firewall or other security software which could interfere.

That's pretty good news, actually. It means the issue isn't the server or the install of nginx itself, but most likely just the cipher configuration. Our version of openssl is most likely just too old.. (We opted to downgrade rather than to upgrade during the heartbleed ordeal).

I'll need to recompile some stuff in order to get to another point. I'll probably try to take care of that sometime when I can bring the site down for a few minutes.

Char said:
I think what you're seeing is that implicated/aliased tags aren't showing up in the initial tag history, but are showing up once there's a new entry in the tag history. The tags on the post itself SHOULD be showing up correctly (all aliases/implications applied) on initial upload, but the first entry in the tag history itself will show only the tags that the uploader entered in.

Sounds exactly it. pokemon -> pokémon, multiple_legs -> multi_limb, etc.

Lance_Armstrong said:
Sounds exactly it. pokemon -> pokémon, multiple_legs -> multi_limb, etc.

Blame me for that one. Sorry. I'll fix it someday. :(

Just a quick heads up regarding Safari.

All iOS 8 users now can't visit e621.net anymore with the same protocol error.

Can you folks visit e621 on iOS8 now?

Well... I don't know why the Safari issue is happening... I'm still working on it, but I'm not confident that it will be solved in the immediate future. :(

Ok I need to read again "how to comment" because I wrote something useless, so I want to say Sorry for my last silly comment , I didnt read all the previous information, but on matter of runing the site again with this issue with iOS 8, I want to say: so don't worry about it or anything it will be solved by you (I guess) with time and work , no hurry anyhow
Sorry for the potato grammar :v , Have a good day everyone.

spight said:
Well... I don't know why the Safari issue is happening... I'm still working on it, but I'm not confident that it will be solved in the immediate future. :(

All I can say is that is is a protocol error. And maybe related to the OpenSSL being outdated.

All new Versions of Safari - iOS 8 or Yosemite - are affected.

I know it's a fairly new format for this site, but it appears webm has a slight problem. Not sure if it's a bug, or related to it being new

Hello everyone!

I've been working with Spight and our sysadmin on this safari and iOS8 problem and we might have something, finally.

We are currently looking at Cloudflare having an issue with dealing with the protocols from the new system, and either haven't updated or are unaware.

We have placed a ticket to them to have it resolved, and we are hopeful that we'll have an answer soon.

Thank you so much for your patience!

EDFDarkAngel1 said:
Thank you so much for your patience!

Thank you for your great work and quick assistance!

Bug: post/show API doesn't return arrays

Expected behavior: it returns the fields sources, tags_array and pools as described in the API documentation

Actual behavior: It doesn't contain either of them in xml or json

Steps to duplicate: load up /post/show.xml or /post/show.json with any id

usr said:
Bug: post/show API doesn't return arrays

Expected behavior: it returns the fields sources, tags_array and pools as described in the API documentation

Actual behavior: It doesn't contain either of them in xml or json

Steps to duplicate: load up /post/show.xml or /post/show.json with any id

I had to disable that shortly after implementation because it was using way more resources than I thought it would, I just haven't reverted the API docs yet D:

tony311 said:
I had to disable that shortly after implementation because it was using way more resources than I thought it would, I just haven't reverted the API docs yet D:

And is there a way to reverse-search pools by API, or only by parsing the html?

Bug: Previewing when posting comments with accented (or with other diacritics) or non english letters
Expected behavior: Preview shows up
Actual behavior: Preview does not show up, with URI error as result in console
Steps to duplicate: Write "any" "non-ascii-letter" in comments box and preview it, e.g. éáöïåäöø
Platform: OS X 10.9.4
Browser: Safari 7.0.6

Chessax said:
Bug: Previewing when posting comments with accented (or with other diacritics) or non english letters
Expected behavior: Preview shows up
Actual behavior: Preview does not show up, with URI error as result in console
Steps to duplicate: Write "any" "non-ascii-letter" in comments box and preview it, e.g. éáöïåäöø
Platform: OS X 10.9.4
Browser: Safari 7.0.6

Reported that bug ages ago, mods said it wasn't important :-/

Munkelzahn said:
Reported that bug ages ago, mods said it wasn't important :-/

Ok, tried to search for any report but I guess I failed at that.

It's usually not a big deal, it's mostly when responding to comments with users whom have diacritics in their names, and the preview is just such a handy little thing. But I guess since one can now edit posts it's not really as invaluable.

Since SSL has been a current issue, and since I just got this error about SSL failure three times in a row, I'm gonna copypaste the error message here in the hopes that this is relevant and important for devs to know about:

ETA: I'm using Opera, just in case browser is relevant in some way.

Some of the recent implications aren't being applied correctly.

antelope -mammal
soda -beverage
female_ejaculation -orgasm
loincloth -clothing
hair_over_eye -hair
knee_socks -legwear
translation_request -text
digimon -bandai
swimsuit -clothing
oryx -antelope

Mienshao said:
Some of the recent implications aren't being applied correctly.

antelope -mammal
soda -beverage
female_ejaculation -orgasm
loincloth -clothing
hair_over_eye -hair
knee_socks -legwear
translation_request -text
digimon -bandai
swimsuit -clothing
oryx -antelope]

Ohhh, I ran into this bug earlier. I had no idea it was so wide-spread! Here's what I've discovered about it:

It's not just that it's failing to apply the implications to those images. It won't let you add the implicated tag that's missing either. You can try, but it won't add it all the way, just to the tag history and not to the actual tags on the image itself. After you add the implicated tag, it will list the implicated tag as already being there in the tag history, but on the actual tag list the implicated tag is not actually there. So it's still stuck in limbo with the implicated tag both there and not-there at the same time, it just makes the bug more obvious. So searches say it isn't there, the tags on the image say it isn't there, but it won't let you add it either because it thinks there's nothing to add because the tag is "already there" so it does nothing. Almost like there's two tag lists and the one it references is different than the one that we see. Or something. Or like it got interrupted when the original implication was run and now it doesn't know how to complete the task. Obviously, I don't know actually why it's borked so I shouldn't guess.

Also, it's not just recent implications that have some images stuck in tag purgatory. Earlier today I ran into three of them that had group_sex tag but had failed to add the implicated group tag as well. They were bugged exactly like these are, but I fixed them manually. But still, that implication was run quite awhile ago. So it's not just recently-processed-implications that could be affected here.

I did figure out how to manually fix it, but it's not really possible/practical to do for very many images. Since there's so many, things will have to be fixed in some other more efficient way. But in case it helps to understand the bug, this is what I found did manage to fix it manually: The only way I found was to remove ALL the tags which implicate the missing tag (the site will remove the missing tag along with them). Save it. And then edit it again to re-add back on all of the tags that you just temporarily removed, and this time it will add the implicated tags correctly. You can see me doing it here: https://e621.net/post_tag_history/index?post_id=337794 I tried to add "mammal" twice. Didn't work. (it lists three entries for that and I don't have any idea why it shows three for only two actual edit sessions, maybe that's somehow related to the bug, I don't know.) So anyways, that just proves it's stuck in limbo because the actual tags still didn't show the "mammal" tag yet. So then I removed the "antelope" tag, and it automatically removed the phantom "mammal" tag along with it. And then I added "antelope" back on and that time it correctly applied the "mammal" implication for real this time. That is the only way I've found to manually fix these stuck/phantom implicated tags. And that would work if it was only a handful. But there's too many to do that for all the instances Mienshao found. Especially since I'm sure there's a lot more "stuck" phantom implication examples than just those. It's a big bug.

furrypickle said:
Especially since I'm sure there's a lot more "stuck" phantom implication examples than just those. It's a big bug.

A while ago I sent Char a list of broken implications (which was derived from all the post and implication data available at the time -- about 400k posts IIRC). about 100,000 of all posts are affected by stuck implications, IIRC. Most of these are missing 'mammal' tag.
Char commented at the time that he was aware that implication approval breaks sometimes, particularly for implications affecting a lot of posts.
I haven't heard any update on this issue since then.

A copy of the list I sent him is here , FWIW.

savageorange said:
A while ago I sent Char a list of broken implications (which was derived from all the post and implication data available at the time -- about 400k posts IIRC). about 100,000 of all posts are affected by stuck implications, IIRC. Most of these are missing 'mammal' tag.
Char commented at the time that he was aware that implication approval breaks sometimes, particularly for implications affecting a lot of posts.
I haven't heard any update on this issue since then.

A copy of the list I sent him is here , FWIW.

That's really good to know. Reading both lists (yours and mienshao's) is a little disconcerting to see how many searches and blacklists are undoubtedly affected by this.

Any idea if these missing implicated tags were always acting "stuck" like they are now? Because my theory is that admins have been aware of incomplete/broken implications for ages, but that it's only been recently that this new bug has compounded the issue by making the broken implication resist being added manually when they're discovered and someone tries to fix it.

In fact, I'm pretty sure in the past I'd sometimes get done editing an image and have it spontaneously add a few implicated tags (from existing tags, not ones that I'd added) onto the edit history because the implication hadn't been added to that one yet. So for example, I might edit a post and add "nude" and "solo" and then notice it also tacked on "mammal" for me because that was an implication that had been processed awhile ago but broke partway so it was automatically catching itself and fixing it now in the background. So it used to be self-correcting of the broken implications whenever it got the chance to (the next time it was edited). But now it's not self-correcting and, even worse, is that it won't let you add those tags yourself either.

Actually, I wonder if it's at all linked with the other bug (there's a newish bug where it does the implications and aliases for the tags you added, not when you add them but it waits to do it on the next edit session after yours) and I wonder if that other bug is now blocking these broken implications from being able to correct themselves like it used to? Just a thought. Either way, it's a much bigger issue now that they're "stuck" instead of the comparatively minor previous problem of implications being incomplete but slowly self-correcting and easily fixable when found. Kind of frustrating to watch a minor problem turn into a major one. Those would have been so easy to fix with tag scripting before this second bug made them behave "stuck" like this and functionally uncorrectable. If they fix the "stuck" issue, then I'd gladly go through and fix all the ones on both of these lists with tag scripting. I wouldn't mind doing that, fwiw.

Regarding all implication issues, they should be fixed now. The issue was that I was implicating old tags when calculating diffs and not actually adding the implicated tags (since it looked like it already existed in the old tags. :x) woops.

Fixed that and I'm reloading all posts' tags now.. should be fixed in the next few hours all around

spight said:
Regarding all implication issues, they should be fixed now. The issue was that I was implicating old tags when calculating diffs and not actually adding the implicated tags (since it looked like it already existed in the old tags. :x) woops.

Fixed that and I'm reloading all posts' tags now.. should be fixed in the next few hours all around

I waited to make sure it all had time to finish the processing and then I didn't get back to until this afternoon. But it looks fixed now (the tags are unstuck and it lets me add the missing tags now). So thank you! (And you fixed it so fast too!) *high five*

--------------

@mienshao and @savageorange:
As promised, I'm now going through the broken implications and adding the missing tags with tag scripting. I'm working on the ones Mienshao and savageorange listed, and I'm keeping track of which ones are handled on my profile. So if anyone knows of any others, (or wants to go and dig up any others that might be around but aren't on the list yet) just let me know and I'll add it to the list and make sure they get completed.

ETA: both of those lists are done. If you find any others, let me know and I'll get those done as well.

Tagme is no longer added automatically to posts without any tags. Maybe the rating is counted as a tag now?

And since the tag count searches have been broken for months, it's now rather hard to find users who break the 'minimum of four tags per upload'-rule.

Genjar said:
Tagme is no longer added automatically to posts without any tags. Maybe the rating is counted as a tag now?

And since the tag count searches have been broken for months, it's now rather hard to find users who break the 'minimum of four tags per upload'-rule.

I agree both of those problems need fixing. In the meantime though, I know I have been able to find the images with fewer than the minimum tags by using order:tagcount_asc because for some reason it still works even after all of the other forms of tag count searching broke. It does give timing-out-errors often so that's a little frustrating. But when you catch it in a working mood, it's the only method I know of that still works.

So I've been having an error where the site will be loading fine here on my phone for a time then the pictures will just stop loading. Both thumbnails and full images will fail to load leaving only an empty box in place. If I try to view an image via the download button I get the following. I am running Opera Classic on a T-Mobile My Touch running Android 2.3.6. E921.net loads fine.

Unable to complete secure transaction

You tried to access the address http://static1.e621proxy.ru/data/ce/9d/ce9dd8257c9c8cb457fb0559d3281fe2.png, which is currently unavailable. Please make sure that the web address (URL) is correctly spelled and punctuated, then try reloading the page.

Secure connection: fatal error (1066)

http://static1.e621proxy.ru/data/ce/9d/ce9dd8257c9c8cb457fb0559d3281fe2.png

Unable to verify the website's identity (OCSP error).

The response from the online certificate validation (OCSP) server was too old.
Make sure your internet connection is active and check whether other applications that rely on the same connection are working.
Need help?
Open the Opera Help.
Go to Opera's online support desk.

So I just noticed that right now both fæces AND feces are showing up in the tag list on this image. https://e621.net/post/show/313240

Now fæces is aliased --> to feces, so they shouldn't both be there at the same time. However I suspect it's because of some quirk with the order that implications and aliases are applied in, because this image also has the pooping tag which implicates faeces, so maybe that's adding that form of the tag back in. Actually both fæces and feces have tags implicated off of them, so no matter what order it's done in, it would probably cause some problems.

I really don't know if this is a case where the implication needs to be moved from the aliased-away tag [fæces] and be made to implicate off of the actively used tag instead [feces]? OR if this is actually a bug of some kind that needs addressing? Since I really don't know what kind of solution will be best for it, I'm sticking it here to let both the devs and the admins be able to look at it and decide what needs to be done. I know it's not everyone's favorite tag, but it probably still needs straightened out.

ETA: Noticed it happening on another tag vore/vorarephilia as well, with some images having both tags because the aliased-away tag still has implications off of it that were never transferred and so it adds itself back on.

Bug:Site fails to load under iOS 8.x or Mac OS 10.10
Expected behavior: Site Should load.
Actual behavior: Safari can't open the page "https://e621.net". The Error is: "The operation couldn't be completed. Protocol error." (NSPOSIXErrorDomain:100)
Steps to duplicate: Navigating to https://e621.net duplicates the error. Of note, site loads fine using either Chrome or Firefox. Additionally, the site e926.net loads fine under all browsers. It appears the issue is related to https/socks.

A quick search found this: http://nshipster.com/nserror/
"The SOCKS server rejected access because it does not support connections with the requested SOCKS version.Query kCFSOCKSStatusCodeKey to recover the status code returned by the server."

Just adding to the whole iOS8 issue: the site DOES work in Chrome on iOS8. It does not, however, work in any of the other apps I've tried (Safari, the integrated web browser in the Twitter app, or AnimeBoxes).

Bug
Can't remove empty tag from post #538925

Expected behavior
You should be able to remove the first general tag from post #538925
? X 1

Actual behavior
You can't.

Steps to duplicate
Go to post #538925 and try to remove it.

Firefox says you guys have SSL problems,

e621.net uses an invalid security certificate. The certificate is only valid for the following names: ssl2000.cloudflare.com , *.redpitt.mobi , redpitt.mobi , cloudflare.com , *.cloudflare.com (Error code: ssl_error_bad_cert_domain)

Rating tab under thumbnail displays "↕0" been like this for quite a while now, been ignoring is for like months.

Esme_Belles said:
Rating tab under thumbnail displays "↕0" been like this for quite a while now, been ignoring is for like months.

I have noticed this a few times (though not recently but it's probably still around). The one thing I did notice about it was that seemed to show up on images which were currently at a zero rating. As soon as the image got rated up or down, it would go back to displaying things normally. I figure it's probably just a tiny thing in a line of code somewhere that's making it include the "↕" part instead of just showing "0" by itself. Just in case that helps narrow down when that minor information display error is occurring and why.

furrypickle said:
I have noticed this a few times (though not recently but it's probably still around). The one thing I did notice about it was that seemed to show up on images which were currently at a zero rating. As soon as the image got rated up or down, it would go back to displaying things normally. I figure it's probably just a tiny thing in a line of code somewhere that's making it include the "↕" part instead of just showing "0" by itself. Just in case that helps narrow down when that minor information display error is occurring and why.

yeah its why I haven't reported it till now. Was not a big issue so I figured it would be fixed quickly but.. I think I'm the first to report it since the update that caused it. Strange.

Eleganta said:
Firefox says you guys have SSL problems,

e621.net uses an invalid security certificate. The certificate is only valid for the following names: ssl2000.cloudflare.com , *.redpitt.mobi , redpitt.mobi , cloudflare.com , *.cloudflare.com (Error code: ssl_error_bad_cert_domain)

Not for me (Firefox 32.0.3, Arch Linux x86_64). It gives no error, and when I click on the security symbol (lock) It says 'you are connected to e621.net, which is run by (unknown), Verified by: GlobalSign nv-sa'

It's possible (indeed, it is the most likely case) that your certificate database is out of date.

savageorange said:
Not for me (Firefox 32.0.3, Arch Linux x86_64). It gives no error, and when I click on the security symbol (lock) It says 'you are connected to e621.net, which is run by (unknown), Verified by: GlobalSign nv-sa'

It's possible (indeed, it is the most likely case) that your certificate database is out of date.

It is due to the whole,

One more step
Please complete the security check to access e621.net

system e621 has (detects proxy access). Doesn't work well, sometimes blocks images.
I got this trying to view an image.

Esme_Belles said:
Rating tab under thumbnail displays "↕0" been like this for quite a while now, been ignoring is for like months.

furrypickle said:
I have noticed this a few times (though not recently but it's probably still around). The one thing I did notice about it was that seemed to show up on images which were currently at a zero rating. As soon as the image got rated up or down, it would go back to displaying things normally. I figure it's probably just a tiny thing in a line of code somewhere that's making it include the "↕" part instead of just showing "0" by itself. Just in case that helps narrow down when that minor information display error is occurring and why.

I think I know what's going on here. I don't recall seeing it like that myself, but I did notice that posts with a score of 0 got a double-ended vertical arrow before the score, just like positive and negative scores have arrows pointing up and down, respectively. So I guess ↕ is the html escape sequence for a vertical arrow, but your browser doesn't recognize it.

EsalRider said:
I think I know what's going on here. I don't recall seeing it like that myself, but I did notice that posts with a score of 0 got a double-ended vertical arrow before the score, just like positive and negative scores have arrows pointing up and down, respectively. So I guess ↕ is the html escape sequence for a vertical arrow, but your browser doesn't recognize it.

You might be onto something there. I know I was using a different (crappier) browser for awhile and around the same time was when I noticed that display error more. And that now that I use opera, I don't remember seeing that display error in awhile, and instead see the double-ended arrow you mentioned. So maybe the other browser was even crappier than I'd thought, and meanwhile opera handles it without any problem. The only thing is that it seems like I've seen the "↕" part occur randomly and intermittently over the last two months but... maybe I just wasn't paying attention to which browser I was using when I saw it, or I could be remembering wrong. Your explanation of it certainly makes sense. You know, on a related note, I wish all browsers recognised all characters/symbols automatically, because it would certainly make a lot of things a hell of a lot simpler...

furrypickle said:
I wish all browsers recognised all characters/symbols automatically, because it would certainly make a lot of things a hell of a lot simpler...

Yeah, this happens when something's added to an already established system. It won't work on older browsers. Keeping up-to-date is pretty much the only solution.

Bug
https://e621.net/artist/create?name=blazecake => error

Expected behavior
no error

Actual behavior
error!

Munkelzahn said:
Bug
https://e621.net/artist/create?name=blazecake => error

Works for me. Here's the result: https://e621.net/artist/show/17277

i get this same ocsp .. certificate too old.. error from opera 12.16 on linux for any e621.net url ... whats wrong?

the theme keeps changing to Spring.

Make it stahp pls :(

Ryuzaki_Izawa said:
the theme keeps changing to Spring.

Make it stahp pls :(

This, also shows it was posted 7 hours ago again. =~.~=

Something just started going funky with the timestamps on forum threads. I posted on two threads within the last hour (I wasn't keeping track of exactly when). The first post I made today, was probably 40 minutes ago at the most but now it lists it as happening 7 hours ago. Which makes me a time traveling ninja but is obviously screwy. Other thread posts I read before I posted are now listed as happening after I posted. Ryuzaki's post just above mine in this thread keep shifting from saying it was posted 7 minutes ago or 7 hours ago, varying when I refresh the page. As far as I can tell, three symptoms are happening here: 1, things are suddenly mismarked of when they happened, 2, that they're changing when pages are refreshed, (back and forth between hours ago, to minutes ago, to hours ago again), and 3, that the chronological order has somehow gotten shuffled. Screencap of the shuffled order even according to the timestamps that are displayed: http://i.imgur.com/GFmniqS.png

Just though you guys might like the heads up that it's behaving so weirdly.

ETA: also, I don't know if it's even possible to be related to the above bugs, but there were an awful lot of "heavy load" errors going on at the moment. About one in three pages loaded were giving me a heavy-load error around the same time the forum timestamps went screwy. Probably unrelated, but thought I'd mention on the oddball chance it had some effect on something which is related to the above bugs.

Post timing is off. Posts I posted mere minutes ago were apparently posted several hours ago.

Bug
Prior to today, (~1-2 months ago) a bunch of OSCP errors have been popping up

Something along the lines of:

Secure connection: fatal error (1066) [url]. Unable to verify the website's identity (OCSP error). The response from the online certificate validation (OCSP) server was too old.

Expected Behaviour
Certificate approval and site access I guess

Steps to duplicate
No idea; it seems to have been fixed since I'm able to access the site

https://e621.net/forum/show/139901

Timestamp off!? I just posted it.

BadKitsu said:
Bug:Site fails to load under iOS 8.x or Mac OS 10.10
Expected behavior: Site Should load.
Actual behavior: Safari can't open the page "https://e621.net". The Error is: "The operation couldn't be completed. Protocol error." (NSPOSIXErrorDomain:100)
Steps to duplicate: Navigating to https://e621.net duplicates the error. Of note, site loads fine using either Chrome or Firefox. Additionally, the site e926.net loads fine under all browsers. It appears the issue is related to https/socks.

A quick search found this: http://nshipster.com/nserror/
"The SOCKS server rejected access because it does not support connections with the requested SOCKS version.Query kCFSOCKSStatusCodeKey to recover the status code returned by the server."

Seems to be fixed. Working on both Safari for Mac and iOS.

Screenshots of the timestamp glitch:
http://imgur.com/wR4AJKU
http://imgur.com/gFNgzl7
http://imgur.com/3bZfdVi
In each of these, the posts "from several hours ago" are inaccurate and the others seem to be accurate.

Btw, are posts supposed to instantly bump threads to the top, because neither my first post nor JoeX's post bumped the thread.*

*Actually, I might just be an idiot, nevermind.

hmmm...well i WAS going to make a post about a download issue i've been seeing for a while now but maybe i just needed to let the pages load and sit there for a bit for it to work. seems to happen after loading multiple pages at the same time. it'd either cause windows to act up when trying to drag and drop a pic into a folder or it would start a download with unknown time remaining - 0kb downloaded, either way it'd get stuck. in the case of windows acting up it'd have to reset something due to the error resulting in my having to reopen any and all folders i might've had up at the time.

I hate to nag about this, but the wiki history comparison is still broken:

Sorry if this has been brought up recently, but I didn't see a bug report on it yet. This might just be from Dragonfru.it's end as well, but I figured I'd document it at least.

Bug:
Ads aren't displaying properly.

Expected behavior:
Ads should display some kind of image, an actual advertisement.

Actual behavior:
Ads are sometimes only showing a blank, flat gray image instead of any actual picture. It happens inconsistently (about every 1/10 times I load a page) and waiting for the image to load doesn't seem to help. Tried waiting for a minute or so to no avail. My Adblock is disabled for e6, and I also tried it in safe mode so that other plugins etc aren't interfering. I'm using Chrome browser, if that helps.

Steps to duplicate:
Load a page with ads a few times and it should pop up.

Genjar said:
I hate to nag about this, but the wiki history comparison is still broken:

Full error

NoMethodError in WikiController#diff

undefined method diff' for #<Array:0x19173590> RAILS_ROOT: /home/e621/e621-production/release-2014-06-08 Application Trace | Framework Trace | Full Trace /home/e621/e621-production/release-2014-06-08/app/models/wiki_page.rb:53:in diff'
/home/e621/e621-production/release-2014-06-08/app/controllers/wiki_controller.rb:146:in diff' Request Parameters: {"title"=>"avoid_posting", "from"=>"254", "to"=>"255"} Show session dump[/section] It was handy for various things (especially checking for wiki vandalism), so I hope it gets fixed soon. And I used to check avoid_posting from time to time to see who have been added (or removed) to the list. Can't do that anymore, so it's hard to keep up-to-date on who's dnp. [/quote] Sorry it took so long but it's fixed now. We assumed it was a bug with the diff library itself and didn't dig any deeper, but I took a deeper look today and realized it was something we were doing. Still no idea why it suddenly stopped working (or why it still worked on my dev copy), but at least it works again.

tony311 said:
Sorry it took so long but it's fixed now. We assumed it was a bug with the diff library itself and didn't dig any deeper, but I took a deeper look today and realized it was something we were doing. Still no idea why it suddenly stopped working (or why it still worked on my dev copy), but at least it works again.

Thanks for digging to the bottom of it and fixing it. I used to use that all the time (and the workaround way I found was hard to apply on larger entries like the DNP list) so it'll be nice to be able to use that again. Thanks!

Excellent, thank you!
I can finally check up on a couple of edits that I've wondering about..

tony311 said:
Sorry it took so long but it's fixed now. We assumed it was a bug with the diff library itself and didn't dig any deeper, but I took a deeper look today and realized it was something we were doing. Still no idea why it suddenly stopped working (or why it still worked on my dev copy), but at least it works again.

Excellent! This'll make checking for changes to the DNP a lot easier!

furrypickle said:
Thanks for digging to the bottom of it and fixing it. I used to use that all the time (and the workaround way I found was hard to apply on larger entries like the DNP list) so it'll be nice to be able to use that again. Thanks!

What did you use? I used this diff checker for changes to the DNP. Had to copy the one I have against the newest (I use the same to provide changelogs for the script too)

Xch3l said:
What did you use? I used this diff checker for changes to the DNP. Had to copy the one I have against the newest (I use the same to provide changelogs for the script too)

I figured out the URLs. So here's the normal wiki entry url: https://e621.net/wiki/show?title=avoid_posting and just add "&version=" and then the version number and it will display that version of the page. So it becomes: https://e621.net/wiki/show?title=avoid_posting&version=254 or https://e621.net/wiki/show?title=avoid_posting&version=253 etc. And then, by opening the different versions in different tabs, I could go back and forth and manually compare them. And I'd get the version number by attempting to compare the entries I wanted, because even though it gave me the error, the url still contained the version numbers for those entries. But I only just realised that's the longer way around, when I could have just clicked the date it was edited in the edit history (which I somehow didn't notice was a link) and that way I wouldn't have to type it in. Oh well. lol It's funny sometimes when you just don't notice there's an easier way. Either way, it still got there.

But like I said, it was hard to apply on larger entries like the DNP list, because the comparing part was all manual. Still, it got the job done and useful to know about. Using a diff checker would have been a lot easier, that's for sure. I'm gonna bookmark the one you linked, because that looks handy for a lot of different uses. Nice!

The source field is broken. On nearly all of my recent uploads the field does not pick up anything past either the www. or .com in the link address.

Peekaboo said:
The source field is broken. On nearly all of my recent uploads the field does not pick up anything past either the www. or .com in the link address.

Are you sure? Maybe it's doing that thing it does on comments/forum-posts where it displays it funny surrounding punctuation, often cutting off urls and anything after them. But it's only cosmetic, only displays it funny right after posting it, and only just to the person who posted it and then it clears up and displays normal after a refresh? Because both of your not-yet-approved most recent uploads have a source in the source field that works great, and adding as second source to both of them also behaved normally and saved fine. So they appear to work, or at least they do now. It does sound similar to that cosmetic display glitch that they haven't been able to completely track down, though it's a new thing to have it show up in sources. Or maybe the glitch is in the uploading screen itself? but I'd have to go find something to upload fast to check that...

furrypickle said:
stuff

I've had to edit and paste in the source again for it to work, I'll try and see if the refresh thingy works on my next uploads.

EsalRider said:
I think I know what's going on here. I don't recall seeing it like that myself, but I did notice that posts with a score of 0 got a double-ended vertical arrow before the score, just like positive and negative scores have arrows pointing up and down, respectively. So I guess &amp;varr; is the html escape sequence for a vertical arrow, but your browser doesn't recognize it.

Aaah! That explains why it only shows up on my kindle fire v1. Thanks.

treos said:
hmmm...well i WAS going to make a post about a download issue i've been seeing for a while now but maybe i just needed to let the pages load and sit there for a bit for it to work. seems to happen after loading multiple pages at the same time. it'd either cause windows to act up when trying to drag and drop a pic into a folder or it would start a download with unknown time remaining - 0kb downloaded, either way it'd get stuck. in the case of windows acting up it'd have to reset something due to the error resulting in my having to reopen any and all folders i might've had up at the time.

This so much. It's at the point now that im seriously considering dropping E621 from my "daily websites to check", because it is extremely irritating.