Hacker News new | comments | ask | show | jobs | submit login
Ask HN: Why can't I login to HN via Firefox?
150 points by SeanDav on May 3, 2013 | hide | past | web | favorite | 101 comments
I can login via Chrome or IE, just not Firefox. This behaviour started very recently - a matter of a few hours ago.

Also, for future reference, what is the best way to ask about this sort of problem?

Just had to re-login to HN for the first time in weeks on Chromium. Android required a refresh (something about a user id conflict). Something must have changed on the backend.

Regarding the browser logout I had the same happen here, I was actually concerned I had been shadow-banned or something when I realised I had been automatically logged out.

Same here. Brief moment of panic when I couldn't get logged in and immediately started mentally going through my recent comments say "OMG, I didn't say anything that controversial did I?!!??" Luckily, that does not appear to be the case.

I had similar thinking. I was excited, almost. If I couldn't use HN anymore, I would have no choice but to move on!

Alas, it looks like I am trapped again.

add Safari on OS X to that list. Made me do a bit of a double take since the login was taken until I clicked the comments link at the top of the page.

Me too (Win7 x64 Chrome), but I could login without problems

Since upvotes are invisible, I guess I'll have to make a "me too" post. So yep, I also had to login for the first time in as long as I remember on Win7 x64, Chrome 24.

ditto on Safari OS X

Oh that's neat! I usually have to re-login every day on Chromium, HN never remembers my session past a few hours. I think it's got something to do with multiple devices - HN only allows one or two logged in sessions I think, and tablet/phone/laptop/desktop over the course of a day messes that up.

I think that the multi-device-thing is not a problem, I've been logged in for weeks (or months) from my laptop (Chrome), work PC (IE 6-8) and 2 different phones (android) without problems

If you explicitly logout from one browser, you are logged out from all. If you just close the window and discard the cookies, you remain logged in elsewhere.

Yep, both my home and work Chrome browsers required me to login this morning for the first time in about a year.

We've had a bunch of reports of problems logging in. We're investigating.

I changed header parsing last night to make header name processing case insensitive.

Due to an oversight, the first cookie sent was being dropped. As long as your session cookie wasn't first in the cookie header, everything worked fine, which is why I didn't notice (despite testing in Safari, FF, and Chrome). This is also why clearing cookies sometimes worked - it shuffled the order in some browsers.

Sorry everyone.

No problem, thanks for posting the details instead of just "try it now." It's just a matter of time before I would have done something similar. I appreciate the opportunity to learn from someone ELSE's mistakes for a change---one of the benefits of HN.

You fixed header processing? That's great!

I wrote an HTTP proxy server. Every site on the internet I ever tried worked through it except HN, which went into a redirect loop. I investigated (insofar as HN's hair-trigger banning allows such things) and think this was due to the fact that the proxy was normalising header case. If this is fixed, I'll be very happy.

You should test with elinks: it failed consistently ;)

It was as if the homepage was cached and wouldn't reflect login state.

Accessing a story's comments showed I was logged in, and only then could I return to the home page and it would also show that I was logged in.

I can't speak officially for anyone at YC, but I happen to know that they have been experimenting with nginx 1.4 with SPDY support and there was an incompatibility with the Arc web server and specifically with cookie handling.

An intelligent guess says this has something to do with what you are seeing.

I can confirm that there was a technical change on HN today that affected other browsers. I have been just turning on my browser (Chrome running on Windoze) and coming over to HN already logged in for months now. Just today I had to log in again, which because HN's new emailed link for password updates allowed me to specify a password, has let me choose a memorable password specific to HN. (The former way HN updated passwords by email was to send a new, random password to my email address, as for example when I logged in via a different computer without an HN cookie set.) So far, so good.

Does not look like it's enabled right now - http://spdycheck.org/#news.ycombinator.com

Check again now ;)

I deleted the HN cookies, then it allowed me in again.

Even after deleting all HN related cookies, it still doesn't work in Firefox (Iceweasel 17.0.1).

EDIT: it worked after I re-deleted all the cookies and re-force-reloaded the page.

I had to refresh the page after cleaning the cookies and before the login, too.

Ditto, on FF20, except I had to delete them twice.

This worked for me as well. I haven't had any trouble since (not even when using Firefox on a different machine).

Cleared cookies. Still can't login. FF 20.0.1 on Windows 7.

Works here too, thanks for the tip!

Clearing browser cookies in IE9 does not work for me. After login it just puts me back to home page and I still see the login link on the right.

Worked like a charm, thanks for the tip!

Yeah, WTF. I noticed it this morning on my work computer (Firefox Aurora on Windows 7). Logging in here from Chrome to upvote.

Logging in seems to work; I don't get an error. But the session doesn't stick, after the redirect I don't get my custom navigation bar color.

P.S. Removing news.ycombinator.com cookies didn't help.

> P.S. Removing news.ycombinator.com cookies didn't help.

Same here. FF 20 on Linux.

It didn't work for me at first, so thinking that I had somehow failed to clear the cookies I tried it again and was able to log in.

Didn't help here either.

Try using "Private Browsing" - it sounds like you have the same problem I have, reported here:


No one seems to be taking it seriously.

I'm having the same issue, and also wondering what the official support channel for HN is.

  Please don't post on HN to ask or tell us something (e.g. to
  ask us questions about Y Combinator, or to ask or complain
  about moderation). If you want to say something to us, please
  send it to info@ycombinator.com. 
-- http://ycombinator.com/newsguidelines.html

...and yet you posted this: https://news.ycombinator.com/item?id=5649704

Indeed - at the time no one else had reported problems, and it was a taster/teaser to see if it was just me, or a system-wide problem. I wasn't asking for support, or reporting a problem to HN, just seeing if it was a local SNAFU.

"at the time no one else had reported problems"

How do you know others didn't do the right thing (by emailing info@ycombinator.com) ? AFAICT emails sent to the address aren't publicly available.

You're right, I didn't know that he hadn't already had a flood of hundreds of emails. However, it was about 4 in the morning in California, so I figured even if he had hundreds of emails, I wouldn't find out quickly. More, I figured that most of the other times something goes wrong at least someone posts to HN asking, and no one had done so.

I was hoping to find out if the problem was unique to me, and I suspected it wasn't. I was half hoping to head off the usual flood of posts that something was wrong and let there be just the one place, and I was certainly not expecting this reaction. I am bewildered at the hostility, and going away to try to understand it. I feel a bit like I've been kicked.

"I am bewildered at the hostility, and going away to try to understand it. I feel a bit like I've been kicked."

IF there's any hostility, it's directed at people and acts that are seen as inconsistent: You point to the rule that says "Please don't post on HN to ask or tell us something ... If you want to say something to us, please send it to info@ycombinator.com. " after you made a post telling them that the site is having issues.

"More, I figured that most of the other times something goes wrong at least someone posts to HN asking, and no one had done so."

That's not an excuse for you to do something which, as you pointed out, violates a rule.

"I was half hoping to head off the usual flood of posts that something was wrong and let there be just the one place,"

Incidentally there is one place that did head off the usual flood of posts: https://news.ycombinator.com/item?id=5650171 Are you complaining that it wasn't your post that was the one place?

Look. It was four in the morning in California, so I posted a question here in essence to see if anyone else was also having problems. Doing so isn't against any rule, it's against a guideline. I assessed the situation, and decided that at the time a posting was more appropriate than an email.

Subsequently someone asked if there was a support line, and I answered the question. I don't see that as inconsistent, I think it was the most prudent thing to do.

    Incidentally there is one place that did head off the usual
    flood of posts: https://news.ycombinator.com/item?id=5650171
    Are you complaining that it wasn't your post that was the one
What? I posted three hours before that. No one noticed, probably because it was 4am in California. A later posting got the attention. So what? I have no idea how you can see that as me complaining - I even posted cross-links so people could go to the one that had got the attention!

This avails us of nothing. You seem to have issues with actions that I see as having been perfectly logical and reasonable. You don't accept my version of what happened, and genuinely can't see why you think they were inconsistent.

Screw it. It's the weekend, there's a holiday Monday, and I'm walking away.

In defence of ColinWright, it was partly due to his quick comment, upvote and crosslink that got this story onto the front page. Otherwise it might have floundered and never reached the front page, which would possibly have meant a whole flood of posts asking the same question, until one gained traction.

Having said that, I too found the logic of pointing out an email address for support, while posting a submission, somewhat contradictory - but certainly nothing to get aggressive about.

I don't think dfc's first reply or niggler's first reply contain any aggression or hostility whatsoever. As you said, pasting the guideline having done the opposite hours earlier was contradictory, so that was pointed out.

I should say it doesn't help that the first part of his explanation was that if he had followed the guideline -- using email instead of posting -- he wouldn't have gotten an answer as quickly as he would like. This immediately implies that exceptions are to be made as soon as the recommended action infringes on personal convenience, which would of course make the guideline meaningless.

To my reading the guidelines didn't apply given what I knew at the time. As far as I knew the problem might have been purely local, in which case I didn't want to hassle YC/PG with what would amount to a support issue. I thought it prudent to try to discover if it was wider than just my situation. So I posted.

If you honestly can't see the difference between posting to tell YC/PG about a problem, or to ask them a question, as opposed to posting to ask if other people are having the same problem, then I'm at a loss.

I can understand that distinction, although I'd argue that pasting the guideline without comment to another user in the context of this particular issue implied that your stance was that this issue was one where the guideline applied. You might argue that his or her question was general but I think it's clear from the context that he or she was asking: "what channel should I use when something _like this_ happens?"

Anyway, not a big deal and no hostility intended - have a good weekend.

"I was half hoping to head off the usual flood of posts that something was wrong" by being the first post in the flood.

What is the difference between reporting a problem and seeing if it was a local SNAFU?

For example, checking downforeveryoneorjustme.com or asking if a co-worker can reach a web site is different from contacting the site's operators to tell them about the trouble you're having.

How is this related to anything we are talking about?

It's related because it's another example of discovering a local symptom and attempting some investigation before emailing YC/PG.

I attempted an analogous investigation before posting. I tried Private Browsing mode, which worked. I tried different browsers, which didn't. The next step in my investigation before emailing YC/PG was to find out if it was just me on this local network, and the only way I had to investigate that was to ask someone else. The only way I could do that was to post to HN. It seemed the obvious next investigatory stage.

Colin, how is submitting a story in order to figure out if it was a local SNAFU different from submitting a story in order to report a problem? Especially in light of the FAQ suggesting/recommending/asking for people to email instead of submiting stories about problems/questions...

Some people are incapable of introspection. Sadly, including some of the more active people on HN.

I had thought you had asked about the difference between seeing whether a problem was just some odd, isolated thing and informing the people running a system that you were having a problem. If that's not it, I'm unclear as to what you were actually asking.

Colin said:

"I wasn't asking for support, or reporting a problem to HN, just seeing if it was a local SNAFU."

I asked how submitting a story to report a problem was different than submitting a story to see if it was a local SNAFU. And you brought up isup.me which would have been useless since the issue was related to logins?

As we have seen time and time again, internal consistency is for the plebeians

This is my favorite of Colin's:

  "Isn't HN amazing? Here is the same article submitted on no less than four previous occasions:

  * http://news.ycombinator.com/item?id=4185226
  * http://news.ycombinator.com/item?id=4170972
  * http://news.ycombinator.com/item?id=4355548
  * http://news.ycombinator.com/item?id=4167834

  Total upvotes: 29.

  Total discussion: 0.

  No wonder people try to game the system and find optimal times to submit things.

  No wonder people think HN is broken."*[1]
What he failed to mention is that his submission was the fourth and most recent duplicate (oldest to newest):

4167834: zmanji

4170972: VeXocide

4185226: moobirubi

4355548: ColinWright

And obviously he had a justification for his duplicated submission....

[1] https://news.ycombinator.com/item?id=4638756

Yes, I did have a justification:


But it now seems clear that this has a personal component:

    I do have a problem with Colin's crusade
    against duplicates.

    -- https://news.ycombinator.com/item?id=4638986
Should I stop just because you and a few others say so? many others have repeatedly expressed support.

But now that I have a context for your criticisms I know how much weight to give them.

Your justification was another, when the rules apply to me there is a built in special case exception.

There is nothing personal about it, I have no idea who you are, I am sure that you are a decent human being. I have a problem with hypocrisy. In the case of the duplicates you said "no wonder people think HN is broken" because someone submitted a duplicate story...and you did not mention that the fourth submission was yours. It seemed to me that your behavior was just as worthy of a "no wonder people think HN is broken" as the person who submitted the story for the fifth time.

"I have a problem with hypocrisy."

I'm glad to see that I'm not the only one :)

Someone clearly cares deeply about HN karma. Not sure why he cares so much.

It's not karma, it's the functioning (and occasional lack thereof) of the "community." Since I run more than one early-stage groups, I'm continuing to learn all I can about what works, what doesn't, what seems broken, and what can be done to make them better.

Don't assume you understand my motivations.

A foolish consistency is the hobgoblin of little minds.

Everything in moderation, including moderation.

PG has repeatedly said that HN, despite its size, is still an experiment, so there is no support channel, I'm sure when he has time he will look into it.

It didn't help in my case. Nor did the cache and cookie purging.

The cookie variant didn't work at first for me either. I had to delete the cookies, refresh the page, and then try to login. Then it worked for me.

This worked for me after clearing cookies alone didn't work, even several times. Thanks!

Finally worked for me when I purged the cookies and restarted with no cache.

For me too.

Me too. Firefox 20.0.1 on Win XP, at the moment.

By comparison, login works fine in Chrome 26.0.1410.64 .

P.S. Deleting cookies in Firefox did not help.

P.P.S. Firefox Private Browsing window also does not work for me. Chrome continues to work fine (multiple logins, each time using an Incognito window).

Nuking my HN cookies fixed it. Using FF Aurora.

This worked on my home desktop but not my workstation. I get a new set of cookies when I try to log in, but I'm still not logged in. (Tried repeatedly.)

Edit: clearing cookies, then refreshing the page before trying to log in again, did work. (Thanks phaylon!)

Yep. It worked for me as well. Thanks! =)

Cleared cookies on FF in linux - Worked

Cleared cookies on FF in windows - Did not work

Versions of Firefox were the same, Linux version has various dev related extensions installed but I cannot figure out why this would change things.

[EDIT] Site just died for a few minutes with NGINX error pages. Probably HN fixing it.

No Opera fans in this thread yet so I'll contribute a report. Opera wasn't logged in to HN when I started the browser this morning, but simply logging in again worked, no cookie clear needed.

Not working on Firefox. Tested and working on latest versions of Chrome, Opera and IE8 for Windows XP.

Edit: Removing HN cookies (3 of them) made Firefox work again for me.

If you erase your cookies for news.ycombinator.com, does it work then? See the menu: Tools; Preferences; Privacy; Remove individual cookies (or something like that, I translated it).

I used a few different browsers and compared the raw responses.

I noticed when I used Firefox (20.0.1 linux), there was an additional redirect to heapanalytics.com.

Using Firefox(20.0.1 linux ), I obtained a valid session. It was simply the login status that was displayed incorrectly.

I deleted the cookie from heapanalytics.com, refreshed my Hacker News page and the correct login state message displayed (user name followed by logout link).

Hope that helps.

Ack. Just go back in via IE, thanks for the hint.

However I could log back in on my other machine with Firefox after, but only after getting a different external IP from the ISP.

My bet is that they are experimenting with some sort of anti-spam/sock-puppet protection thing and it's misfiring.

I actually could never log in on Chrome and Chromium and today, suddenly, I was already logged in! It used to work fine on Firefox (don't have it on office machine, so can't check right now).

I too have not been able to use my original id, I made this one and I can logout / login with it but not my original, shivetya

firefox / safari in mountain lion, windows 7 at work was firefox

Still annoyed HN doesn't allow logging in via Google any more, myself. The login functionality just seems to get worse and worse every time they touch it.

This is seriously oddpants, but I could not log into HN until clicking on this link, at which point I was automatically logged in.

Took a few times of clearing my cookies, then it eventually just worked, no idea why though and I guess it'll stop working again.

Chrome works fine for me.

But I can't login via IE9 (Vista SP2) either.

Though from the last browser stats share on here, my issue only potentially affects 3 users.

Whenever I type in the correct password, I'm redirected but I'm still not logged in.

Debian Linux 7.0 Iceweasel 10.0.12

I'm using Konqueror as a workaround.

No Problems with FF 20 on Win7 and OS X

Really glad this wasn't just me.

I too had to re-login for the first time in, well ever, today. I'm running Chrome on OSX.

Working for me on FF 20.0.1 / Win 7, after the suggestions to delete HN cookies.

Chrome on Windows 7 with ABP here - had to disable ABP to be able to log in.

No problem with Firefox 20.0

I was having trouble on Safari/Mac OS X but it's all better now. Thanks!

I had to clear my cookies a couple days ago on Chrome to log back in.

Seems to be working for me in: FF 19

(Really need to rebuild that ...)

FF20 on Mac works fine for me.

nightly 23 -> no problem

Glitches in Safari and IE10 as well.

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact