Distrowatch

[DavidMcCann]

Since updating to version 31.1.1, distrowatch.com can not longer be accessed. I get

400 Bad request
Your browser sent a request that this server could not understand.

[gepus]

User Agent sniffing!
With User Agent Mode set to Native the site can be accessed.

[Night Wing]

Distrowatch changed their site sometime in the last two days since I also get the:

"400 Bad request
Your browser sent a request that this server could not understand" prompt".

Since I use Firefox compatibility mode as default in Pale Moon, if I have to change to Native mode for Distrowatch, then when I visit other sites which have problems with Native mode, then I have to switch back to Firefox Compatibility mode. So I've learned to quit fighting the problem.

In other words, Waterfox and Firefox are my two backup browsers to Pale Moon when a site (like Distrowatch) gives me a problem. When that happens; I just use Waterfox to access those sites and in this case, Distrowatch.

[gepus]

In other words, Waterfox and Firefox are my two backup browsers to Pale Moon when a site (like Distrowatch) gives me a problem.

All you have to make is a Site Specific User Agent Override for Distrowatch.
If for every User Agent sniffing site you use other browsers then you can as well quit using Pale Moon.

[mr tribute]

In other words, Waterfox and Firefox are my two backup browsers to Pale Moon when a site (like Distrowatch) gives me a problem.

All you have to make is a Site Specific User Agent Override for Distrowatch.
If for every User Agent sniffing site you use other browsers then you can as well quit using Pale Moon.

The user agent for Distrowatch will probably be updated for the next release of Pale Moon.
I stay with defaults because I don't want to keep track of my own changes.

[andyprough]

This is why Pale Moon has so many site-specific user agent switchers, that make it super easy to change the UA for a single problematic site. Sasuga, PermissionsPlus, and User Agent Status are the ones I've tried. My current favorite is Sasuga.

[Moonchild]

An error 400 based on user-agent seems pretty severe and is likely deliberate behaviour of the site. There's obviously nothing wrong with the request the browser makes or it would never work no matter the UA string, so the error 400 is thrown on purpose as a "go away" signal, most likely.
An SSUAO is a workaround but this feels to me like distrowatch needs to ease up on the UA filtering of clients. Have you contacted them yet?

[andyprough]

An error 400 based on user-agent seems pretty severe and is likely deliberate behaviour of the site. There's obviously nothing wrong with the request the browser makes or it would never work no matter the UA string, so the error 400 is thrown on purpose as a "go away" signal, most likely.
An SSUAO is a workaround but this feels to me like distrowatch needs to ease up on the UA filtering of clients. Have you contacted them yet?

I will, I know them relatively well.

[andyprough]

An error 400 based on user-agent seems pretty severe and is likely deliberate behaviour of the site. There's obviously nothing wrong with the request the browser makes or it would never work no matter the UA string, so the error 400 is thrown on purpose as a "go away" signal, most likely.
An SSUAO is a workaround but this feels to me like distrowatch needs to ease up on the UA filtering of clients. Have you contacted them yet?

I received a good response, an apparent fix, and a recommendation for our devs for the Firefox compatibility string:

Hi Andy,

Thanks for the detailed report.

I looked at our settings and I think what was causing the problem is we'd started filtering software which was using old (and possibly inaccurate) user agent strings. These are typically used by spam bots and we end up filtering them out to prevent attacks and spam on our site. The old PaleMoon user agent, which refers to itself as Firefox 68, is one of those strings that bots use but Firefox hasn't used for a long time.

I've removed this filter and it should, I think, allow Pale Moon to work with its normal agent string. However, I'd suggest to the developers they should probably update it to match a more modern browser as a lot of sites stop working (or will nag users to upgrade) when they see older UA.

Please let me know if the issue continues.

Hopefully that fixes it, and if not they seem happy to work with us.

[Night Wing]

All you have to make is a Site Specific User Agent Override for Distrowatch.
If for every User Agent sniffing site you use other browsers then you can as well quit using Pale Moon.

I'm with Mr Tribute on this one. I don't like to constantly change things in "about:config" user agent wise because I can make a change user agent wise to make the site work again for Pale Moon. But this is not the point. The point is below.

And the point is, if the fix is in the next version of Pale Moon for a site that makes the site work again, then I have to go in and delete the change in "about:config". This is okay if you have one hard drive. But I've got 5 computers (2 desktop towers, 3 laptops).

Now comes the work part for me for a problem site.

Each desktop tower computer has three hard drives in each tower for a total of six hard drives. Then each laptop has a hard drive in it so three laptops equals three more hard drives for a total of nine hard drives to make a user agent change for Pale Moon for a problem site.

Then we have the the PITA sites like "Citibank" which every few weeks makes a change and Pale Moon won't work in it (again). So changing specific user agents for Citibank is like playing "whack-a-mole" which is counter productive in my opinion. I have a Citi credit card, but I don't have an online account with Citibank. I handle all of my Citibank credit card account transactions with a land line phone. No browsers.

I could use an extension like the ones mentioned in other posts. But then I would have to choose an extension to make a specific user agent for Pale Moon for a problem site. Then I would have to hope the extension I choose keeps on working when Pale Moon goes to another major version. Like what happened with Pale Moon when some extensions which worked in Pale Moon 29, did not work in Pale Moon 31.

All of my Pale Moon browsers run one extension in them and that extension is uBlock Origin. I do not use any other extensions in Pale Moon. I do use two lightweight themes which are Micro Moon and Moonscape. This is all I use in Pale Moon. For me, when it comes to extensions, I like the "less is more approach".

So with my nine hard drives running Pale Moon, it is a lot easier for me to use a different browser for the problem site which used to work in Pale Moon and now does not than fiddling around with changing user agents for Pale Moon with sites that cause problems in conjunction with those nine hard drives. Ditto for extensions.

In other words, I don't like to make unnecessary work for myself if I can help it.

As for me quitting the use of Pale Moon, that is not going to happen. I started using windows Pale Moon back in April of 2011 with Windows 7 and then started using linux Pale Moon in January of 2014 when I started to experiment with linux distros (Debian, Zorin, Linux Lite, Chalet OS, SolydX and Mint).

I like the way Pale Moon looks and it is very easy for me to customize using the Customize Toolbar right out of the box since coding is not my forte'. If the Distrowatch site starts working again with Pale Moon, fine. But if Distrowatch changes something again where Pale Moon does not work, like with Citibank, then it's Waterfox for me for Distrowatch.

[andyprough]

I can confirm that Distrowatch is now working with Pale Moon's default "Firefox compatibility" UA on my system. It appears that the fix that the Distrowatch dev implemented for us has worked.

[Moonchild]

a recommendation for our devs for the Firefox compatibility string

The issue is that other sites will use user-agent sniffing as a "feature detection" method[1] which is in use a lot more often than a filter. The problem with that, in turn, is that if a site uses feature detection this way and we feed them a Firefox version that is current/current-1, they will assume we have all the shinies Firefox does, which we don't -- meaning it will break the sites. We have been forced to use a lowest common denominator for a long while in that respect, which is what landed us on 68.0 (latest ESR version prior to Google Webcomponents were implemented in Firefox)

Maybe it's time to update this, but I'm reluctant to do so without people testing this fairly broadly first.
You can blanket change this compatibility version easily though if you want to help me test this: general.useragent.compatMode.version -- set this to 102.0 if you want to change the Firefox version it "pretends to be" in the user agent string to 102

[1] This is a really, really bad idea, but it's been the practice for decades and even the biggest companies that are supposed to have very capable web devs still keep using this, despite its many pitfalls! Feature detection can be done without sniffing the UA. Everyone has ignored that memo, though, despite there being APIs in browsers specifically for it.

[andyprough]

Maybe it's time to update this, but I'm reluctant to do so without people testing this fairly broadly first.
You can blanket change this compatibility version easily though if you want to help me test this: general.useragent.compatMode.version -- set this to 102.0 if you want to change the Firefox version it "pretends to be" in the user agent string to 102

I've been trying this out since last night, don't notice any difference at all so far. I don't go to youtube or facebook or those kinds of sites, so it will probably need more testing by other users. My most complex sites are video sites like odysee.com and piped.kavin.rocks, and I haven't noticed any trouble there.

[Moonchild]

Yeah it won't really help if you aren't visiting a lot of mainstream sites. But first impressions are good.

[andyprough]

Yeah it won't really help if you aren't visiting a lot of mainstream sites. But first impressions are good.

I tested it on Youtube to listen to a couple of versions of Fleetwood Mac's 'The Chain' on @mr tribute's latest 'Your Theme Song Today' recommendation. Worked perfectly as far as I can tell.

Also, @Michaell was having trouble getting to the log-in screen at gab.com, but I was able to with the 102.0 user agent setting.

[Moonchild]

I tested it on Youtube

Youtube uses a site-specific override so won't be affected by the version change.

[Falna]

Maybe it's time to update this, but I'm reluctant to do so without people testing this fairly broadly first.
You can blanket change this compatibility version easily though if you want to help me test this: general.useragent.compatMode.version -- set this to 102.0 if you want to change the Firefox version it "pretends to be" in the user agent string to 102

Well it's only been 12 days, but I switched over to 102.0 the day you posted this (with web components enabled too) and all the sites I've visited - lots of them, including eBay, Amazon, but no social media - seem to work as expected so far...

[andyprough]

Well it's only been 12 days, but I switched over to 102.0 the day you posted this (with web components enabled too) and all the sites I've visited - lots of them, including eBay, Amazon, but no social media - seem to work as expected so far...

Same here, I haven't had to downgrade from 102.0 for any sites. If anything, I've actually had to do far less fiddling with site specific user agents to get things to work. My poor Sasuga extension is getting lonely.

[back2themoon]

The problem with that, in turn, is that if a site uses feature detection this way and we feed them a Firefox version that is current/current-1, they will assume we have all the shinies Firefox does, which we don't -- meaning it will break the sites.

No issues with 102.0 so far. I guess it comes down to knowing what those shinies might be, and which websites actually use them so they can be tested.

[Moonchild]

Well that's good to know! I guess the frameworks moved on to just not bothering to check and just using stuff so it'll break regardless of UA, then. Or maybe we're caught up enough to satisfy most of them and can avoid bad sniffing that way then.
I'll consider it for 31.3.0