Errors/Warnings/Bugs

[Silverhawke]

and if a new problem is published, the progress would be adjusted as well, right?

[euler]

Yes, that is correct. It is a dynamic measure of the most recent problems. But the good news is that, along with only a few others, they are permanent awards. That is, once you've earned it, you keep it.

[user202729]

This is the first I am hearing about this. How recently have you seen this behaviour? Is it the whole website that seems to struggle or just viewing the latest problem?

* ... I can't recall? I only start competing in the "fastest solvers" table recently (although it has been a long time...), and if the server is "fast like you expect", I wouldn't remember that much...

That having said, it has been happening for several recent problems (which mean a few weeks/months).

* The same problem occur when I try to access the server with Tor. (just a datapoint)
* I think it's only the latest problem, although I usually don't try to visit other pages during that minute.
* Latest minimal page has the same issue.
* Sometimes I get "problem not found" even if the page load start time is strictly after the publication time (and the delay is high).

[user202729]

About the new tags feature: nice work! It will (hopefully) reduce the number of users who want to cheat.

(my idea was, that if the background knowledge listed in the tags is all the background knowledge required to solve the problem, then the user have gained nothing by viewing the actual solution, because that part of the solution for that specific problem is not applicable to any other problems)

(I intended to do something similar -- listing the list of common-knowledge required to solve all PE problems, but it's just too much work, and I definitely didn't solve the whole problem set)

But there's a typo on the page https://projecteuler.net/about=tags . "sovler" ?

[philiplu]

I just noticed that the information over fastest solvers for a particular problem is displayed incorrectly whenever someone who has solved it in more than one week but less than two weeks enters the scoreboard, i.e. this occurs whenever less than 100 people have solved the problem within a week. What happens is that the counter for number of days is not reset when entering the second week, so an entry corresponding to day eight after the problem was released will display as 1 week, 7 days, x hours, where it should be displayed as 1 week, x hours, etc.

This counter does not reset until the end of the second week, see e.g. https://projecteuler.net/fastest=747 for an example.

Hi, @euler. I noticed a variation of this timestamping bug. I was looking through the Problem Archives list, and noticed a problem that said the last post was 11 hours ago, but the link and time was not in red as is normal for posts less than a day old. Instead, the poster name and time were in the normal black for older posts. This was on problem #355, last post by @SNAIL-ERATO, reported as 11 hours ago as I type this.

When I actually went to that post, I see that it was actually posted 1 week, 11 hours ago. I assume that's the same bug as @jonado1 noticed, just in a different place.

[euler]

Thanks for reporting that. I think I've found the problem and fixed it. Please let me know if you spot it happening elsewhere.

[philiplu]

There's a minor bug in the Archives pages. Looks like it's possible to create a user name that's not displayable, and which breaks some of the page functionality. Take a look at this screenshot



Problem 726 doesn't show the number of posts, nor the time since that last post, and the user-name isn't a link to the last post in the forum. When I look at the forum, the last post is from 14 Dec 2020, and has no user name attached. I'm pretty sure I just noticed this a few weeks ago, though, and I don't think it's been this way since December.

[euler]

As always, I am very grateful to you for reporting issues. The problem was caused because the member who made the last post had deleted his/her account, but chose to leave the posts. It should be fixed now.

[jaap]

The security certificate of projecteuler.chat seems to have expired today.

[euler]

Yes, it was a bit of a headache. The new certificate was renewed one month before it expired, but it seems that it has become a two-step process now with our hosts thanks to the the joys of GDPR. It should all be working now and I'm better aware of the new process that means it shouldn't happen again in future. Until they change the process again!

[DeKlod]

Hello,
I wonder whether there is a bug on the stats page - when I hover with the mouse over some user names, it says that they have solved a problem 'within the past day'. However, the number of problems solved has not changed in ages.
Nothing major, just thought I'd mention it
Cheers,
Claude

[euler]

Hi, Claude. It used to say how recently each member had solved a problem, but unless you found somewhere I missed, we changed it some time ago so that it now displays "____ was active ____ ago" or "____ was active within the past day". Active simply means they have visited the site. If you found somewhere it says "____ solved a problem within ____", then could you let me know where you saw it?

[DeKlod]

Oh, my mistake! I didn't notice the changed working and only looked at the date, i.e. 'days since last'.
Apologies for the confusion

[euler]

Please don't apologise. I'd never spot bugs otherwise, so I really appreciate members taking their time to offer feedback.

[GNFS]

Does the show all problems page (https://projecteuler.net/show=all) functionality still exist? This was changed sometime between last and this summer.

[euler]

No, the functionality was removed some time ago. The number of problems in the DB had grown so large that generating so much data (problem text and images) had become unreasonable.

Instead we introduced a much better and more manageable feature that allows you to "pin" up to twenty-five problems that you are working on and you can filter based on those.

[CBurkhart]

I was trying to update my "Extra Info" profile text and was getting some surprise truncation on text I thought was below the 200 character limit.

Some rearranging ended up with a single character comma being left as the partial truncated string ,. Thus is would seem that a comma takes up 7 of the 200 characters instead of 1 of the 200. Almost certainly the same for other punctuation. Is this the expected behavior?

[euler]

Thanks for bringing this to my attention and your analysis was spot on. I was using a fairly strong encoding function primarily for security purposes, which is way stronger than needed for that. So I am wondering if I'd done that at some point in the past to ensure that some special characters from our wide base of members were displaying properly. I've changed it to a less zealous function, which still protects against exploits, but I'll have to wait to see if it causes problems elsewhere. Most standard punctuation entities should no longer encode, which means you should be able to enter a longer string now. There will be unavoidable instances where the number of characters entered are less than 200, but the encoded version is longer, but they should be incredibly rare now.

[GNFS]

https://projecteuler.net/problem=124 (and maybe others) still has a leftover spacer.gif which 404s.

[euler]

Thank you for pointing that out. It should now be fixed. I did a DB search and there don't seem to be any more instances of that relic.