Errors/Warnings/Bugs

[TitanCode]

Try this tread:
viewtopic.php?p=73689

[AndrewMcCall]

[Bug on Forum Registration Page]

The link on the register page (projecteuler.chat/ucp.php?mode=register) back to the main site has a typo. The link goes to projecteuleT.net which of course, doesn't exist. This is a lovely way to start a phishing attack though.
Just log out, go to register and accept the conditions to see the link.

ss-20251012-162557.png

Thanks for reading,
I'm still new so apologies if this is not the correct place for this sort of report.

[euler]

Thank you for reporting this. I have no idea how long it's been like that. Good spot!

It should now be fixed.

[philiplu]

I made a short post to the forum for PE965, at https://projecteuler.net/thread=965#448070. When I open the forum thread, or refresh the page, the code block in the post is defaulting to opened instead of closed. Also, minor nit, but the code block is now including an extra blank line at the start.

Edit: Seems to be on purpose; very short code blocks apparently now default to opened.

[euler]

It was an intentional change some time ago. It seemed pointless to expand a code box to find short blocks of code, so code that does not exceed 10 lines or 80 characters in width will show open by default. It used to be 5 lines, but I increased slightly a couple of weeks ago, so maybe that's why you noticed it now. It's also likely that other recent problems required much longer blocks of code to solve.

I've also fixed the empty line at the start of the code block.

[i707107]

The times that problems were solved in History section is not coherent with my time zone. Also, the problem title fonts became too small.

It was okay about 3 days ago.

[euler]

Some changes were made the other day that clearly impacted that. Could you check to see if it is displaying correctly now?

[i707107]

The fonts are still small, but the times are correct now.

[Incompleteusern]

The difficulty bar for problem 957 is orange despite it being at less than 100 solves, in comparison to say 949, 943, 910, and 942. I don't know if this is intentional, or a consequence of the leaderboard being closed early.

[euler]

Yes, that is intentional. The current difficulty metrics use the FS times, so once the table closes, the difficulties are finalised, albeit not as reliable as using 100 members. This is why the spoilers have had such a harmful impact on Project Euler.

[byhill]

I believe then that problem 957 is not appearing in the progress bar graph. With 957, there are 19 problems rated at 90%, but it only says there are 18.

Thank you!

[universalset]

Yes, that is intentional. The current difficulty metrics use the FS times, so once the table closes, the difficulties are finalised, albeit not as reliable as using 100 members. This is why the spoilers have had such a harmful impact on Project Euler.

Are the first (up to) 100 solve times still being used for the difficulty calculation, with the only difference being whether the difficulty is considered finalized and the bar shown in orange, or are only the solves that are displayed in the table being used for the difficulty calculation? With the current rating of the problem I can imagine it being either way. (I hope it's the former...)

[euler]

Once the table is closed it means we have no confidence in any further data. The bar shows orange because we have received all the data we are able to use to finalise the problem's difficulty. Fortunately, it is the early data that has the greatest influence on the metric, so it is not meaningless, just not as reliable as it could be.

[euler]

I believe then that problem 957 is not appearing in the progress bar graph. With 957, there are 19 problems rated at 90%, but it only says there are 18.

Thank you!

Fixed.

[Derfellios]

Hi, I think I found a small bug in the forums. If a post has two pieces of code, of which one is quoted, then extending the code sometimes breaks. See, for example, the post of jhn_nln in https://projecteuler.net/thread=89;page=6. If I open the quoted code first, everything is alright. But if I open the non-quoted code first (and then possibly close it first) and then open the quoted code, the code displays correctly, but the posts below do not move down to make room for the code.

[euler]

Great spot! Thank you for reporting. It should now be fixed.

The error was in the JavaScript and I'm guessing it crept in when I made some recent changes to trigger auto-expansion of code blocks that are short. The code wasn't properly checking the class of parent nodes in the case of nested collapsibles and it was messing up the quote division. I'm just hoping that the fix hasn't had some unexpected knock-on effect to other posts. It shouldn't as the changes should have made the script far more stable, so it was definitely a job worth doing. I'll just deal with those issues separately if they should ever arise.

[tec1234]

mathjax render failed on problem pages caused by CSP violation after recent cdn redirect:

Refused to load the script 'https://fastly.jsdelivr.net/npm/mathjax ... l-chtml.js' because it violates the following Content Security Policy directive: "script-src 'self' https://cdn.jsdelivr.net/gh/highlightjs/ https://cdn.jsdelivr.net/npm/mathjax@3/". Note that 'script-src-elem' was not explicitly set, so 'script-src' is used as a fallback.

[amic]

I just realized that now, but it seems that some posts on the last forum (for problem 970) do not have author associated to them. When you try to quote one of these messages, it writes "Someone" instead.

I do not know if it is intentional, or a bug, or related to the posters which may be anonymous. But I did not remember having seen that in any other solution thread.

[euler]

@amic: Yes, it is due to a recent change. Those members have private profiles. In the past, usernames still showed in the threads, but it makes more sense to hide the details of members with private profiles.

mathjax render failed on problem pages caused by CSP violation after recent cdn redirect:

Is this still happening for you? I have not seen the error. I'm also uncertain why it would be using the subdomain fastly?

[tec1234]

mathjax render failed on problem pages caused by CSP violation after recent cdn redirect:

Is this still happening for you? I have not seen the error. I'm also uncertain why it would be using the subdomain fastly?

Sorry. After recheck, it is due to my local extension. invalid issue.