13

My typical workflow is to search for something like this (tags change over time):

[c++20] or [c++23] or [c++26]

This worked up until today. Now, while a single tag search works fine, the multi-tag search with ors turns into [c][20] or [c][23] or [c][26], which then fails to find anything at all because [20], [23], and [26] aren't valid tags.

11
  • 3
    Seems related to Doing a tag search and changing the sorting via the navigation buttons breaks queries with operators - the wrong behaviour I reported there is fixed. It was something to do with the lack of + in the URL. However, apparently after the fix the + which is part of the tag is now broken. The character is escaped correctly to %2b but apparently ignored. Commented yesterday
  • Interestingly (?) a search for [c++*] still works, despite expanding to a query you cannot run directly (due to length and this bug). Commented yesterday
  • @jonrsharpe that goes to /questions/tagged/c%2b%2b* and seems to be interpreted fine. Searching [c++20] also works. It seems it's when there is more than one tag that the issue triggers. Commented yesterday
  • 6
    The handling of different things is just doubleplusungood: search for [c++] -[c++20] and you get a search for [c+++-c++20]. But swap the positions and which one is negated, then [c++20] -[c++] leads you to [c] [20] -[c] Commented yesterday
  • I have noticed that C# is also redirected to C Commented yesterday
  • @Canyon I don't get this behaviour. With or without using operators. The only problem I notice is that searching with a wildcard: [c#*] ends up searching for only [c#] Commented yesterday
  • @VLAZ I should have done more testing. This happens when I click C# under the "Popular tags" of the community activity widget. It does not happen when I click a C# badge under a post Commented yesterday
  • @Canyon Yes, that's broken: Popular tags links not navigating to C# questions Commented yesterday
  • 6
    Definitely related to my work on Doing a tag search and changing the sorting via the navigation buttons breaks queries with operators, though I'm not sure how it happened because I explicitly had unit tests covering C++. Strange! I'll fix it today. Commented yesterday
  • 7
    Okay, the source of the bug is that the URL c%2b%2b20+or+c%2b%2b23 gets parsed as c+++or+c++23, which makes it equivalent to c or c 23, whoops. Commented yesterday
  • 5
    Alright, I made an incorrect assumption that our tag query encoding was supposed to be encoding spaces as + when I made the fix in Doing a tag search and changing the sorting via the navigation buttons breaks queries with operators. The intended encoding was to encode spaces as %20 instead. I've got a fixed developed and will get it built out for y'all as soon as it gets reviewed and approved. Commented yesterday

1 Answer 1

4

This has been fixed. A regression was introduced by my fix in Doing a tag search and changing the sorting via the navigation buttons breaks queries with operators where spaces in tag queries were being encoded with +, which broke lots of other assumptions in the code. They're now encoded with %20 instead and work as expected.

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.