Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Stats Traffic: Card information fails to load when date cannot be parsed when returned from API #22859

Closed
staskus opened this issue Mar 19, 2024 · 3 comments
Assignees

Comments

@staskus
Copy link
Contributor

staskus commented Mar 19, 2024

self.periodDataQueryDateFormatter.date(from: "2023-10-01") returns nil resulting in "An error occurred" messages shown on Stats Traffic cards

From #22826 (review)

An error occurred when viewing certain weeks

I thought it might be related to Oct having six calendar weeks, but it looks to be related to DST kicking in in my timezone on Oct 1st.

On the team P2 with the week of "Sep 25 - Oct 1, 2023" selected, the screen doesn't load and `An error occurred" is shown:

self.periodDataQueryDateFormatter.date(from: "2023-10-01") returns nil for me, which ends up being due to clocking moving forward one hour at that time. Adding df.timeZone = TimeZone(secondsFromGMT: 0) seems to fix the issue.
This is unrelated to this PR, so not a blocker.

@dangermattic
Copy link
Collaborator

Thanks for reporting! 👍

@staskus
Copy link
Contributor Author

staskus commented Mar 25, 2024

The bug is reproducible and looks to be tightly related to TimeZones and DTS. The one used in Paraguay has a DTS change at midnight 10-01 which is ambiguous to DateFormatter and nil is returned. Even similar examples from StackOverflow reference Paraguay timezone with this particular bug.

DateFormatter:

private var periodDataQueryDateFormatter: DateFormatter {
    let df = DateFormatter()
    df.locale = Locale(identifier: "en_US_POSIX")
    df.dateFormat = "yyyy-MM-dd"
    return df
}
po df.timeZone = TimeZone(identifier: "America/Asuncion")
po df.date(from: "2023-10-01")

nil
po df.timeZone = TimeZone(identifier: "Europe/Vilnius")
po df.date(from: "2023-10-01")

▿ Optional<Date>
  ▿ some : 2023-09-30 21:00:00 +0000
    - timeIntervalSinceReferenceDate : 717800400.

Solution

In documentation Apple suggests to set the timezone to UTC:

When working with fixed format dates, such as RFC 3339, you set the dateFormat property to specify a format string. For most fixed formats, you should also set the locale property to a POSIX locale ("en_US_POSIX"), and set the timeZone property to UTC.

RFC3339DateFormatter.timeZone = TimeZone(secondsFromGMT: 0)

However, this is not a straightforward solution, since in Stats we're working with the dates that are in site's timezone.

@staskus
Copy link
Contributor Author

staskus commented Jun 25, 2024

Closing as won't fix. See #22910 (comment)

@staskus staskus closed this as not planned Won't fix, can't repro, duplicate, stale Jun 25, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment