Can't parse 1970-01-01T00:00:00.000Z as a date

[analytophile]

Trying to plot a time series that uses the following ISO date stamp in Google Chrome:
1970-01-01T00:00:00.000Z

It returns the following error message:
"Couldn't parse 1970-01-01T00:00:00.000Z as a date"

This works for any other date in the format yyyy-mm-ddT00:00:00.000Z

This is apparently a known issue in js on different browsers:
https://stackoverflow.com/questions/2587345/why-does-date-parse-give-incorrect-results

It can be recreated by providing the following data object to a new dygraph plot. If the problematic date is the first in the time series, it does not draw a plot. If it is in the middle, it just skips plotting that data point

Date,Data
1969-12-31T00:00:00.000Z,5.44
1970-01-01T00:00:00.000Z,5.35
1970-01-02T00:00:00.000Z,5.18

[mirabilos]

On Thu, 19 Jun 2025, James Craig wrote: analytophile created an issue (danvk/dygraphs#1062) Trying to plot a time series that uses the following ISO date stamp in Google Chrome: 1970-01-01T00:00:00.000Z It returns the following error message: "Couldn't parse 1970-01-01T00:00:00.000Z as a date" This works for any other date in the format yyyy-mm-ddT00:00:00.000Z This is apparently a known issue in js on different browsers: https://stackoverflow.com/questions/2587345/why-does-date-parse-give-incorrect-results

Interesting, but not related. I’m looking for it, as I have an idea. I found this… riddle me this: dateStrSlashed = dateStr.replace("-", "/", "g"); while (dateStrSlashed.search("-") != -1) { dateStrSlashed = dateStrSlashed.replace("-", "/"); } Should the first line not obsolete the other three?

[mirabilos]

On Thu, 19 Jun 2025, James Craig wrote: It can be recreated by providing the following data object to a new dygraph plot.

CSV only, but yes. For now, please apply the following patch locally: Change lines… 518 if (d && !isNaN(d)) return d; and 533 if (!d || isNaN(d)) { … to… 518 if (d != null && !isNaN(d)) return d; and 533 if (d == null || isNaN(d)) { … (yes, that is != and ==, NOT !== and ===). This should fix your issue. Before I commit this, a question to the original author: @danvk I read about some framework returning null instead of NaN on invalid dates (hence the check above, also catching undefined), but is there ever a chance for dateStrToMillis to return 0 for invalid dates, and if so, does the environment do it consistently (i.e. could we check if dateStrToMillis("invalid") === 0 and if so, compare for 0 as well, if not don’t)? Or might even be better to just let 0 always pass, and it’ll complain if the axis is unsorted. Thanks, //mirabilos -- FWIW, I'm quite impressed with mksh interactively. I thought it was much *much* more bare bones. But it turns out it beats the living hell out of ksh93 in that respect. I'd even consider it for my daily use if I hadn't wasted half my life on my zsh setup. :-) -- Frank Terbeck in #!/bin/mksh

[analytophile]

thanks for the fix!