Comments on: An Interesting "Feature" In Ruby’s DateTime.Parse

By: Eric Anderson

Eric Anderson — Thu, 09 May 2013 21:55:00 +0000

Great post! Our team discovered this very same problem. Totally a pain the ass. What’s the best way to check for garbage when parsing dates? What did you end up coming up with?

By: workmad3

workmad3 — Fri, 17 Dec 2010 09:04:33 +0000

Just found another interesting ‘feature’ of DateTime here:

DateTime.parse(“10″) => #

This works for all numeric strings >= “10″ and less than “31″ (although this depends on month, I expect). It gives you an invalid date if you just do

DateTime.parse(“1″)

By: Adam Meehan

Adam Meehan — Fri, 17 Dec 2010 08:59:02 +0000

It is pretty crazy and it’s one of the reasons I created my timeliness gem https://github.com/adzap/timeliness. It is also much faster and allows greater control and custom formats.

By: Adam Meehan

Adam Meehan — Fri, 17 Dec 2010 08:55:46 +0000

It is pretty crazy and one of the reasons I created my timeliness gem https://github.com/adzap/timeliness. The gem is also a lot faster and allows much greater control and custom formats.

By: derick.bailey

derick.bailey — Thu, 16 Dec 2010 22:59:38 +0000

you were never going to use ruby, anyways, john. :)

By: John Sonmez

John Sonmez — Thu, 16 Dec 2010 22:17:22 +0000

Thats it I’m not using Ruby.

By: Paul Keeble

Paul Keeble — Thu, 16 Dec 2010 00:26:17 +0000

Java’s Date parsing is horribly overspecified and SimpleDateFormat is anything but Simple. But accepting any string that contains anywhere the first 3 characters of a day of the week is dangerous. No one would expect it to work that way!

Thanks for the info, I’ll make sure to defend my inputs to parse better.