File.utime returns Invalid argument on Windows (winXP, ruby 1.8.5/1.8.6).

[Matt Scilipoti]

I am receiving an "Invalid argument" error from File.utime.

irb> testfile = 'd:/temp/2007/06/test.txt'
=> "d:/temp/2007/06/test.txt"
irb> File.utime(0, Time.now, testfile)
Errno::EINVAL: Invalid argument - d:/temp/2007/06/test.txt
from (irb):4:in `utime'
from (irb):4
irb> File.exist?(testfile)
=> true
irb> File.writable?(testfile)
=> true

My local user group suggested:
irb> require 'time'
=> false

False?

I tried this on two computers:
| WinXP SP2 | WinXP SP2 |
| Ruby 1.8.6 | 1.8.5 |
| NTFS | FAT32 |

A friend has these same specs, but it works for him.
Google for 'ruby, windows, File.utime, "Invalid Argument"' (and various versions
of this) doesn't help. It does appear in a nice Ruby poetry jam.

Any ideas?

 

[Daniel Berger]

I am receiving an "Invalid argument" error from File.utime.

irb> testfile = 'd:/temp/2007/06/test.txt'
=> "d:/temp/2007/06/test.txt"
irb> File.utime(0, Time.now, testfile)
Errno::EINVAL: Invalid argument - d:/temp/2007/06/test.txt
from (irb):4:in `utime'
from (irb):4
irb> File.exist?(testfile)
=> true
irb> File.writable?(testfile)
=> true

Is D: a hard disk? Or is it a CDROM, usb drive, etc? Hey, gotta ask.

My local user group suggested:
irb> require 'time'
=> false

False?

Unrelated. It means it was already loaded, probably by rubygems. An
actual failure would raise an error.

Regards,

Dan

 

[Matt Scilipoti]

Solved (mostly), thanks to:

Nicholas Evans, who suggested that Windows may not accept a date of 0
(the first param sets access time). It doesn't. I knew that. Bad
brain. Thanks for asking an "is it plugged in" type of question.

Daniel Berger, for reminding me that "require 'x' -> false" doesn't
mean "we can't find 'x'" - it just means that it is ALREADY required.
I knew that too. Funny how, in this context, I easily accepted an
incorrect translation of the results, because it was corroborating
evidence - exactly what I was looking for. And thanks for another "is
it plugged in?" question. Yes, D: drive is a hard drive.

Why "mostly"? This code is derived from a battery of tests for a
project in a User Group. These tests pass for another Windows user.
We will investigate next meeting.

Thanks to all,
Matt