doist / cronwrap Goto Github PK

The time used in the cronwrap email is UTC. My time-zone is UTC+12 and now +13 because of daylight savings.

It is confusing and error-prone having to derive the local time from UTC. So I suggest to use the local time instead of the UTC time when adding the time to the email.

I don't see a button to attach a patch, so I included it as text below.

/usr/local/bin# diff cronwrap cronwrap2
123c123
<     result_str.append('%s UTC\n' % (datetime.utcnow() - timedelta(seconds=int(cmd.run_time))))

---
>     result_str.append('%s\n' % (datetime.now() - timedelta(seconds=int(cmd.run_time))))
126c126
<     result_str.append('%s UTC\n' % datetime.utcnow())

---
>     result_str.append('%s\n' % datetime.now())

I remember using this earlier and thought I would use it again, but it requires Python 2 due to its print statements not using parenthesis. Would you consider releasing a new version that is runnable on non-deprecated versions of Python?

I could of course update the script and submit a PR.

Edit: turns out someone already did the work.

Hi,
I have started using cronwrap a couple of months ago and really liked it. I wanted to have more freedom with the from address of the emails I send out, so I began modifying it a bit. I added additional command line parameters for this and also allowed killing of commands that exceeded their deadline. At that point, this was still supposed to be a pull request.

I then ran into some issues with that because mail doesn't work with the same parameters on all my systems. I hadn't thought of that when I started out. I subsequently switched the mail backend over to the email / smtp library that comes with Python. From there, I realized re-integration with cronwrap would be difficult. I have thus re-branded the project to croncoat and shared it again under the same BSD license. I hope this is fine with everyone here and in line with the open source spirit.

I am also writing here to request some help with the attribution. I could only find "2010 by Plurk" in the script files. Please reply here and/or contact me if you want your name added to a license file. If you consider the differences I list in the readme unfair or if you find bugs, let me know as well.

Kind regards,
Matthias Kauer

[edit] I suppose the code that kills commands after they timed out could be re-integrated with cronwrap if anyone is interested in that.

If I'm reading the code right, timeout is implemented by waiting for the command to complete (as os.system blocks), and only then sending a mail about the timeout.

This doesn't really handle stalled processes right, does it?