If to-file fails in the middle, should the file still be created?

[qiemem]

No description provided.

[nicolaspayette]

Good question. I'd be tempted to say yes. If you're trying to save the results of a long running experiment, some data might be a lot better than no data (and potentially help you diagnose what happened).

[arthurhjorth]

Definitely yes. As long as the file is closed properly and we don't get
weird I/O issues, I see no reason why people shouldn't get to "keep" what
they've already written to the file.

On Wed, Jul 27, 2016 at 11:00 AM Nicolas Payette [email protected]
wrote:

Good question. I'd be tempted to say yes. If you're trying to save the
results of a long running experiment, some data might be a lot better than
no data (and potentially help you diagnose what happened).

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#10 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/ABxk60tnmMBjYx9eABzhNfgWMdJQCfVHks5qZ4CegaJpZM4JWWIE
.

[qiemem]

Fine with me. That's the current behavior so I'll leave as is!