Code review comment for lp:~edwin-grubbs/bzr/bug-384158-passing-body-to-mutt

So I think to do this correctly, we would need to have a way to define the lifetime of the NamedTemporaryFile, versus the lifetime of running the mutt command.

I *think* if you just assigned the temp file to 'self' then it would survive long enough for mutt to find it.

If you just did

temp_file = NamedTemporaryFile(...)
temp_file.write()
...

args.extend(...)
...
return args

Then the temp file will be destroyed as soon as the function returns.

However if you did:

self._temp_file = NamedTemporaryFile(...)

then the file will be around until the Mutt class is destroyed, which should be long enough.

I'll also note that NamedTemporaryFile won't work on windows (it gives you the path to a file, but that file cannot be opened even within the same process... :(

Even so, I'm not worried about people using mutt on Windows, so I'd still like to try.

Anyway, I think this is good enough to merge, but it would be nice if you could try the NamedTemporaryFile trick again, with a longer lifetime.