Getting happier ;-), but wondering if I'm thinking pythonically

[Brian Quinlan]

Timothy Delaney wrote:
> In CPython. Except that if you don't explicitly close files that you
are
> *writing* then you may find that data isn't flushed.

How so? Unless the file object is participating in a cyclic data
structure, when will it not be collected?

> Get used to explicitly closing files - and do it in a finally: block!

No.
 
> But there's no harm, and can be useful as documentation - especially
if
> you have multiple exit points in a function - it can be very useful to
> have the consistency of *every* exit point being a 'return'.

I disagree. The more code that I can see on a screen, the better I can
perceive its logic. Redundant statements just reduce the amount of code
that I can see at once.

> > the decision to exit an application should be made close to the
> > top level (this is true of other languages with exception handling)
> > (e.g. line 52)

> Not necessarily. If there is a fatal condition, choosing to exit at
that
> point is fine. 

How does library code know what constitutes a "fatal condition"? 

> > 5. you should generate more descriptive error messages (e.g. line
30)
> 
> This is related to my comments for point 4. A stack trace is much more
> descriptive ;)

But probably scary and useless to the user. 
 
Cheers,
Brian