On 10/14/10 02:10 PM, [email protected] wrote:
On 10/14/10 07:48 AM, Jan Damborsky wrote:
On 10/14/10 01:05 PM, [email protected] wrote:
On 10/14/10 03:02 AM, Jan Damborsky wrote:
Hi Keith,
On 10/13/10 06:36 PM, Keith Mitchell wrote:
Hi Jan,
It'd be better to check status and len() of the output before
splitting the output - if the command errors out and returns an empty
string for status, the "co[0].split()[1]" command will fail.
Good point. I went with what Bart suggests, since I think it handles
all kind of potential failures.
If you are interested, I have updated the webrev with those changes:
http://cr.opensolaris.org/~dambi/bug-6991546/
Thank you for review !
Jan
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
Hey Jan,
Would it be better to trap on IndexError instead of StandardError?
Hey Joe,
to be honest, I am not sure. I chose StandardError, as it has broader
scope
and we want to catch everything.
Looking at the exception hierarchy [1], the related portion looks like
StandardError
+--TypeError
+-- LookupError
+-- IndexError
+-- KeyError
Is IndexError the only exception which could be raised ?
If this is the case, then I agree it is better to go with more specific
IndexError.
Thank you,
Jan
[1] http://docs.python.org/library/exceptions.html
Hey Jan,
A while back Keith had told me the way to determine what exception to
trap is to induce the expected error. So I tried this:
boomvangLT > python
Python 2.6.4 (r264:75706, Sep 12 2010, 16:30:49) [C] on sunos5
Type "help", "copyright", "credits" or "license" for more information.
Sourcing ${HOME}/.pythonrc
py 1> co = ''
py 1> x = co[0].split()[1]
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
IndexError: string index out of range
py 1>
Since the exception raised is IndexError I thought I would raise the
issues.
Yep, understood. The code covers that scenario, since 'IndexError' is
derived
from 'StandardError'. I am trying to understand if 'IndexError' is the only
exception which could occur.
In looking at the Python documentation regarding exceptions:
http://docs.python.org/tutorial/errors.html
The safest thing may be to have an except IndexError and an
unspecified catch all:
2255 try:
2256 preferred_publisher = co[0].split()[1]
2257
2258 except IndexError:
2259 prerror('Could not determine name of preferred
publisher from '
2260 'following input : ' + repr(co))
2261 prerror('Failure. Returning: ICT_PKG_RESET_UUID_FAILED')
2262 return ICT_PKG_RESET_UUID_FAILED
except:
prerror:('Unexpected failure')
raise
2263
2264 _dbg_msg('Preferred publisher: ' + preferred_publisher)
I believe that for purposes of evaluating potential failure, it is useful
to have 'repr(co)' always captured in log file. Then it could be
easily determined why failure happened. So I am not sure if
splitting 'except' clause into two parts buys us anything.
Let's see what Keith suggests too.
Yep. I agree :-)
Jan
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
