c/profile refactoring

[giampaolo]

BPO	17197
Nosy	@birkenfeld, @pitrou, @giampaolo
Files	profile-refactoring.diff profile-refactoring.diff

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = 'https://github.com/giampaolo'
closed_at = <Date 2013-02-25.10:43:23.273>
created_at = <Date 2013-02-12.22:40:09.472>
labels = ['easy', 'library']
title = 'c/profile refactoring'
updated_at = <Date 2013-02-25.10:43:23.272>
user = 'https://github.com/giampaolo'

bugs.python.org fields:

activity = <Date 2013-02-25.10:43:23.272>
actor = 'giampaolo.rodola'
assignee = 'giampaolo.rodola'
closed = True
closed_date = <Date 2013-02-25.10:43:23.273>
closer = 'giampaolo.rodola'
components = ['Library (Lib)']
creation = <Date 2013-02-12.22:40:09.472>
creator = 'giampaolo.rodola'
dependencies = []
files = ['29053', '29106']
hgrepos = []
issue_num = 17197
keywords = ['patch', 'easy']
message_count = 6.0
messages = ['181980', '182008', '182010', '182026', '182281', '182928']
nosy_count = 4.0
nosy_names = ['georg.brandl', 'pitrou', 'giampaolo.rodola', 'python-dev']
pr_nums = []
priority = 'normal'
resolution = 'fixed'
stage = None
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue17197'
versions = ['Python 3.4']

[giampaolo]

profile and cProfile modules define the same utility functions (run() and runctx()) which use the same code except the profiler class.
Considering that we're going to add 2 new utility functions (runcall() and runblock(), see bpo-9285 and bpo-17130) I think we should refactor the code in order to avoid this code duplication.
Patch in attachment does that.
The approach I came up with looks a bit hackish though so I'd like to get some feedback. Is it acceptable?

[pitrou]

The patch doesn't look right to me. If you import cProfile, profile will always invoke the cProfile profiler.

[giampaolo]

No, it's the other way around. It's from cProfile which I import profile.

diff --git a/Lib/cProfile.py b/Lib/cProfile.py
--- a/Lib/cProfile.py
+++ b/Lib/cProfile.py
...
+import profile as _pyprofile

[pitrou]

No, it's the other way around. It's from cProfile which I import
profile.

diff --git a/Lib/cProfile.py b/Lib/cProfile.py
--- a/Lib/cProfile.py
+++ b/Lib/cProfile.py
...
+import profile as _pyprofile

That's exactly what I'm saying. Once you import cProfile, the attributes
on the profile functions are overriden.
Either way, a module shouldn't mutate another module's functions.

[giampaolo]

Of course you're right. I didn't realize that.
How about this (in attachment)?

[python-dev]

New changeset 422169310b7c by Giampaolo Rodola' in branch 'default':
Fix bpo-17197: profile/cProfile modules refactored so that code of run() and runctx() utility functions is not duplicated in both modules.
http://hg.python.org/cpython/rev/422169310b7c