Message40386
Logged In: YES
user_id=80475
In the interests of saving time and having maximum
accuracy, I recommend reverting and then re-applying first
patch and then making the changes on my list. Try not to
make any other changes.
It took half a day to thoroughly review that 6000 line diff.
It doesn't make sense for me to re-check everything from
scratch. To keep the integrity of the review, make as few
changes as possible.
Also, it does make sense for you to give a self-review to
your final diff (remember, GvR believes in public floggings
for those who break working code).
Here are a couple of fast checks:
-- scan the diff for ''.join and make sure it came from a join
(strdata, "")
-- grep the final files for >= 0: or < 0: to see if the 'in'
operator applies.
-- grep the final files for expand( to make sure an
expandtabs( was not missed.
-- grep the final files for join("\n"), join(" "), or join(' ') to see
if there was an error moving the separator to the front
A slower self check is to read the diff file line-by-line and
mentally recompute each change and compare with the
actual change. The diff review will certainly catch
unintended changes like the merge conflict or the she-
bang change.
I know you can't run each script, but do as much as you
can by importing each script to see if it gets a compilation
error. This won't catch runtime errors but it will catch
syntax glitches.
One important group of scripts can be run. Try to give
IDLE a thorough exercise.
Then, when you're feeling brave, commit.
Good luck! |
|
Date |
User |
Action |
Args |
2007-08-23 15:13:47 | admin | link | issue572113 messages |
2007-08-23 15:13:47 | admin | create | |
|