[Python-ideas] Historical revisionism and source file formatting
talin at acm.org
Mon Apr 6 10:17:15 CEST 2009
As you know, the current Python source tree contains a mixture of
various different tab/space conventions and source formatting styles.
The Python community has opted not to perform any sort of mass
conversion of source file to a uniform style, because the community
values time-wise consistency (of versions and branches) over space-wise
consistency (between files in the same revision).
However, I have often wondered if it would be reasonable to reformat
*all prior* versions of the source code. Imagine some kind of tool that
would iterate over all previous versions, copying each version from one
repository to another, and doing an automatic reformat on each version.
Of course, in order for this to work, the automatic reformatter would
have to be very, very trustworthy, since any errors would likely go
undetected for a long time.
One objection to this plan is that the history thus created would be
"revisionist" - that it would be, in fact, an unauthentic historical
record of what had happened. I don't have a good answer to that other
than to say that this is an argument of theory over practicality.
In any case, I don't mean this as a serious suggestion that we reformat
all Python revisions in the current SVN repository.
However - since we're going to be setting up a new DVCS anyway, I wonder
if something along these lines would make sense as part of the
migration. It might be a one-time opportunity to deal with this matter
once and for all.
More information about the Python-ideas