[Python-checkins] truncation limit
Mon, 7 Feb 2000 19:03:18 -0800 (PST)
On Mon, 7 Feb 2000, Fred L. Drake, Jr. wrote:
> Greg Stein writes:
> > I found that was a bit too small. A lot of simple changes can easily beat
> > 100 lines. Just think about moving a few functions in a file. Or maybe
> > updating a structure and its uses. Or ...
> The description of the checkin should be sufficient for such things
> most of the time.
People are using python-checkins for one of two purposes:
1) monitoring what/how Python is changing
2) code reviewing the changes
Sending the diffs within the email supports (2). Aggressive trimming
simply removes that possibility, so you may as well not send any diffs.
Note that sending full diffs does not obviate (1); it merely provides
additional detail that people may not be interested in.
IMO, I think it is better to enable both purposes.
Guido is excellent at ensuring code is solid before checkin, but mistakes
can and do happen. In addition, there have been proposals to open up
checkins to more individuals. Or, to do some more up-front work to assist
Guido which may end up with the case of Guido not personally reviewing the
code as well and letting something slip by. Additional code review can
> Let's trim the diffs more agressively and cut our losses in this
> thread. It taken way to long to decide on this.
It took zero time. It was decided a month or so ago :-). You mean it is
taking too long to try and *change* the decision :-).
Let's get down to the underlying problem. People don't like the
possibility of large emails arriving in their inbox. That is what this
thread is all about. There isn't anything about "it is best to remove them
because <X>." There is *no* reason to remove the diffs, other than to
reduce the size of the emails. I say: read the checkin comment, then
ignore the diffs if you don't want to read them.
Greg Stein, http://www.lyra.org/