[squeak-dev] preserving heritage (was: filtered menu in 3.11)
asqueaker at gmail.com
Fri Feb 5 05:53:12 UTC 2010
Ohh, that sounds pretty good.
Until then, though, thankfully Monticello at least informs with the
comment, "source same but revision changed" in the "Patch Browser" (I
assume everyone reviews their changes like me before saving! :) ).
Those methods can be right-clicked, and then select "revert". MC will
load the prior version with the old accounting info.
For reverting to methods prior to the last version:
- note timestamp of the method to revert
- from the Monticello browser,select the package the method is in,
and pressy History.
- Find the version in the list dated just after the date of the method change.
- Spawn history on that one.
- Browse changes, select the method, revert.
On Thu, Feb 4, 2010 at 1:11 PM, Eliot Miranda <eliot.miranda at gmail.com> wrote:
> On Thu, Feb 4, 2010 at 10:10 AM, Ken Causey <ken at kencausey.com> wrote:
>> On Wed, 2010-02-03 at 20:38 -0600, Chris Muller wrote:
>> > Squeak has been such a great journey, and its own historical path is
>> > somewhat captured in the timestamp / author-stamp information of the
>> > individual methods.
>> > I think it is worthwhile to revert methods properly; meaning to truly
>> > revert them not just to the prior code, but to the prior accounting
>> > record associated with that.
>> > With Monticello, it is relatively easy to retrieve the original method
>> > from the prior version. Proper reversion is one reason for keeping
>> > around all the old versions of code.
>> > More importantly, however, I think we should take care to revert
>> > methods properly so that the original history and heritage of the
>> > object-model that is the latest and greatest Squeak image, be
>> > preserved as much as possible.
>> > I do it even with just my own code, even when I'm the same initials,
>> > because it is still worth it to me to know, _when_ was that method
>> > written. How long has that method been sitting there, unchanged.
>> > Again, I think it's worth it, and I hope others will consider the
>> > value of adopting this practice as well.
>> > Thank you,
>> > Chris
>> I wonder if there is some reasonable change to the tools that could be
>> made to make this more natural and therefore require less forethought.
>> I share your goal but often forget.
> Monticello could be made to ask and/or warn if one should automatically
> revert unchanged methods if the previous version is in the package cache.
More information about the Squeak-dev