[squeak-dev] review of my inbox items

Chris Muller ma.chris.m at gmail.com
Mon Apr 22 01:18:02 UTC 2019

Hi all,

This is just a courtesy note to let you know I'd like to push these
items into trunk sometime soon, perhaps this week.  A couple of these
items received no feedback, which may be a good sign since they were
all well tested.  Here's a summary:


We have a couple of places where we're able to check the
package-cache.  The intention is to improve offline access with a side
benefit of reducing network and server load even when on-line.  There
is no intent to subvert any benefits of the proxy server, it still
helps us when it's not in the cache.


Jakob, the last question we had about this was about restoring the
underscore assignment preference.  At first I struggled to type the
code to subvert other code in the system, going on/off just to avoid
it while leaving underscores in the image.  But I can see that at
least we could ensure _new code_ doesn't get further written with
underscore assignment.  This 431 restores the prior setting.

Message:  Let Strings have the same #name as the object they name,
sans Smalltalk syntax. Allows the #name of either to be used as a
lookup key.

I submitted this in January but, Levente, you recently made a comment
about not liking Object>>#name -- which I can *totally* understand,
but fear it may be too late to go back.  I don't know, so, I didn't
want to advance the system further into dependence on all objects
having a #name before getting your thoughts.  We probably should
discuss in a separate thread.

Keep and access Classes and Traits by their #name's logical value
instead of the physical identity of the String object.

I probably won't push this one this week but was curious whether
anyone has loaded this.  There should be no noticeable impact, please
let me know if that was the case.


More information about the Squeak-dev mailing list