[squeak-dev] Removing ScriptLoader legacy

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

[squeak-dev] Removing ScriptLoader legacy

NorbertHartl
Are all updateFromXXXX methods in ScriptLoader supposed to stay
where they are? Or will they be removed by any major version
change?

ScriptLoader seems to collect some cruft. There are things
referenced that aren't in the image anymore.

FixVisible

is referenced in updateFrom7024 but is not in the image. There
is ReleaseTest which tests if Undeclared is empty. For 3.10 it
seems that this test has been fixed by manually emptying the
Undeclared dictionary. If you e.g. file out ScriptLoader and
file it back in you will see that there are 2 failures running
the test suite (instead of 1).

So what is updateFrom7024 useful for if an update to 7025 already
happened. Shouldn't any further update delete the former one?
Otherwise many delete operations in updates will lead to cruft,
won't it?

Norbert




Reply | Threaded
Open this post in threaded view
|

Re: [squeak-dev] Removing ScriptLoader legacy

Adrian Lienhard
Hi Norbert

I checked your SLICE-ReleaseTestFixUndeclared-NorbertHartl.1.
Looks good; testUndeclared now runs green.

I agree, ScriptLoader needs some cleanup...

Adrian


On Jun 13, 2008, at 00:36 , Norbert Hartl wrote:

> Are all updateFromXXXX methods in ScriptLoader supposed to stay
> where they are? Or will they be removed by any major version
> change?
>
> ScriptLoader seems to collect some cruft. There are things
> referenced that aren't in the image anymore.
>
> FixVisible
>
> is referenced in updateFrom7024 but is not in the image. There
> is ReleaseTest which tests if Undeclared is empty. For 3.10 it
> seems that this test has been fixed by manually emptying the
> Undeclared dictionary. If you e.g. file out ScriptLoader and
> file it back in you will see that there are 2 failures running
> the test suite (instead of 1).
>
> So what is updateFrom7024 useful for if an update to 7025 already
> happened. Shouldn't any further update delete the former one?
> Otherwise many delete operations in updates will lead to cruft,
> won't it?
>
> Norbert
>
>
>
>