Hi Nao,
> Hi all,
>
> There has been no bug report since the release of iMops 1.0.
> But I have found 5 bugs and fixed them.
Good work!!
I've done absolutely no Mops stuff for a year or more. Just too many
other things going on. I'm giving some help to another software project
that some friends of mine are doing, but it's all in C++. I still
prefer Mops but I'm stuck with C++ for this other project :-(
>
> Although not many improvements nor extensions have been done
> on iMops 1.0, the new version should be uploaded,
> for already 3 months have passed since the last release.
> The version number will be 1.01.
>
> By the way,
> the present word search order in "FIND" is,
> 1. locals
> 2. methods
> 3. temporary objects
> 4. ivars
> 5. public words
> 6. try number -> abort.
>
> When "ANSI?" flag is on, method search will be skipped.
> At present, it is the only meaning of the flag "ANSI?".
>
> I tried to change the word search order like following,
> 1. locals
> 2. temporary objects
> 3. ivars
> 4. public words
> 5. methods
> 6. try number -> abort
>
> Then words whose name ends with colon become findable.
> This will allow to define and call ANSI forth standard words
> having the last colon names.
> On the other hand, that causes a restriction on method names.
> And we cannot unconditionally consider a last-colon word
> as a method selector, any more.
>
> I don't know which is better.
> Any opinions?
I don't have any strong opinions in this area (although locals first
seems best to me) -- I've always tried to keep names I'm using unique so
the search order wouldn't make any difference. So I don't object to
having methods lower on the list, so names ending in colon could be used
as normal words. I'd never use a word ending in colon like that myself,
since for me it would make the code less readable, making something look
like a method call when it isn't. But somebody might want to port some
ANSI code which uses a name like that, and doing it the new way would
make this porting easier. So it's probably a good idea.
>
> Anyway, the next version (1.01) will try the new search order.
> The reversion is easy.
>
> Another topic.
> I am thinking over a new message binding mechanism
> which doesn't use PARSE (tick).
> That is, at least theoretically, possible.
> But Mops has rather rich variety of the binding syntax,
> so that implementing it will takes some time.
> (... I am a poor man, so my real life tortures me. :-( )
> Additionally, it is not really an improvement.
> However, that may be an interesting experimentation.
>
>
> Anyway,
> the next version of iMops will be uploaded in a week or so.
Great work!! I'll be away for a few weeks but I'll see if I can do some
testing with it when I get back.
Cheers, Mike.
--
---------------------------------------------------------------
Mike Hore mike_hore@...
---------------------------------------------------------------
|