Issue 5484 in pharo: Unicode depending on zinc

Status: FixReviewNeeded
Owner: [hidden email]
Labels: Type-Cleanup

New issue 5484 by [hidden email]: Unicode depending on zinc
http://code.google.com/p/pharo/issues/detail?id=5484

In

UnicodeClass>>initializeCaseMappings

It's downloading the lowercase-uppercase mappings  
from 'http://www.unicode.org/Public/UNIDATA/CaseFolding.txt'.

But since Unicode will not change a lot in time, we can initialize it via a  
single initialize method with all the unicode data...

Proposal attached

Attachments:
        Unicode class-initializeCaseMappings.st  11.5 KB


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Comment #1 on issue 5484 by [hidden email]: Unicode depending on zinc
http://code.google.com/p/pharo/issues/detail?id=5484

Have a look at what I did with  
http://code.google.com/p/pharo/issues/detail?id=4187

The spec tables are constants, like in your latest example, but I added a  
method that can generate these from an internet accessible resource (using  
Zn as HTTP client). That way, there is some kind of dependency, but not  
really at run time. I think it is cool to literally use these public spec  
documents.



_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Comment #2 on issue 5484 by [hidden email]: Unicode depending on zinc
http://code.google.com/p/pharo/issues/detail?id=5484

Hmm, I don't know, since that introduces dead code and it smells for me


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Comment #3 on issue 5484 by [hidden email]: Unicode depending on zinc
http://code.google.com/p/pharo/issues/detail?id=5484

Better version with the previous version as comment so that we can use  
potentially ring.



Attachments:
        Unicode class-initializeCaseMappings.st  12.0 KB


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Updates:
        Labels: Milestone-1.5

Comment #4 on issue 5484 by [hidden email]: Unicode depending on zinc
http://code.google.com/p/pharo/issues/detail?id=5484

(No comment was entered for this change.)


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Updates:
        Labels: Seed

Comment #5 on issue 5484 by [hidden email]: Unicode depending on zinc
http://code.google.com/p/pharo/issues/detail?id=5484

We need that for Seed to ;)
(and we also need it on CombinedChar)

Ben & Guillermo


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Updates:
        Status: Workneeded

Comment #7 on issue 5484 by [hidden email]: Unicode depending on zinc
http://code.google.com/p/pharo/issues/detail?id=5484

When I execute it I get a DNU "undefined object are not indexable"


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Comment #8 on issue 5484 by [hidden email]: Unicode depending on  
zinc
http://code.google.com/p/pharo/issues/detail?id=5484

I am playing to encode the unicode specification into a gzipped byte array  
like:

fs := FileStream fileNamed: 'UnicodeData.txt'.
fs converter: Latin1TextConverter new.
gzipped := ByteArray streamContents: [ :stream |
        (GZipWriteStream on: stream) nextPutAll: fs contents asByteArray; close ].

method := WriteStream on: String new.
method nextPutAll: 'unicodeData
        ^'.
gzipped printOn: method.
Unicode class compile: method contents classified: 'class initialization'


But gzipped it reaches 200kb, and the image freezes when trying to browse  
it :/.


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Updates:
        Status: FixReviewNeeded
        Cc: [hidden email]
        Labels: -Type-Cleanup -Seed Type-Seed

Comment #9 on issue 5484 by [hidden email]: Unicode depending on  
zinc
http://code.google.com/p/pharo/issues/detail?id=5484

Slice in inbox:

- added unicode spec in a new class with chunked pieces of string, so it  
can be browsed and your image does not get freezed by mistake.
- fixed unicode so it does not depend on zinc.  Old code is still in  
comments for ring I guess.

Name: SLICE-Issue-5484-Unicode-depending-on-zinc-GuillermoPolito.1
Author: GuillermoPolito
Time: 21 May 2012, 2:08:11.714 pm
UUID: f42d8919-a185-4ede-8b42-5b7c11112ccf
Ancestors:
Dependencies: Multilingual-Encodings-GuillermoPolito.24


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Comment #10 on issue 5484 by [hidden email]: Unicode depending on  
zinc
http://code.google.com/p/pharo/issues/detail?id=5484

I latest Pharo Kernel we handle this tables by Fuel. It is quite elegant  
and fast. In Seed I used shorter version of this tables with values up to  
255 that are stored as methods.

I think that the construtors of this tables should be in separtate  
unloadable package that should not be in standard Pharo image by default.


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Comment #11 on issue 5484 by [hidden email]: Unicode depending  
on zinc
http://code.google.com/p/pharo/issues/detail?id=5484

But the other way it is depending on fuel... :/


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Comment #12 on issue 5484 by [hidden email]: Unicode depending on  
zinc
http://code.google.com/p/pharo/issues/detail?id=5484

Fuel will be unloadable but default part of Pharo Kernel. The shrinked  
versions of tables will not be dependent on Fuel and if someone will want  
full tables and not use Fuel then may load full initialization code.


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Updates:
        Status: FixToInclude

Comment #13 on issue 5484 by [hidden email]: Unicode depending  
on zinc
http://code.google.com/p/pharo/issues/detail?id=5484

But why are you using just 255 values and not all of them? space,  
simplicity, speed?

I'd like to have a complete initialization available :).  Making it  
unloadable is a matter of refactor the complete encodings to make them all  
initialized from outside...
I mean, the unloadability is probably cross-encoding, not just unicode.

And this update fixes the zinc/network dependency.  One step at a time :)


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker

Updates:
        Status: Integrated

Comment #14 on issue 5484 by [hidden email]: Unicode depending on  
zinc
http://code.google.com/p/pharo/issues/detail?id=5484

in 2.0 090


_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker