gmane.comp.lang.eiffel.smalleiffel

Raphael Mack | 8 Jun 21:46

Unicode

Hi,

I started to use UNICODE_STRING (not that I need any unicode support, but I 
want to use the new XML code in se libs). A few questions come up:

Will inspect be extended to unicode strings? - I fear using "inspect 
some_unicode.to_utf8" is not very efficient.

I need to convert UNICODE_STRINGs to integers. What is the way to go? 
convert to a utf8 string and then to a integer as 
unicode.to_utf8.to_integer or will these convertion features make it into 
UNICODE_STRING once upon a time?

Are there any timing comparisons (estimates?) between STRING an 
UNICODE_STRING?

How good is the current svn-compiler? (I got c compilation errors using 
svn-9151)

Cheers,
Rapha

headers

Dominique Colnet | 9 Jun 18:07

Re: Unicode

Raphael Mack wrote:
> Hi,
>   
Hi,
> I started to use UNICODE_STRING (not that I need any unicode support, but I 
> want to use the new XML code in se libs). A few questions come up:
>   
So, you tell us that using XML code of se libs raise the need for inspect on
UNICODE_STRING?
> Will inspect be extended to unicode strings? - I fear using "inspect 
> some_unicode.to_utf8" is not very efficient.
>   
For sure it cannot be as efficient as for other types. Even STRING inspect
are fast.
> I need to convert UNICODE_STRINGs to integers. What is the way to go? 
>   
> convert to a utf8 string and then to a integer as 
> unicode.to_utf8.to_integer or will these convertion features make it into 
> UNICODE_STRING once upon a time?
>
> Are there any timing comparisons (estimates?) between STRING an 
> UNICODE_STRING?
>   
I have not made comparison, but it is probably much more slow than
simple STRINGs.
> How good is the current svn-compiler? (I got c compilation errors using 
> svn-9151)
>   
Not in a good shape. We have started a huge modification (inlining
of all dynamic dispatch calls with Eiffel "inspect").

(Continue reading)

headers

Raphael Mack | 9 Jun 22:19

Re: Unicode

Hello,

thanks Dominique for your answer,

Am Montag, 9. Juni 2008 schrieb Dominique Colnet:
> Raphael Mack wrote:
> > I started to use UNICODE_STRING (not that I need any unicode support,
> > but I want to use the new XML code in se libs). A few questions come
> > up:
>
> So, you tell us that using XML code of se libs raise the need for
> inspect on UNICODE_STRING?

not directly, but my sax parser based on the XML code in se cries for it...

> > Will inspect be extended to unicode strings? - I fear using "inspect
> > some_unicode.to_utf8" is not very efficient.
>
> For sure it cannot be as efficient as for other types. Even STRING
> inspect are fast.

mh. Maybe, if the unicode raw data is handled as the bytes in STRINGs and 
the literal strings are converted to unicode raw data... Just an idea, I 
know you do not have much spare time.

> > I need to convert UNICODE_STRINGs to integers. What is the way to go?

Any suggestions on these

> > How good is the current svn-compiler? (I got c compilation errors

(Continue reading)

headers

Frederic Merizen | 10 Jun 11:54

Re: Unicode


> Hello,
>
> thanks Dominique for your answer,
>
> Am Montag, 9. Juni 2008 schrieb Dominique Colnet:
>> Raphael Mack wrote:
>> > How good is the current svn-compiler? (I got c compilation errors
>> > using svn-9151)
>>
>> Not in a good shape. We have started a huge modification (inlining
>> of all dynamic dispatch calls with Eiffel "inspect").
>> I have the hope to reset the whole test_suite for, let's say, one month.
>
> I feared that, on reading some of the last commit messages. As I already
> use some parts of Cyrils work on the libraries I need to work with a later
> version than the last release. Could you suggest a svn revision, before
> the ongoing major rework?

The big change came with revision 8964, so rev. 8963 is where I'd start
looking.
--

-- 
Frederic Merizen
If you keep an open mind people will throw a lot of garbage in it.