gobo-eiffel-develop — Development discussions

Re: [gobo-eiffel-develop] Preconditions in creation procedures of DT_FIXED_OFFSET_TIME_ZONE

[Colin P. A. <co...@co...>]

>>>>> "Eric" == Eric Bezault <er...@go...> writes:

    Eric> Colin Paul Adams wrote:
    >> I have changed the m_small_enough precondition in the creation
    >> procedures from h <= 0 to h < 0, as otherwise it was impossible
    >> to create offsets of zero hours and a few minutes.

    Eric> I understand the problem, but I don't think that the fix is
    Eric> correct. Indeed, now it will not be possible anymore to have
    Eric> offsets of zero hours and minus a few minutes. In case of
    Eric> `make_hours_minutes' for example I think that we should
    Eric> rather have:

    Eric>   h_large_enough: h > -Hours_in_day h_small_enough: h <
    Eric> Hours_in_day m_large_enough_positive: h > 0 implies m >= 0
    Eric> m_small_enough_positive: h > 0 implies m < Minutes_in_hour
    Eric> m_large_enough_negative: h < 0 implies m > -Minutes_in_hour
    Eric> m_small_enough_negative: h < 0 implies m <= 0
    Eric> m_large_enough_null: h = 0 implies m > -Minutes_in_hour
    Eric> m_small_enough_null: h = 0 implies m < Minutes_in_hour

OK. I have made that change, and similar in others.
-- 
Colin Adams
Preston Lancashire

Re: [gobo-eiffel-develop] Preconditions in creation procedures of DT_FIXED_OFFSET_TIME_ZONE

[Eric B. <er...@go...>]

Colin Paul Adams wrote:
> I have changed the m_small_enough precondition in the creation
> procedures from h <= 0 to h < 0, as otherwise it was impossible to
> create offsets of zero hours and a few minutes.

I understand the problem, but I don't think that the fix is
correct. Indeed, now it will not be possible anymore to have
offsets of zero hours and minus a few minutes. In case of
`make_hours_minutes' for example I think that we should
rather have:

   h_large_enough: h > -Hours_in_day
   h_small_enough: h < Hours_in_day
   m_large_enough_positive: h > 0 implies m >= 0
   m_small_enough_positive: h > 0 implies m < Minutes_in_hour
   m_large_enough_negative: h < 0 implies m > -Minutes_in_hour
   m_small_enough_negative: h < 0 implies m <= 0
   m_large_enough_null: h = 0 implies m > -Minutes_in_hour
   m_small_enough_null: h = 0 implies m < Minutes_in_hour

-- 
Eric Bezault
mailto:er...@go...
http://www.gobosoft.com

[gobo-eiffel-develop] Preconditions in creation procedures of DT_FIXED_OFFSET_TIME_ZONE

[Colin P. A. <co...@co...>]

I have changed the m_small_enough precondition in the creation
procedures from h <= 0 to h < 0, as otherwise it was impossible to
create offsets of zero hours and a few minutes.
-- 
Colin Adams
Preston Lancashire

Re: [gobo-eiffel-develop] Gobo example for reading attributes in XML?

[Colin P. A. <co...@co...>]

>>>>> "Eric" == Eric Bezault <er...@go...> writes:

    Eric> Bru...@ao... wrote:
    >> Eric:
    >> 
    >> I am looking to read and XML file with Eiffel and it seems that
    >> your Gobo classes can do this.  I looked at the examples for
    >> Gobo and I don't see any example that actually read AND USE an
    >> XML file.
    >> 
    >> Can you point me to any example that show how to use Gobo to
    >> read the attributes of an XML file?  Ideally, it would include
    >> the input XML file as well.

Hello Bruce,

Some of the examples in the $GOBO/example/xml tree do actually read the
attributes of an XML file. What they don't do is to specifically
process the attributes within the example class itself - this is done
in the tree building class XM_CALLBACKS_TO_TREE_FILTER.

The general approach to processing attributes is as Neal has
described.

If you want a more detailed answer, I would suggest describing your
scenario in more detail.
-- 
Colin Adams
Preston Lancashire

Re: [gobo-eiffel-develop] Gobo example for reading attributes in XML?

[Neal L. <ne...@3d...>]

At 02:13 PM 11/18/2007, Eric Bezault wrote:
>Bru...@ao... wrote:
> > Eric:
> >
> > I am looking to read and XML file with Eiffel and it seems that your
> > Gobo classes can do this.  I looked at the examples for Gobo and I don't
> > see any example that actually read AND USE an XML file.
> >
> > Can you point me to any example that show how to use Gobo to read the
> > attributes of an XML file?  Ideally, it would include the input XML file
> > as well.


Hi Bruce,

The two approaches I've used are:

1) Create a descendent of XM_CALLBACKS and process the attributes as 
the parser sends them to your class.  An example of this approach may 
be found at

http://swete.svn.sourceforge.net/viewvc/swete/trunk/swete/test/swete_script_callbacks.e?view=markup

2) Use XM_TREE_CALLBACKS_PIPE to create the xml tree in memory and 
then obtain the information you are looking for from that tree.  For example:


         update_virtual_domain_hosts: BOOLEAN is
                         -- Update table containing all virtual 
domain hosts from file
                         -- Returns true if table was updated
                         -- false if there was a problem reading the 
file and the hash table was not updated
                 local
                         temp_virtual_domain_hosts: DS_HASH_TABLE 
[VIRTUAL_DOMAIN_HOST, STRING]
                         file: KL_TEXT_INPUT_FILE
                         hosts: DS_LIST [XM_ELEMENT]
                         aliases: DS_LIST [XM_ELEMENT]
                         new_host: VIRTUAL_DOMAIN_HOST
                         host_element: XM_ELEMENT
                         read_exception: BOOLEAN
                 do
                         if not read_exception then
                                 create temp_virtual_domain_hosts.make (15)
                                 create host_parser.make
                                 create callbacks.make
                                 host_parser.set_callbacks (callbacks.start)
                                 create file.make 
(configuration.virtual_hosts_file_name)
                                 file.open_read
                                 host_parser.parse_from_stream (file)
                                 hosts := 
callbacks.tree.document.root_element.elements
                                 from
                                         hosts.start
                                 until
                                         hosts.after
                                 loop
                                         host_element := hosts.item 
(hosts.index)
                                         create new_host
                                         new_host.set_host_name 
(host_element.attribute_by_name ("virtual_host_name").value)
                                         new_host.set_company 
(host_element.attribute_by_name ("company").value)
                                         new_host.set_company_abbreviation 
(host_element.attribute_by_name ("company_abbreviation").value)
                                         new_host.set_banner_text 
(host_element.attribute_by_name ("banner_text").value)
                                         new_host.set_url 
(host_element.attribute_by_name ("url").value)
                                         new_host.set_style_sheet 
(host_element.attribute_by_name ("style_sheet").value)
                                         new_host.set_logo 
(host_element.attribute_by_name ("logo").value)
                                         new_host.set_logo_width 
(host_element.attribute_by_name ("logo_width").value.to_integer)
                                         new_host.set_logo_height 
(host_element.attribute_by_name ("logo_height").value.to_integer)
                                         new_host.set_default_state 
(host_element.attribute_by_name ("default_state").value)
                                         new_host.set_required_fields_color 
(host_element.attribute_by_name ("required_fields_color").value)
-- 
new_host.set_program_ready_message (host_element.attribute_by_name 
("program_ready_message").value)
                                         new_host.set_faq_price_answer 
(host_element.attribute_by_name ("faq_price_answer").value)
                                         new_host.set_price 
(host_element.attribute_by_name ("price").value.to_integer)
                                         if equal 
(host_element.attribute_by_name ("use_ssl").value.as_upper, "YES") then
                                                 new_host.set_use_ssl
                                         end
--                                      if equal 
(host_element.attribute_by_name 
("show_requirement_at_greeting").value.as_upper, "YES") then
-- 
new_host.set_show_requirement_at_greeting
--                                      end
                                         aliases := 
host_element.element_by_name ("aliases").elements
                                         from
                                                 aliases.start
                                         until
                                                 aliases.after
                                         loop
                                                 temp_virtual_domain_hosts.force_last 
(new_host, aliases.item (aliases.index).text)
                                                 aliases.forth
                                         end
                                         hosts.forth
                                 end
                                 virtual_domain_hosts.copy 
(temp_virtual_domain_hosts)
                                 file.close
                         end
                         Result := not read_exception
                 rescue
                         read_exception := True
                         retry
                 end

         host_parser: XM_EIFFEL_PARSER

         callbacks: XM_TREE_CALLBACKS_PIPE

Here is the associated XML:

   <?xml version="1.0" encoding="UTF-8" ?>
- <host_list xmlns="http://www.freesafetyprogram.com/hosts">
- <!--


Virtual Domain Host Content
Author:         Neal L Lester [nea...@3d...]
date:           $Date: 2005/03/07 14:45:35 $
revision:       $Revision: 1.3.2.3 $
copyright:      (c) Neal L Lester


   --> - <host virtual_host_name="www.mysafetyprogram.com" 
company="mysafetyprogram.com" company_abbreviation="MSP" 
banner_text="Click here for Bay Area safety consulting services" 
url="http://www.mysafetyprogram.com/resume.htm" style_sheet="fsp.css" 
logo="ff_hepa.gif" logo_width="56" logo_height="80" 
default_state="CA" required_fields_color="Green" 
faq_price_answer="Viewing the final created program costs $249. You 
may view, print, and modify the program (without limitation) for up 
to one year at no additional cost." price="249" use_ssl="yes">
- <aliases>
   <alias>www.mysafetyprogram.com</alias>
   <alias>mysafetyprogram.com</alias>
   <alias>www.freesafetyprogram.com</alias>
   <alias>freesafetyprogram.com</alias>
   <alias>default</alias>
   </aliases>
   </host>
   </host_list>


Neal 

Re: [gobo-eiffel-develop] Gobo example for reading attributes in XML?

[Eric B. <er...@go...>]

Bru...@ao... wrote:
> Eric:
> 
> I am looking to read and XML file with Eiffel and it seems that your 
> Gobo classes can do this.  I looked at the examples for Gobo and I don't 
> see any example that actually read AND USE an XML file.
> 
> Can you point me to any example that show how to use Gobo to read the 
> attributes of an XML file?  Ideally, it would include the input XML file 
> as well.
> 
> Thanks,
> 
> --Bruce Mount
>   Cambridge, MA, USA

Colin or Franck, can you help Bruce?
I'm sure you have some examples like that in the Gobo package.

-- 
Eric Bezault
mailto:er...@go...
http://www.gobosoft.com

Re: [gobo-eiffel-develop] UNIX_FILE_INFO

[Colin P. A. <co...@co...>]

>>>>> "Eric" == Eric Bezault <er...@go...> writes:

    Eric> Colin Paul Adams wrote:
    >> Is there a way to get the information provided from this class
    >> within Gobo? I want to be able to use the functionality in the
    >> XPath library.
    >> 
    >> Initial searching suggests the answer is no.
    >> 
    >> But SE has FILE_INFORMATION - I should think it will be easy to
    >> add a class in Gobo unifying the two.  I can do this if it is
    >> needed.

    Eric> Perhaps we should decide first what to do with SE.  I
    Eric> suggest that we wait until the end of the month to see if
    Eric> Daniel resumes his work on SE 1.2 or not.  Or we could ask
    Eric> him now what his current plans are.

I'm copying this message to the SE 1.2 list:

Hello Daniel,

We need to know what your plans are for SE 1.2.
Ultimately, we will have to drop support for SE at some time if it
doesn't conform to ECMA. We understand that this conflicts with the
goals of the SE transitional project. Perhaps you might consider
changing the goals (if so, we would like to know) to support ECMA (or
at least use free elks).

If not, then we have to make a decision as to how long we continue to
support SE 1.2.  Right now, the question of support for STRING_GENERAL
has become urgent. Already at work, I have had to duplicate a Gobo
routine with one minor edit (that it accepted a STRING_GENERAL instead
of a STRING_8). As it was just one routine, such a bodge is acceptable
for the short term, but we will be faced with doing this a lot. This
will not be acceptable, so Gobo must very soon support STRING_GENERAL.

Last time I asked you, you said you had not decided on whether you
would support this or not. You also said you might well be making
another release in November 2007.
-- 
Colin Adams
Preston Lancashire

Re: [gobo-eiffel-develop] UNIX_FILE_INFO

[Eric B. <er...@go...>]

Colin Paul Adams wrote:
> Is there a way to get the information provided from this class within
> Gobo? I want to be able to use the functionality in the XPath library.
> 
> Initial searching suggests the answer is no.
> 
> But SE has FILE_INFORMATION - I should think it will be easy to add a
> class in Gobo unifying the two.  I can do this if it is needed.

Perhaps we should decide first what to do with SE.
I suggest that we wait until the end of the month
to see if Daniel resumes his work on SE 1.2 or not.
Or we could ask him now what his current plans are.

-- 
Eric Bezault
mailto:er...@go...
http://www.gobosoft.com

[gobo-eiffel-develop] UNIX_FILE_INFO

[Colin P. A. <co...@co...>]

Is there a way to get the information provided from this class within
Gobo? I want to be able to use the functionality in the XPath library.

Initial searching suggests the answer is no.

But SE has FILE_INFORMATION - I should think it will be easy to add a
class in Gobo unifying the two.  I can do this if it is needed.
-- 
Colin Adams
Preston Lancashire

Re: [gobo-eiffel-develop] Callback from C to Eiffel

[Eric B. <er...@go...>]

I tend to agree with Manu. First, when calling Eiffel from C, I think
that we should make the difference between a C program which calls some
Eiffel code, and an Eiffel program which calls some C code which then
calls some Eiffel code. I think that the CECIL mechanism has been
designed for the first case. In the second case, Manu's example works
fine without using CECIL. I think that there is a typo though in Manu's
example:

 > 	c_set_address (your_eiffel_routine_callback)

should probably be:

  	c_set_address ($your_eiffel_routine_callback)

Anyway, I prefer this technique better because the other one with
`eif_type_id' and `eif_procedure' has the drawback that we don't
get a compilation error when we rename the class or the feature.
And if I remember correctly, this technique of passing $feature_name
to the C side is what has already been used to implement Expat
callbacks in the XML library.

As for external features, you should note that gec also has a
limitation when passing $feature_name to the C side. If the
feature has non-expanded argument types then you may experience
some weird behavior. But with expanded arguments, it seems to
work. At least the example below works with gec under Windows:


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
class FOO

create

	make

feature

	make is
		do
			c_set_address ($f)
			c_set_object ($Current)
			foo_c
		end

	c_set_address (p: POINTER) is
		external
			"C inline use <foo_c.h>"
		alias
			"signal_callback = $p;"
		end

	c_set_object (a: POINTER) is
		external
			"C inline use <foo_c.h>"
		alias
			"signal_object = eif_protect($a);"
		end	

	f (i: INTEGER) is
		do
			print (i)
		end

	foo_c is
		external
			"C"
		end

end
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#include "ge_eiffel.h"
#include "ge_gc.h"

extern EIF_REFERENCE signal_object;
extern void (*signal_callback)(EIF_REFERENCE, EIF_INTEGER_32);
extern void foo_c(Void);
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#include "foo_c.h"

EIF_REFERENCE signal_object;
void (*signal_callback)(EIF_REFERENCE, EIF_INTEGER_32);

void foo_c()
{
	(signal_callback) (eif_adopt(signal_object), 1);
}
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

-- 
Eric Bezault
mailto:er...@go...
http://www.gobosoft.com


Emmanuel Stapf [ES] wrote:
> The easiest way for doing this kind of callback is to store the eiffel
> routine address by doing:
> 
> c_set_address (p: POINTER) is
> 	external
> 		"C inline use <your.h>"
> 	alias
> 		"signal_callback = $p"
> 	end
> 
> where you do:
> 
> 	c_set_address (your_eiffel_routine_callback)
> 
> 
> To store the object needed for the call:
> 
> c_set_object (a: POINTER) is
> 	external
> 		"C inline use <your.h>"
> 	alias
> 		"signal_object = eif_protect($a);"
> 	end
> 
> where you do:
> 
> 	c_set_object ($Current) -- Current or any other entity
> 
> Then the C code should simply look like:
> 
> 	(signal_callback) (eif_adopt(signal_object), sig)
> 
> That way there is no need to worry about freezing objects.
> 
> Manu 
> 
>> -----Original Message-----
>> From: gob...@li... 
>> [mailto:gob...@li...] On 
>> Behalf Of Berend de Boer
>> Sent: Thursday, November 15, 2007 3:01 PM
>> To: Gobo Developers Mailing List
>> Subject: [gobo-eiffel-develop] Callback from C to Eiffel
>>
>> Hi Eric,
>>
>> Can I callback from C to Eiffel? Need this at a few places in 
>> eposix, for example to support signals.
>>
>> For example this the code used for ISE:
>>
>>   EIF_TYPE_ID tid;
>>   EIF_OBJECT temp;
>>   if ( signal_switch != NULL ) {
>>     temp = signal_switch;
>>     signal_switch = NULL;
>>     eif_unfreeze (frozen_signal_switch);
>>     frozen_signal_switch = NULL;
>>     eif_wean (temp);
>>   }
>>   if ( a_switch != NULL ) {
>>     tid = eif_type_id ("STDC_SIGNAL_SWITCH");
>>     signal_callback = eif_procedure ("switcher", tid);
>>     if (signal_callback == 0)
>>       { eif_panic ("switcher feature not found."); }
>>     frozen_signal_switch = eif_freeze (a_switch);
>>     signal_switch = eif_adopt (a_switch);
>>   }
>>
>> I.e. the callback pointer is "calculated" at set time and 
>> stored in signal_switch. When a signal occurs, this call is made:
>>
>>   (signal_callback) (frozen_signal_switch, sig);
>>
>>
>> SmartEiffel has a different strategy. When a feature is 
>> exported, SE simply generates a prototype for it in the cecil.h file:
>>
>>   /* Available Eiffel routines via -cecil:
>>   */
>>   void stdc_exit_switch_at_exit(void* C);
>>   void stdc_signal_switch_switcher(void* C,T2 a1);
>>
>> And within C I can simply call stdc_signal_switch_switcher to 
>> make the callback to Eiffel.
>>
>> --
>> Cheers,
>>
>> Berend de Boer
>>

Re: [gobo-eiffel-develop] Callback from C to Eiffel

[Emmanuel S. [ES] <ma...@ei...>]

The easiest way for doing this kind of callback is to store the eiffel
routine address by doing:

c_set_address (p: POINTER) is
	external
		"C inline use <your.h>"
	alias
		"signal_callback = $p"
	end

where you do:

	c_set_address (your_eiffel_routine_callback)


To store the object needed for the call:

c_set_object (a: POINTER) is
	external
		"C inline use <your.h>"
	alias
		"signal_object = eif_protect($a);"
	end

where you do:

	c_set_object ($Current) -- Current or any other entity

Then the C code should simply look like:

	(signal_callback) (eif_adopt(signal_object), sig)

That way there is no need to worry about freezing objects.

Manu 

> -----Original Message-----
> From: gob...@li... 
> [mailto:gob...@li...] On 
> Behalf Of Berend de Boer
> Sent: Thursday, November 15, 2007 3:01 PM
> To: Gobo Developers Mailing List
> Subject: [gobo-eiffel-develop] Callback from C to Eiffel
> 
> Hi Eric,
> 
> Can I callback from C to Eiffel? Need this at a few places in 
> eposix, for example to support signals.
> 
> For example this the code used for ISE:
> 
>   EIF_TYPE_ID tid;
>   EIF_OBJECT temp;
>   if ( signal_switch != NULL ) {
>     temp = signal_switch;
>     signal_switch = NULL;
>     eif_unfreeze (frozen_signal_switch);
>     frozen_signal_switch = NULL;
>     eif_wean (temp);
>   }
>   if ( a_switch != NULL ) {
>     tid = eif_type_id ("STDC_SIGNAL_SWITCH");
>     signal_callback = eif_procedure ("switcher", tid);
>     if (signal_callback == 0)
>       { eif_panic ("switcher feature not found."); }
>     frozen_signal_switch = eif_freeze (a_switch);
>     signal_switch = eif_adopt (a_switch);
>   }
> 
> I.e. the callback pointer is "calculated" at set time and 
> stored in signal_switch. When a signal occurs, this call is made:
> 
>   (signal_callback) (frozen_signal_switch, sig);
> 
> 
> SmartEiffel has a different strategy. When a feature is 
> exported, SE simply generates a prototype for it in the cecil.h file:
> 
>   /* Available Eiffel routines via -cecil:
>   */
>   void stdc_exit_switch_at_exit(void* C);
>   void stdc_signal_switch_switcher(void* C,T2 a1);
> 
> And within C I can simply call stdc_signal_switch_switcher to 
> make the callback to Eiffel.
> 
> --
> Cheers,
> 
> Berend de Boer
> 

[gobo-eiffel-develop] Callback from C to Eiffel

[Berend de B. <be...@po...>]

Hi Eric,

Can I callback from C to Eiffel? Need this at a few places in eposix,
for example to support signals.

For example this the code used for ISE:

  EIF_TYPE_ID tid;
  EIF_OBJECT temp;
  if ( signal_switch !=3D NULL ) {
    temp =3D signal_switch;
    signal_switch =3D NULL;
    eif_unfreeze (frozen_signal_switch);
    frozen_signal_switch =3D NULL;
    eif_wean (temp);
  }
  if ( a_switch !=3D NULL ) {
    tid =3D eif_type_id ("STDC_SIGNAL_SWITCH");
    signal_callback =3D eif_procedure ("switcher", tid);
    if (signal_callback =3D=3D 0)
      { eif_panic ("switcher feature not found."); }
    frozen_signal_switch =3D eif_freeze (a_switch);
    signal_switch =3D eif_adopt (a_switch);
  }

I.e. the callback pointer is "calculated" at set time and stored in
signal_switch. When a signal occurs, this call is made:

  (signal_callback) (frozen_signal_switch, sig);


SmartEiffel has a different strategy. When a feature is exported, SE
simply generates a prototype for it in the cecil.h file:

  /* Available Eiffel routines via -cecil:
  */
  void stdc_exit_switch_at_exit(void* C);
  void stdc_signal_switch_switcher(void* C,T2 a1);

And within C I can simply call stdc_signal_switch_switcher to make the
callback to Eiffel.

=2D-=20
Cheers,

Berend de Boer

Re: [gobo-eiffel-develop] Further problems with regular expressions on Unicode data

[Eric B. <er...@go...>]

Colin Paul Adams wrote:
> The re @.@ matches two @ characters separated by a single character.
> 
> If that character is a non-ascii Unicode character, then the trick
> Franck suggested of replacing both the data and the re with their
> utf-8 byte sequences doesn't work, as the . will match only one byte.

You'll have to replace '.' by it's uft-8 equivalent.
Something like that:

  ([\0-\9]|[\11-\x7F]|([\x80-\xFF][\0-\x7F])|......)

I guess you see what I mean.

-- 
Eric Bezault
mailto:er...@go...
http://www.gobosoft.com

[gobo-eiffel-develop] Further problems with regular expressions on Unicode data

[Colin P. A. <co...@co...>]

The re @.@ matches two @ characters separated by a single character.

If that character is a non-ascii Unicode character, then the trick
Franck suggested of replacing both the data and the re with their
utf-8 byte sequences doesn't work, as the . will match only one byte.
-- 
Colin Adams
Preston Lancashire

Re: [gobo-eiffel-develop] Bug in {RX_REGULAR_EXPRESSION}.replace_all

[Colin P. A. <co...@co...>]

>>>>> "Eric" == Eric Bezault <er...@be...> writes:

    Eric> Colin Paul Adams wrote:
    >>>>>>> "Colin" == Colin Paul Adams <co...@co...>
    >>>>>>> writes:
    >> 
    Eric> What about: <snip>

    Eric> I think that we really need to allocate a string with a
    Eric> capacity of 'subject_end - subject_start + 1' (and not just
    Eric> 'subject_end - subject_start').

I implemented something that was working, but this looks better. I'll
check it in later this weekend if all the tests pass.
-- 
Colin Adams
Preston Lancashire

Re: [gobo-eiffel-develop] Bug in {RX_REGULAR_EXPRESSION}.replace_all

[Eric B. <er...@be...>]

Colin Paul Adams wrote:
>>>>>> "Colin" == Colin Paul Adams <co...@co...> writes:
> 
>     Colin> I don't know if that is the correct result to return, but I
>     Colin> can't think of any other at the moment.
> 
>     Colin> That wasn't sufficient, as the postcondition all_matched
>     Colin> doesn't hold. I will have to dig deeper.

What about:

	replace_all (a_replacement: STRING): STRING is
			-- Substring of `subject' between `subject_start' and `subject_end'
			-- where the whole matched string has been repeatedly replaced by
			-- `a_replacement'; All occurrences of \n\ in `a_replacement' will
			-- have been replaced by the corresponding n-th captured substrings
			-- if any
		require
			is_matching: is_matching
			a_replacement_not_void: a_replacement /= Void
			same_type: ANY_.same_types (a_replacement, subject)
		do
			Result := STRING_.new_empty_string (subject, subject_end - subject_start + 1)
			append_replace_all_to_string (Result, a_replacement)
		ensure
			all_matched: not has_matched
			replace_not_void: Result /= Void
			same_type: ANY_.same_types (Result, subject)
		end

I think that we really need to allocate a string with a capacity
of 'subject_end - subject_start + 1' (and not just 'subject_end -
subject_start').

-- 
Eric Bezault
mailto:er...@go...
http://www.gobosoft.com

Re: [gobo-eiffel-develop] [Colin Paul Adams] Re: Bug in {RX_REGULAR_EXPRESSION}.replace_all

[Colin P. A. <co...@co...>]

>>>>> "Colin" == Colin Paul Adams <co...@co...> writes:

    Colin> I don't know if that is the correct result to return, but I
    Colin> can't think of any other at the moment.

    Colin> That wasn't sufficient, as the postcondition all_matched
    Colin> doesn't hold. I will have to dig deeper.

This might be better - at least the contract holds:

	replace_all (a_replacement: STRING): STRING is
			-- Substring of `subject' between `subject_start' and `subject_end'
			-- where the whole matched string has been repeatedly replaced by
			-- `a_replacement'; All occurrences of \n\ in `a_replacement' will
			-- have been replaced by the corresponding n-th captured substrings
			-- if any
		require
			is_matching: is_matching
			a_replacement_not_void: a_replacement /= Void
			same_type: ANY_.same_types (a_replacement, subject)
		do
			if subject_end = subject_start then
				Result := STRING_.new_empty_string (subject, subject_end - subject_start)
				append_replace_all_to_string (Result, a_replacement)
			else
				Result := subject
				append_replace_all_to_string (Result, a_replacement)
			end
		ensure
			all_matched: not has_matched
			replace_not_void: Result /= Void
			same_type: ANY_.same_types (Result, subject)
		end
-- 
Colin Adams
Preston Lancashire

[gobo-eiffel-develop] [Colin Paul Adams] Re: Bug in {RX_REGULAR_EXPRESSION}.replace_all

[Colin P. A. <co...@co...>]

    Colin> I don't know if that is the correct result to return, but I
    Colin> can't think of any other at the moment.

That wasn't sufficient, as the postcondition all_matched doesn't
hold. I will have to dig deeper.
-- 
Colin Adams
Preston Lancashire

[gobo-eiffel-develop] Bug in {RX_REGULAR_EXPRESSION}.replace_all

[Colin P. A. <co...@co...>]

In this routine, if_subject_end is zero, and subject_start is 1, then
the pre-condition of {KL_STRING_ROUTINES}.new_empty_string is
violated.

Accordingly, I have re-coded it to read:


	replace_all (a_replacement: STRING): STRING is
			-- Substring of `subject' between `subject_start' and `subject_end'
			-- where the whole matched string has been repeatedly replaced by
			-- `a_replacement'; All occurrences of \n\ in `a_replacement' will
			-- have been replaced by the corresponding n-th captured substrings
			-- if any
		require
			is_matching: is_matching
			a_replacement_not_void: a_replacement /= Void
			same_type: ANY_.same_types (a_replacement, subject)
		do
			if subject_end = subject_start then
				Result := STRING_.new_empty_string (subject, subject_end - subject_start)
				append_replace_all_to_string (Result, a_replacement)
			else
				Result := subject
			end
		ensure
			all_matched: not has_matched
			replace_not_void: Result /= Void
			same_type: ANY_.same_types (Result, subject)
		end

I don't know if that is the correct result to return, but I can't
think of any other at the moment.
-- 
Colin Adams
Preston Lancashire

Re: [gobo-eiffel-develop] [eiffel_software] Re: New release: Gobo 3.7

[Eric B. <er...@go...>]

pahihu wrote:
> Hi,
> 
> You could find the Windows 64bit binaries at
> 
> http://web.t-online.hu/edesignweb/gobo37-bin-win-64bit.7z
> 
> Hope this helps,
> Andras Pahi
> 
> --- In eif...@ya..., Eric Bezault <ericb@...> wrote:
>> Vitaliy Filin wrote:
>>> What about 64 bit version?
>> The Eiffel library classes should work fine in 64 bit.
>> If you want to have the Gobo Eiffel tools compiled in 64 bit,
>> you'll have to do that yourself (using the install procedure
>> below). You're then welcome to send me the binary files generated
>> in your bin directory so that I can make them available
>> in:
>>
>>     https://sourceforge.net/project/showfiles.php?group_id=24591

Thank you. It's now available from the Gobo project download page
on SourceForge.

If other people want to share their binary files generated on other
platforms, they are welcome. I'll make them available as well.

-- 
Eric Bezault
mailto:er...@go...
http://www.gobosoft.com

Re: [gobo-eiffel-develop] [eiffel_software] New release: Gobo 3.7

[Eric B. <er...@go...>]

Vitaliy Filin wrote:
> What about 64 bit version?

The Eiffel library classes should work fine in 64 bit.
If you want to have the Gobo Eiffel tools compiled in 64 bit,
you'll have to do that yourself (using the install procedure
below). You're then welcome to send me the binary files generated
in your bin directory so that I can make them available
in:

    https://sourceforge.net/project/showfiles.php?group_id=24591


-- 
Eric Bezault
mailto:er...@go...
http://www.gobosoft.com

 >> In order to avoid binary file incompatibilities between
 >> operating systems, the Gobo package does not contain executables
 >> anymore. Instead of that, the directory $GOBO/bin contains the
 >> C files of the Gobo Eiffel compiler which will be used by a script
 >> to populate this directory with the binary files for your platform.
 >> When installing the Gobo package, you will now have to run either:
 >>
 >> %GOBO%\install.bat msc
 >>
 >> under Windows, or:
 >>
 >> $GOBO/install.sh gcc
 >>
 >> under Unix/Linux in order to generate the binary files
 >> of the Gobo tools in the directory $GOBO/bin.

Re: [gobo-eiffel-develop] New release: Gobo 3.7

[Eric B. <er...@go...>]

Colin Paul Adams wrote:
>>>>>> "Eric" == Eric Bezault <er...@go...> writes:
> 
>     Eric> Hello, A new release of Gobo is available from the usual
>     Eric> location:
> 
> Hooray!
> 
> But looking at the list of known defects, it is a litttle short.
> 
> There are four known defects in the XML parser, for starters. One of
> which even has an entry on the bug tracker (1727015). 

In fact there are other bugs that I didn't report yet. My plan is
to report them in the bug tracker and take more often advantage of
the bug tracker in the future. I should therefore remove the html
page about known bugs, or modify this page so that it points to
the bug tracker.

-- 
Eric Bezault
mailto:er...@go...
http://www.gobosoft.com

[gobo-eiffel-develop] New release: Gobo 3.7

[Eric B. <er...@go...>]

Hello,

A new release of Gobo is available from the usual location:

     https://sourceforge.net/projects/gobo-eiffel/

This version should work with the forthcoming release of
ISE's EiffelStudio 6.1.

In order to avoid binary file incompatibilities between
operating systems, the Gobo package does not contain executables
anymore. Instead of that, the directory $GOBO/bin contains the
C files of the Gobo Eiffel compiler which will be used by a script
to populate this directory with the binary files for your platform.
When installing the Gobo package, you will now have to run either:

     %GOBO%\install.bat msc

under Windows, or:

     $GOBO/install.sh gcc

under Unix/Linux in order to generate the binary files
of the Gobo tools in the directory $GOBO/bin.

For more details about this new release of Gobo, please
read the Readme file and the online documentation:

     http://gobo-eiffel.sf.net/gobo/Readme.txt
     http://gobo-eiffel.sf.net/gobo/index.html

-- 
Eric Bezault
mailto:er...@go...
http://www.gobosoft.com

Re: [gobo-eiffel-develop] Precondition violated in XPath test

[Colin P. A. <co...@co...>]

>>>>> "Colin" =3D=3D Colin Paul Adams <co...@co...> writes:

>>>>> "Colin" =3D=3D Colin Paul Adams <co...@co...> writes:
    Colin> But on the XSLT tests, I'm getting:

    Colin> Compiling C code in C9 In file included from
    Colin> big_file_C9_c.c:30: in273.c:2865:1: warning: integer
    Colin> constant is so large that it is unsigned In file included
    Colin> from big_file_C9_c.c:30: in273.c: In function =81eFbyq6jo=81f:
    Colin> in273.c:2865: warning: this decimal constant is unsigned
    Colin> only in ISO C90 {standard input}: Assembler messages:
    Colin> {standard input}:820975: Warning: end of file not at end of
    Colin> a line; newline inserted gcc: Internal error: Killed
    Colin> (program cc1) Please submit a full bug report.  See
    Colin> <URL:http://bugzilla.redhat.com/bugzilla> for instructions.
    Colin> Internal error: Killed (program cc1) Please submit a full
    Colin> bug report.  See <URL:http://bugzilla.redhat.com/bugzilla>
    Colin> for instructions.  make[1]: *** [big_file_C12_c.o] Error 1
    Colin> make: *** [C12/Cobj12.o] Error 2 make: *** Waiting for
    Colin> unfinished jobs....  make: *** [C11/Cobj11.o] Error 2

    Colin> which isn't good (that's with ISE 6.0 - I'm going to try
    Colin> with gec now).

    Colin> The gec tests work.

    Colin> So this is the ISE problem with gcc 4.0 I presume. I think
    Colin> that's fixed in 6.1, but that's no good to use if Gobo 3.7
    Colin> is supporting 5.7 and 6.0.

I was seeing the machine hanging (presumably with kernel errors).

After rebooting, the xslt tests compile and run fine (with ISE 6.1)
--=20
Colin Adams
Preston Lancashire

Re: [gobo-eiffel-develop] Precondition violated in XPath test

[Colin P. A. <co...@co...>]

>>>>> "Eric" == Eric Bezault <er...@go...> writes:


    >> >> Eric> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ [VDRS-1] class
    Eric> XM_XPATH_LAST_POSITION_FINDER (19,29):
    Eric> `as_last_position_finder' is not the final name of a feature

I deleted this class.

    Eric> And what about the other two?

I've fixed the errors.

In the case of the XM_XPATH_ARRAY_ITERATOR errors, my "fix" was to
comment out the assertion, replacing it by True. I'm assuming gec was
complaining because of the possibility of an expanded generic, and I
don't know a better fix (I understand that the reference keyword is
deprecated). Perhaps you can do better.
-- 
Colin Adams
Preston Lancashire