Forum

feature 

    open_file_sample
        local
            l_file: UNIX_FILE_INFO
            l_path: STRING
        do
            make
            l_path := "/var/log/syslog"
            l_file.update (l_path)
            if l_file.parent_directory.exists and then l_file.parent_directory.is_writtable then
                create l_file.make
            end
             -- AS the above statement doesn't exist!
            check
                syslog_file_exists_and_is_readable: l_file.exists and then l_file.is_readable
            end
        end

Is this the proper way to check for file existence in Eiffel?

I was wondering if there is a way not to create 2 objects. I'll complete my check with following statement:

• define path `l_file_path := "/some/path/with_file.log"
• check if parent directory exists and has rights to write into
• create log file

I didn't find any notion of timezone DATE_TIME class and TIME class. Am I wrong?

How do I handle timezones with Eiffel library?

Is there a way to specify it from the {DATE_TIME}.formated_out ("YYYY-MM-DD [0]hh:[0]mi:[0]ss")

Is there a setter for timezone?

I already got this problem with the logger which is logging with UTC...

The only workaround I found was l.hour_add (-3) before the formated_out which is really dirty

Does this work for removing first two children from a node?

remove_children(current_node : CURSOR_TREE[EXPRESSION]) 
        do
            current_node.down (1)
            current_node.remove
            current_node.forth
            current_node.remove
        end

Or does that only remove the current_node?

Reading Design by Contract tutorial I stumbled upon the following line:

Contracts in Eiffel are not just wishful thinking. They can be monitored at run time under the control of compilation options.

followed by explanations that they will throw exceptions when fail. It makes me think that all the require ensure invariant all checks can be either performed at runtime or turned off. Is this correct? Or they can be enforced at compile time as well using appropriate compiler options?

I'm calling a procedure with an argument which is an integer_64. I implemented a WATT class which can create it from an INTEGER_64 and it seems the execution stops when reached this point, where am I wrong?

Catcall detected for argument#1args': expected TUPLE [!WATT] but got TUPLE [INTEGER_64]`

Attached case (Update)

Actually when checking with syntax

attached {INTEGER_64} my_watt_object as l_int

it doesn't pass either... is it the expected behaviour?

Actually it seems for me that the semantic cases are the same which have to validate the conformity step... for me (but seems not to be the case for the definition of the language between conformance/conformity) which says

Conformance and convertibility are exclusive of each other, p.87

Is the conformance rule valid for a type which defines as convert a type to another which is my case from WATT to INTEGER_64?

As I can see into the debugger it's possible to get the operands, and name of procedure, is there a way to get it?

• PROCEDURE=>operands returns a detachable that seems return the operands only when they have been setted into the agent
• Do I have pass through any REFLECTOR class because the PROCEDURE class doesn't have this function and in this case why?

Seems that estudio has access to informations as ROUTINE client don't have, why is he a privileged one? is he cheating?

Eiffel: What's the difference between an "eiffel" library and a gobo library? for example with xml, which should I choose and why?

I can't really use Io.read_integer as that just ignores everything except the first number.

I can use Io.read_line to get something like 15 14 59 86. How would I split these into integers now?

Javascript has split(), C++ has stringstream, something similar would be ideal.

Whats the difference between

• object test

  if attached {DOG} an_animal as a_dog then
     a_dog.eat (meat)
  end
  

• operator / of the class TYPE

  if an_animal / a_dog then
     an_animal.eat (food)
  end
  

• reference equality =

  if a_dog = an_animal then
      a_dog.eat (meat)
  else
      an_animal.eat (food) 
  end
  

• object equality ~

  if a_dog ~ an_animal then
      a_dog.eat (meat)
  else
      an_animal.eat (food) 
  end
  

And where can I find documentation about that?

As default_create is always a an example of creation method which can be used for instance creation as create some_instance if I don't want to allow the descendants of a class to use it as a creation method is there a way to do that?

A

deferred class A
feature
   make (a_db_connection: DB_CONNECTION)
      do
          default_create
          db_connection := a_db_connection
      end

B

class B
inherit
    A
create
    make
    -- default_create -- I'd be able to do that and want to avoid it

Trying to create an object which is decided @runtime I have a similar situation as:

class ZOO

feature

    animals: LINKED_LIST[ANIMAL]

    default_create
        do
            create animals.make
            animals.extend(create {LION})
            animals.extend(create {SERPENT})
            animals.extend(create {BIRD})
        end

    open
        local
            l_sector: ZOO_SECTOR[ANIMAL]
        do
            across
                animals as animal
            loop
                create {ZOO_SECTOR[animal.item.generating_type]} l_sector
            end
        end

on create {ZOO_SECTOR[animal.item.generating_type]} l_sector the compiler disagrees with me, I tried with l_type: TYPE[ANIMAL] and create {ZOO_SECTOR[l_type]} l_sector which neither works. Am I obligated to do something like that? which will be for me a contradiction with polymorphism flexibility, I think I'm missing a mecanism/statement

open
    local
        l_sector: ZOO_SECTOR[ANIMAL]
    do
        across
            animals as animal
        loop
            if attached {LION} animal.item then
                create {ZOO_SECTOR[LION]} l_sector
            else if attached {SERPENT} animal.item then
                create {ZOO_SECTOR[SERPENT]} l_sector
            else
                .....
        end
    end

How do I compare the type of an object to a given type (instanceOf statement in Java)?

do_stuff (a_type: TYPE)
    local
        an_object: ANY
    do
        an_object := get_from_sky
        if an_object.instanceOf (a_type) then
            io.putstring("Object is same type as parameter")
        else
            io.putstring("Object is NOT same type as parameter")
        end
    end

Animal

deferred class ANIMAL
inherit
    ANY
        redefine
            default_create
        end
feature

    creator: like Current

    guts: GUTS

    default_create
        do
            create guts
        end

    make_malformed
        do
            default_create
        end

end --class

PIG

class PIG

inherit
    ANIMAL
        redefine
            make_malformed
        end

create
    default_create,
    make_malformed,
    make_from_insemination

feature
    guts: GUTS

    make_malformed
        do
            Precursor
            set_left_eye (create {MALFORMED_EYE})
        end

    make_from_insemination (some_humain: HUMAIN)
        do
            default_create
            creator := some_humain
        end

end --class

Into my vision of best practices, I'll say that

• If there is no particular sense of making a creation procedure (like my make_malformed example) redefine the default_create
• All creation procedure should call default_create and add specific behavior (like my make_from_db example)
• So what is the purpose of many libraries in Eiffel which are adding a make like create {LINKED_LIST}.make

Correct me if I'm wrong. Thanks in advance!

Recommendations I understood in Java (which has a lot of restrictions, @ least for me), even with hibernate was to have separated layers

• Entities like persons, children, users, etc...
• DAO entities linked to database
• Service providing entities and functionalities, where I'll do the SQL
• WebService providing an interface over needs

As I'm starting with Eiffel and store, I'm facing some difficulties I face since ever in programming (hoping there's somebody in this earth who has not the same problem) I always want to generalize things more than necessary. Every time I do a copy-paste, I refactor and look for a solution which makes me able to write it one time... which takes time and time on the delivery of the software, but for me adds more quality and flexibility to the software. I'm actually working alone in a company where I'm going to be the lead developer and if the future wants we'll be more developers. The goal is to develop a platform of services in Eiffel, postgresql-odbc, and an Angular-web front-end.

I'd like to have the more generic pattern to be able to manage entities in the future with typical situations as:

• Database entities
• Relationships
  • one_to_one
  • one_to_many
  • many_to_one
  • many_to_many

@ The point I'm now, I'm about to develop an architecture which ideally for me has:

• DB_ENTITY which as relations: BAG[RELATIONSHIP[P,S]] where P=Primary and S=Secondary
• Primary is P->DB_ENTITY when ONE and BAG[P] when MANY
• A COMPANY on my design will inherit from DB_ENTITY and add relationships as a BRANCH. So I was thinking having in my COMPANY class branches: RELATIONSHIP[like Current, BRANCH]

The relationship classes would help me to create the CRUD SQL statements into the "service" layer in a more abstract manner.

• when I try something more lightweight I find restrictions in the pattern where I have to repeat operations... thats a bit my difficulty
• Do you think of any disadvantages of such model I'm creating out of the first shot of development?

Following with strange patterns for some, can't I do that? The compiler says Invalid constraint for formal generic parameter

class PARENT[G -> CHILD[like Current]]

feature -- Access

    children: LIST[G]

end

class CHILD[H -> PARENT[like Current]]

feature -- Access

    father: H
end

to be able to do something like

class WIDOW_PARENT

inherit
    PARENT[BLACK_CHILD]

end

class BLACK_CHILD

inherit 
    CHILD[WIDOW_PARENT]

end

If I don't do it with genericity, I'd have to redefine the children collection from

• children: LIST[CHILD] to children: LIST[BLACK_CHILD] into the WIDOW_PARENT class
• father: PARENT to father: WIDOW_PARENT into the BLACK_CHILD class

instead of only specify it in the inherit clause... Hope it makes sense

Update

As I solved it with Alexanders answer, I'm stuck further doing a conformity check. I'm trying to set an HTTP router depending on entities and if its a child entity it should be able to do a http://host:port/entity/child_entity/id to get all child entities from entity. For that I'd like to add to the generic router a check. On something like ANY_PARENT_DB_ENTITY such as

if ({G}).conforms_to ({CHILD_DB_ENTITY[ANY_PARENT_DB_ENTITY]}) then
    friend.act_like_a_father 
else
    friend.act_like_a_mate
end

Is there a way to test a given Generic parameter of a Class without any attached instance of it?

class BAG[G -> MOUSE]

feature -- 

    discriminate
        do
            if G.conforms_to (MAMMAL) then
                io.putstring ("you gave me a real animal")
            elseif G.conforms_to (COMPUTER_ACCESSORY) then
                io.putstring ("Seems you don't like animals such as computers")
            else
                io.pustring ("Still dont know what you are dealing with")
            end
        end

After starting it with F5, it seems that when I stop an EWF application (from estudio stop button) it takes up to 10 seconds to free the defined port on linux debian machine, Don't know if its only on this OS but is there a way to make it stop quicker? I always have to wait for some seconds to start it again after some corrections.

As there is an initialize function, is there an exit/on_exit function such as I can close my database connection into it?

Trying to write some logger enhancements I'd like to deal with ANSI codes interpretation, and it seems that it's not working with the standard io.putstring method neither with print I wonder there is a way to do something such as

echo -e "\u001B[31mSome_red_txt" in bash

Seems other languages can do it but I can't in Eiffel.

I'm trying to create some terminal logging colors to be able to see clearer my errors. Doing that it seems to me obvious that I'd have to create constants like followings. As I don't want to create an instance each time I call my constant, it makes sense for me to do something like that, but the compiler doesn't seem to have the same conception as me...

A once function has generic or anchored result what is an anchored result?

As the compiler has always the last word and me the forelast, why am I wrong and is he right??

class
    TERMINAL_COLOR

create
    make

feature -- Initialization

    make (a_fg: like foreground; a_bg: like background)
        do
            foregound := a_fg
            background := a_bg
        end

feature -- Status report

    foreground: INTEGER

    background: INTEGER

feature -- Colors

    Black: like Current
        once -- compiler doesn't agree with me
            create Result.make (30, 40)
        ensure
            instance_free: class
        end

end

as my post describes it, I'd like to create an UTIL class with a never_both function.

class
    UTIL

create
    default_create

feature -- could be in class BOOLEAN

    double_implies, reversible_implies, never_both (a, b: BOOLEAN): BOOLEAN
            -- Into boolean class with never_with
        do
            if a and b then
                Result := False
            else
                Result := True
            end
        end

end

When I use it

invariant
    never_both: {UTIL}.never_both (attached last_error, attached last_success_message)

the compiler complains about a VUNO error

never_both used in the non-object call is not a class feature.

I saw 2 notations about objects creating - {UTIL}.never_both (a, b) - ({UTIL}).never_both (a, b)

Whats the difference between them?

How to create an application wide (could be even world wide once if you want!) object for the use of this UTIL if possible in Eiffel?!

I know this is a TUPLE of questions so I put them in Bold

Defining various names for a same feature seems possible in Eiffel, what for the attributes?

Is it possible to name an attribute like:

background, bk: COLOR
foreground, fg: COLOR

If not why?

I'm trying to get an answer from ODBC postgresql database when doing a simple

DELETE * from some_table

with following code, I'm unable to get the DELETE 1 classical answer? affected_row_count was my guess but seems its not is_affected_row_count_supported. I have the same problem with inserts...

delete_entity (a_primary_key: INTEGER_64)
        -- Delete from given primary_key object
    require
        valid_db_connection: attached db_connection.base_update
                    and attached db_connection.session_control
    local
        l_void: like item
    do
        if attached db_connection.base_update as l_base_update
                and then attached db_connection.session_control as l_session_control then
            l_base_update.modify (delete_from_primary_key_query(a_primary_key))
            if l_base_update.is_ok then
                last_success_message := "Entity with ID " + a_primary_key.out + " deleted from DB successfully"
                if l_base_update.is_affected_row_count_supported then
                    check
                        only_one_affected_row: l_base_update.affected_row_count = 1
                    end
                    if attached last_success_message as l then
                        l.append (" affected rows:" + l_base_update.affected_row_count.out )
                    end
                end
                if attached last_success_message as l then logger.write_information (l) end
            else
                create {INSERT_ERROR} last_error.make ("last delete did not work: " + l_base_update.error_message_32)
            end
        end
    end

    require
        valid_item: attached item as l_i and then l_i.valid_for_insert or l_i.valid_for_update

why do I have an unknown identifier here with l_i??!!!

What is the best practice to do something such as

local
    l_pet: ANIMAL
do
    l_pet := catch_it_from_the_sky
    inspect l_pet
    when attached {DOG} l_pet as l_dog
        l_dog.eat (meat)
    when attached {FISH} l_pet as l_fish
        l_fish.eat (plants)
    else
        io.put_string ("Strange animal how do I feed him???")
    end
do

the compiler is complaining with the attached after when...

Update: why such a need?

because it just happened me to mess up with repeated copy-paste which is what a language tries to help avoiding. In the above case, the l_pet is written one time, with a N times if/else I'd have to write it as much times as ifs...

Like double bind in psychology, is there a way to tell that one expression implies another and reversely?

valid_last_error: attached last_error implies not attached last_success_message
valid_last_success_message: attached last_success_message implies not attached last_error

would be something like

valid_last_error: attached last_error double_binded_not attached last_success_message

which could be equivalent to

valid_last_error: attached last_error never_with attached last_success_message

• T stands for True boolean expression
• F for False
• R for Result

implies (a, b: BOOLEAN): BOOLEAN

a b R
T T T
T F F
F T T
F F T

and (a, b: BOOLEAN): BOOLEAN

a b R
T T T
T F F
F T F
F F F

or (a, b: BOOLEAN): BOOLEAN

a b R
T T T
T F T
F T T
F F F

xor (a, b: BOOLEAN): BOOLEAN

a b R
T T F
T F T
F T T
F F F

double_implies (a, b: BOOLEAN): BOOLEAN

a b R
T T F
T F T
F T T
F F T

as a maybe more explaining example (more known) instead of writing

invalid_index_implies_off: index < 1 implies off
off_implies_invalid_index: off implies index < 1

we could write:

index_coherent_with_off: index < 1 never_both off

which would just add a function to BOOLEAN class such as

alias never_with, alias reversible_implies (v: like Current): like Current
    do
        if Current and v then
            Result := False
        else
            Result := True
        end
    end

Hope now everybody got my point... I don't know if there is such arithmetic operator.

We could extend it for a variable number of parameters

I'm looking for escaping characters for a DB query. As it is a complex mechanism (and also could be hacked), it seems obvious I don't want to rewrite it...

I saw that in the JSON_STRING class there is something for escaping STRING_8 (unescape_to_string_8) is there something comparable for DB Strings (need for postgresql)?

Is there a way to restrict the conformance of a type to be a collection of types?

Let me explain by example:

give_foo (garbage: ANY): STRING
    do
        if attached {STRING} garbage as l_s then
            Result := l_s
        elseif attached {INTEGER} garbage as l_int then
            Result := l_int.out
        elseif attached {JSON_OBJECT} garbage as l_int then
            Result := l_int.representation
        elseif attached {RABBIT} garbage as l_animal then
            Result := l_animal.name + l_animal.color
        else
            Result := ""
            check
                unchecked_type_that_compiler_should_be_able_to_check_for_me: False
            end
        end
    end

Couldn't I do something like (like a convert function could do)

give_foo (garbage: {STRING, INTEGER, JSON_OBJECT, RABBIT}): STRING
    do
        if attached {STRING} garbage as l_s then
            Result := l_s
        elseif attached {INTEGER} garbage as l_int then
            Result := l_int.out
        elseif attached {JSON_OBJECT} garbage as l_int then
            Result := l_int.representation
        elseif attached {RABBIT} garbage as l_animal then
            Result := l_animal.name + l_animal.color
        else
            Result := ""
            check
                unchecked_type_that_compiler_should_be_able_to_check_for_me: False
            end
        end
    end

or something like

not_garbage_hash_table: HASH_TABLE[{INTEGER, STRING, ANIMAL}, STRING]