Forum

    Discussion regarding the Eiffel programming language and various supporting tools.
  • Nov 30
    Re: [EXT] Re: [eiffel-users] Idiomatic Eiffel way for class variables
    Sure, sure -- that is understandable because the binary is running in a single process-thread, so there's that. What I mean is--I have never directly used it because of doing Eiffel-based multi-threading with the thread library and so on. So, my statement still stands, but your point is well
  • Nov 30
    Re: [eiffel-users] AWS ECS & Docker Compose
    Hi Chris! Yes--I had come across that link. It provides a lot of detail without any context from which to understand how to apply it. This morning, I have located a YouTube video that does precisely what I needed as a starting point. It is a "deep dive" into not only setting up a Docker
  • Nov 30
    Re: [EXT] Re: [eiffel-users] Idiomatic Eiffel way for class variables
    once means once ("THREAD") so you have certainly used it. On Mon, 30 Nov 2020 at 12:32, Larry Rix wrote: > I have never used the once ("THREAD"), but I have certainly used both once > and once ("OBJECT"). These are as you have noted: once = system-wide and > once
  • Nov 30
    Re: [EXT] Re: [eiffel-users] Idiomatic Eiffel way for class variables
    I have never used the once ("THREAD"), but I have certainly used both once and once ("OBJECT"). These are as you have noted: once = system-wide and once ("OBJECT") = per created instance. While I may have ultimately found a use-case for once ("THREAD"), I think this has now been relegated to the
  • Nov 30
    Antw: [EXT] Re: [eiffel-users] Idiomatic Eiffel way for class variables
    Hi! Talking about it: I never found the clear reference that explains whether "once" has "object scope" (only called once for the the particular object) or "system scope" (called only once for the system). ETL3 says: "You may also specify “once” to mean “once per thread” or “once per
  • Nov 29
    Re: [eiffel-users] AWS ECS & Docker Compose
    Perhaps https://docs.docker.com/compose/compose-file/#versioning ? On Mon, Nov 30, 2020 at 1:47 AM Larry Rix wrote: > Does anyone here know how to use "docker compose"--specifically--how to > set up and use the YAML files and to do this against AWS ECS? > > I am
  • Nov 29
    Re: [eiffel-users] Idiomatic Eiffel way for class variables
    i would suggest using something like this class EVENT feature -- Basic operations notify do Counter.set_item (Counter.item + 1) end feature {NONE} -- Constants Counter: INTEGER_32_REF once create Result end end For a thread safe counter I
  • Nov 29
    AWS ECS & Docker Compose
    Does anyone here know how to use "docker compose"--specifically--how to set up and use the YAML files and to do this against AWS ECS? I am presently reading through the following website, but I am not quite understanding how to construct the YAML file within the context of AWS to get my Docker
  • Nov 28
    one more thing (p n d in model worlds)
    Just one more thing. For those interested, you can get the current pointer (mouse) position of a drop event by querying the world's enclosing cell (EV_MODEL_WORLD_CELL) from your drop event handler. txy := world_cell.pointer_position This opens up a bunch of possibilities for moving stuff
  • Nov 28
    RE: [eiffel-users] Re: pick and drop in model world
    Hi Larry (et al) As it happens, model worlds behave correctly w/r/t pick-and-drop (even if some programmers don't). Just follow the directions, and don't try to overthink it too much. Just sayin'. The working setup is very simple in the end. For each model in the model world, call set_pebble
  • Nov 28
    RE: [eiffel-users] Re: pick and drop in model world
    Hi Larry My initial roadblock could be attributed, it seems, to a hardware problem (the nut behind the keyboard). With a bit more instrumentation, chasing the hungry hungry hippos of event land, I did find some event eating, BUT, I found also that model world was more cooperative than my previous
  • Nov 28
    Re: [eiffel-users] Re: pick and drop in model world
    As I recall, we encountered this with scroll-bars where we wanted a typical scroll-wheel behaviour, but it turned out that the wheel event was not getting passed to the bar, but was being intercepted early and silently swallowed higher up. Eiffel Software (Manu, specifically as I recall) handled
  • Nov 28
    RE: [eiffel-users] Re: pick and drop in model world
    Hi Larry I added drop event handlers to every class I could find (well, almost). Model world exists within a model world cell, model world cell has a drawing area. They never fired. I even tried an individual model and, while the picked just fine, they too would not respond to the a drop event.
  • Nov 28
    Re: pick and drop in model world
    Is it because a higher level class is swallowing the event? For example--model-world is getting the drop-event and doing nothing with it instead of passing it down to items it contains? Is it possible that an even higher level class than model-world is swallowing the event? Perhaps even a
  • Nov 27
    pick and drop in model world
    Hi All I decided to see what I could do about using pick-and-drop in the context of an EV_MODEL_WORLD. While I can set the pebble to a model in the world, I can never get a drop event from within the world. I tried giving the world and its enclosing/associated world_cell, and even the world_cell
  • Nov 27
    RE: [eiffel-users] Idiomatic Eiffel way for class variables
    There is also non-object call: {TYPE_NAME}.feature_name (possibly_with_arguments) where feature_name denotes a feature (constant attribute or routine, including once function as mentioned by others) that does not use the state. This scheme covers class methods as well as
  • Nov 27
    Re: [eiffel-users] Idiomatic Eiffel way for class variables
    Thanks all, that is most enlightening. Seamus On Friday, 27 November 2020 at 01:14:33 UTC rfo wrote: > If the counter is an integer, then a once function will always return the > value from the first invocation. > A shared integer value that can change can be done by using a CELL > [INTEGER],
  • Nov 27
    RE: [eiffel-users] Idiomatic Eiffel way for class variables
    If the counter is an integer, then a once function will always return the value from the first invocation. A shared integer value that can change can be done by using a CELL [INTEGER], where the instance of CELL is the once function. counter: INTEGER do Result := counter_cell.item end
  • Nov 27
    Re: [eiffel-users] Idiomatic Eiffel way for class variables
    https://www.eiffel.org/doc/eiffel/ET-_Once_routines_and_shared_objects -- ********************************************** Philippe Gachoud Puerto Williams 6657 Las Condes Santiago de Chile +56 934022210 ph.g...@gmail.com ********************************************** On Thu, Nov 26, 2020,
  • Nov 26
    Re: [eiffel-users] Idiomatic Eiffel way for class variables
    hope it helps covering your needs... from memory some_attribute: STRING once Result := "Baz" end my_class_function (a_bar: STRING): FOO do create Result.make(a_bar) ensure instance_free: class end in another class l_foo_instance := {FOO}.my_class_fun
  • Nov 26
    Idiomatic Eiffel way for class variables
    Hi folks As a hoary old ex-Smalltalker, I am aware of the distinction between class side variable and instance variables. Class side variables (statics in Java) can be very useful for information pertaining to the entire population of all instances of a class. I have a class called EVENT which
  • Nov 24
    Re: [eiffel-users] cloud services: amazon vs google
    I am presently working with the Eiffel team to form up some automated ways of putting Eiffel binary executables directly to various cloud services. For example, I reached out to Cloudflare, who responded to me by phone last night. While this is very early in the process, I am hopeful that we can
  • Nov 23
    cloud services: amazon vs google
    Hi, does anybody have experience with cloud services and eiffel? I'm evaluating the option of having hosted our - postgresql DB - eiffel web server - angular web application Could you talk about your experience with any of google/amazon? microsoft is out of the game for us but for the
  • Nov 23
    Re: [EXT] [eiffel-users] JSON_Ext Library
    The JSON-ext library was built years ago in Atlanta. I was not the originator of it, but have become the curator of it. As I recall, the library was built as a higher-level library to help simplify the JSON library. This does not mean that the JSON library does not already facilitate these
  • Nov 23
    Antw: [EXT] [eiffel-users] first step to 20.05
    schrieb am 21.11.2020 um 11:36 in Nachricht <20201121113...@neuchatel.abstraction.ch>: > Some comments on version 20.05 > > 1) At installation time the restoration of previous installed version fail. > It suggested the 16..., I suppled 19.05 but that does works > >
  • Nov 23
    Antw: [EXT] [eiffel-users] JSON_Ext Library
    Hi! Actually I think a JSON method should be in a COLLECTION as ARRAYS and HASH_TABLES both sould be supported. Or maybe have a class JSONIZER. Regards, Ulrich >>> Larry Rix schrieb am 20.11.2020 um 17:18 in Nachricht <575640e0-0e3e-4305-939c-df0f5e1ba27bn@goo
  • Nov 22
    Re: Eiffel and microcontrollers
    No offense taken! :-) The WiringPi library was abandoned by its maker because people started making rather abusive requests of him and he got tired of the abuses (I don't blame him). So, be aware that consumers of your library have a human side and they might say and do things with a cranky
  • Nov 21
    first step to 20.05
    Some comments on version 20.05 1) At installation time the restoration of previous installed version fail. It suggested the 16..., I suppled 19.05 but that does works 2)Following the conversion I got some failures. One of then: check attached system_preferenc
  • Nov 21
    Re: [eiffel-users] Re: Eiffel and microcontrollers
    Raspberry pi is already supported, see the armv7 platform. See https://ftp.eiffel.com/pub/download/20.05/Eiffel_20.05_rev_104521-linux-armv7.tar.bz2 If someone needs armv6, please let ISE know. Le ven. 20 nov. 2020 à 04:48, jjj a écrit : > Ian, > Who is "we" in "we may help"? >
  • Nov 20
    Re: Eiffel and microcontrollers
    Great, was not sure. > What about Raspberry Pi? >> > Larry Rix has been working on it. Check the following repository > https://github.com/ljr1981/WiringPi-Eiffe > Yes, I am aware, but it wraps a [IMO] long, hard-to-understand, no longer maintained, C-library, while maintaining the structure
  • See more ...