| Andrew Cooke | Contents | Latest | RSS | Twitter | Previous | Next

C[omp]ute

Welcome to my blog, which was once a mailing list of the same name and is still generated by mail. Please reply via the "comment" links.

Always interested in offers/projects/new ideas. Eclectic experience in fields like: numerical computing; Python web; Java enterprise; functional languages; GPGPU; SQL databases; etc. Based in Santiago, Chile; telecommute worldwide. CV; email.

Personal Projects

Lepl parser for Python.

Colorless Green.

Photography around Santiago.

SVG experiment.

Professional Portfolio

Calibration of seismometers.

Data access via web services.

Cache rewrite.

Extending OpenSSH.

Last 100 entries

Santiago 30m Bike Route; Mapa de Ciclovias en Santiago; How Unreliable is UDP?; SE Santiago 20m Bike Route; Cameron's Rap; Configuring libxml with Eclipse; Reducing Combinatorial Complexity With Occam - AI; Sentidos Comunes (Chilean Online Magazine); Hilary Mantel: The Assassination of Margaret Thatcher - August 6th 1983; NSA Interceptng Gmail During Delivery; General IIR Filters; What's happening with Scala?; Interesting (But Largely Illegible) Typeface; Retiring Essentialism; Poorest in UK, Poorest in N Europe; I Want To Be A Redneck!; Reverse Racism; The Lost Art Of Nomography; IBM Data Center (Photo); Interesting Account Of Gamma Hack; The Most Interesting Audiophile In The World; How did the first world war actually end?; Ky - Restaurant Santiago; The Black Dork Lives!; The UN Requires Unaninmous Decisions; LPIR - Steganography in Practice; How I Am 6; Clear Explanation of Verizon / Level 3 / Netflix; Teenage Girls; Formalising NSA Attacks; Switching Brakes (Tektro Hydraulic); Naim NAP 100 (Power Amp); AKG 550 First Impressions; Facebook manipulates emotions (no really); Map Reduce "No Longer Used" At Google; Removing RAID metadata; New Bike (Good Bike Shop, Santiago Chile); Removing APE Tags in Linux; Compiling Python 3.0 With GCC 4.8; Maven is Amazing; Generating Docs from a GitHub Wiki; Modular Shelves; Bash Best Practices; Good Emergency Gasfiter (Santiago, Chile); Readings in Recent Architecture; Roger Casement; Integrated Information Theory (Or Not); Possibly undefined macro AC_ENABLE_SHARED; Update on Charges; Sunburst Visualisation; Spectral Embeddings (Distances -> Coordinates); Introduction to Causality; Filtering To Help Colour-Blindness; ASUS 1015E-DS02 Too; Ready Player One; Writing Clear, Fast Julia Code; List of LatAm Novels; Running (for women); Building a Jenkins Plugin and a Jar (for Command Line use); Headphone Test Recordings; Causal Consistency; The Quest for Randomness; Chat Wars; Real-life Financial Co Without ACID Database...; Flexible Muscle-Based Locomotion for Bipedal Creatures; SQL Performance Explained; The Little Manual of API Design; Multiple Word Sizes; CRC - Next Steps; FizzBuzz; Update on CRCs; Decent Links / Discussion Community; Automated Reasoning About LLVM Optimizations and Undefined Behavior; A Painless Guide To CRC Error Detection Algorithms; Tests in Julia; Dave Eggers: what's so funny about peace, love and Starship?; Cello - High Level C Programming; autoreconf needs tar; Will Self Goes To Heathrow; Top 5 BioInformatics Papers; Vasovagal Response; Good Food in Vina; Chilean Drug Criminals Use Subsitution Cipher; Adrenaline; Stiglitz on the Impact of Technology; Why Not; How I Am 5; Lenovo X240 OpenSuse 13.1; NSA and GCHQ - Psychological Trolls; Finite Fields in Julia (Defining Your Own Number Type); Julian Assange; Starting Qemu on OpenSuse; Noisy GAs/TMs; Venezuela; Reinstalling GRUB with EFI; Instructions For Disabling KDE Indexing; Evolving Speakers; Changing Salt Size in Simple Crypt 3.0.0; Logarithmic Map (Moved); More Info; Words Found in Voynich Manuscript

© 2006-2013 Andrew Cooke (site) / post authors (content).

Erlang - Processes, Objects, Protocols

From: "andrew cooke" <andrew@...>

Date: Wed, 2 May 2007 22:13:06 -0400 (CLT)

I was going to write a long post about this code, but I really don't have
the time.  So it's just here in case anyone feels like reading it.  The
first section defines a protocol and supporting functions for chaining
data through different processes (the "unix pipe" idea).  The protocol is
minimal - just data or exit - but each process can have some persistent
state that is preserved across calls.

I'm particularly proud of the formatting here :o)  Took me some time to
work out what is (and isn't) possible with the syntax, scoping, etc.


forward(none, Message) -> Message;
forward(Next, Message) -> Next(Message).

new(Response, Filter, Name, Next) ->

    DoLog =
        fun(Template, Params) -> res_info(Response, Template, Params) end,


    % when the Filter is called with a {data, State, Data} tuple it
    % should return one of:
    % - {data, Data2} containing processed data
    % - {state, State2} containing new state
    % - {both, State2, Data2} containing new state and data
    % new data (Data2) are forwarded to the next filter in the chain,
    % if defined (it may be "none") even if identical to old values.
    % to avoid forwarding anything to the next filter, return state
    % alone.

    DoData =
        fun(Driver, _State, {both, State2, Data2}) ->
                forward(Next, {data, Data2}),
                Driver(Driver, State2);
           (Driver, State, {data, Data2}) ->
                forward(Next, {data, Data2}),
                Driver(Driver, State);
           (Driver, _State, {state, State2}) ->
                Driver(Driver, State2);
           (_Driver, _State, Other) ->
                DoLog("Unexpected result ~p in ~p", [Other, Name])
        end,


    % when the Filter is called with the {exit, State} tuple it should
    % return one of:
    % - {data, Data} containing final data to be forwarded
    % - exit
    % if data are returned they are passed to the next filter (if
    % defined), after which, "exit" is forwarded.

    DoExit =
        fun({data, Data}) ->
                forward(Next, {data, Data}),
                Next(exit);
           (exit) ->
                Next(exit);
           (Other) ->
                DoLog("Unexpected result ~p in ~p", [Other, Name])
        end,


    % driver loop.

    Driver =
        fun(Driver, State) ->
                receive
                    {data, Data} ->
                        DoData(Driver, State, Filter({data, State, Data}));
                    exit ->
                        DoExit(Filter({exit, State}));
                    Other ->
                        DoLog("Unexpected message ~p in ~p", [Other, Name])
                end
        end,

    Pid = spawn(fun() -> Driver(Driver, Filter(start)) end),
    fun(Message) -> Pid ! Message end.


Next, some example processes.  First, a simple one that groups the data by
lines (ended by newline character).  Note the exit strategy handles a
missing final newline.

lines(start) -> [];
lines({data, Partial, Data}) ->
    {Partial2, Lines} = collect(Data, Partial, []),
    {both, Partial2, Lines};
lines({exit, Partial}) -> {data, [lists:reverse(Partial)]}.

collect([], Partial, Lines) -> {Partial, lists:reverse(Lines)};
collect([$\n|Data], Partial, Lines) ->
    Line = lists:reverse([$\n|Partial]),
    collect(Data, [], [Line|Lines]);
collect([C|Data], Partial, Lines) ->
    collect(Data, [C|Partial], Lines).


And second, a process that "gunzips" a stream.  As far as I can tell, the
Erlang zlib library doesn't (directly) support streaming gunzip, so this
accumulates the data beforehand.

unzip(start) -> [];
unzip({data, Acc, Data}) -> {state, [Data|Acc]};
unzip({exit, Acc}) ->
    Compressed = list_to_binary(lists:reverse(Acc)),
    Decompressed = binary_to_list(zlib:gunzip(Compressed)),
    {data, Decompressed}.


I was worried that this buffering would make subsequent chained processes
inefficient (processors waiting on data), so here's something that
fragments the input to a process (and shows that wrapping these processes
as functions pays off).

shape_before(Fun, ChunkLimit) ->
    fun({data, Data}) -> fragment(Fun, ChunkLimit, Data);
       (Message) -> Fun(Message)
    end.

fragment(_Fun, _Chunklimit, []) -> ok;
fragment(Fun, Chunklimit, Data) when length(Data) > Chunklimit ->
    {Chunk, Data2} = lists:split(Chunklimit, Data),
    Fun({data, Chunk}),
    fragment(Fun, Chunklimit, Data2);
fragment(Fun, _Chunklimit, Data) ->
    Fun({data, Data}).


Finally, this is construction of a chain of processes, with fragmentation
and optional compression:

close(Response) ->
    res_info(Response, "Closing"),
    Channel = response:channel(Response),
    socket:close(channel:browser(Channel), channel:proxy(Channel)).

build_chain(Response) ->
    Echo = chain:link(Response, chain:echo(Response), "Echo"),
    Links = chain:new(Response, fun chain_links:links/1, "Links", Echo),
    Lines = chain:new(Response, fun chain_lines:lines/1, "Lines", Links),
    res_info(Response, "Encoding ~p", [response:encoding(Response)]),
    case response:encoding(Response) of
        unknown -> Lines;
        "gzip" -> chain:new(Response, fun chain_unzip:unzip/1, "Gzip",
                            chain:shape_before(Lines, 2000));
        Encoding -> {error, Encoding}
    end.

The processes are in reverse order, so this (optionally) unzips, splits
into lines, searches for hypertext limits, and echoes the result to the
screen.  As you might have guessed, I'm processing HTTP data flows.

Andrew

Re: Objects

From: "andrew cooke" <andrew@...>

Date: Wed, 2 May 2007 22:16:07 -0400 (CLT)

Ooops - see that I left "Objects" in the title from earlier.  I was
originally going to talk about persistent state and the similarities
between method calls and dispatching on tuple tags.  But I'm sure you can
see that anyway :o)

Andrew

First bug

From: "andrew cooke" <andrew@...>

Date: Wed, 2 May 2007 22:22:29 -0400 (CLT)

Staring at that, I've already seen one bug.  No prizes!

Andrew

Comment on this post