Errata

The text example given for badmatch error is unrelated. The discussion is about the return tuple from keysearch/3, and what may happen when the key is not present. However, the example code below the discussion is a trivial example of an attempt to assign to a tuple with an element already bound.

1> Tuple = {1, two, 3}.{1,two,3}
2> {1, two, 3, Four} = Tuple.
** exception error: no match of right hand side value {1,two,3}

The lines

7> lists:delete(2,[1,2,3,2,4,2]).
[1,3,2,4,2]

are ambiguous. They seem to indicate that the deletion is occurring on the second element rather than being pattern matched to the integer 2. If this were written as

7> lists:delete(b,[a,b,c]).
[a,c]

or

7> lists:delete(2, [1,1,1,1,1,1,1,1,1,2,1,1,1,1])
[1,1,1,1,1,1,1,1,1,1,1,1,1]

it would be clearer.

Its seems tome that the second clause of mergeR/3 definition breaks the function definition.
calling merge/2 with first list longer by 1 item than the second,
will not raise an exception.

In the text following this code sample it is stated that the exited function checks to see if the frequency has already been placed in the Allocated list because of a possible race condition where the client process exits before the server process finishes deallocating the frequency and unlinking the client process.

Is this correct? I don't believe the possibility of a race condition exists here.

Since the server process is trapping exit signals, the client's exit signal is delivered to the server's mailbox. Since the deallocate function (including the unlink operation) is executed from the message loop in response to a deallocate message, no other messages (including the exit message) will be processed until the deallocation and unlink completes.

While the code should still check the Allocated list, the reason (I believe) is only because of a client that may have neglected to deallocate a frequency before it terminated.

Is there some command or setting I need to change in the console to
> get it to recognize/load the revised db.beam file?
> I added the patches directory to the path as described (the only
> difference being that I added it as a relative path "./patches", but
> that is also how the "ebin" directory is in the path), and then I run
> the convert routine and get an error:
> =ERROR REPORT==== 26-May-2010::11:59:48 ===
> Error in process <0.45.0> with exit value: {undef,[{db,convert,[list,
> [{rec1,"Steve"},{rec2,"Steve2"}]]},{db_server,loop,1}]}
>
> Which is what happens if I call the (non-existent) convert method on
> the old class. :(
>
> TIA
>

"each incarnation of node, references are allocated to be integers starting at 0"

this is unclear. I think several points could clarify:

- references are used like GUID's, can be paired with any erlang entity in the runtime, not just a node

- references are 4 to 7 words, not integers, per

http://erlang.mirror.su.se/doc/efficiency_guide/advanced.html

- The integer starting at 0 is some kind of seed value, not the reference's value.

The form

prettyIndexNext(Entry, {Word, Lines}=IndexEntry) ->

with a match pattern in the function header is nowhere described. This is the first use of it.

debian squeeze
erlang erts-5.8

What can I do?

>microblog:start().

exception error: {load_driver,"symbol _ZN16wxStyledTextCtrl7SendMsgEill, version WXU_2.8 not defined in file libwx_gtk2u_stc-2.8.so.0 with link time reference"}
in function wxe_server:start/0
in call from wx:new/1
in call from microblog:start/

wx#{id=?UNDO....} ->
{StartPos,EndPos} = lastLineRange(Text),
...

lastLineRange(Text) not declared

where does this exist?


Thank you,
Douglas

The text makes a special point of noting:

---
Note how the get request/3 function receives the binary chunks in batches of 100 bytes.
---

Is it really possibly to know that the chunks are 100 bytes long? It's my understanding that when a message is sent through a TCP connection, you have to assume that a message can be split into arbitrarily sized chunks, so each 100 byte message could be split into multiple chunks.

And, when you specify {packet,0}, which I think is the default (it would be better if that was explicit in the code rather than implicit), Erlang isn't going to bundle the received chunks into a complete 100 byte message for you. Now, if the code had specified {packet, 1|2|4}, then I think gen_tcp:recv(Socket, 0, Timeout) *would* return whole, 100 byte messages.

even_multiple(List, Multiple) ->
Even = lists:filter(fun(X) -> (X rem 2) == 0 end, List),
Multiple = lists:map(fun(X) -> X * Multiple end, Even),
lists:sum(Multiple).

Multiple = lists:map(fun(X) -> X * Multiple end, Even) can't be right. I guess it should be

Multiple1 = lists:map(fun(X) -> X * Multiple end, Even), lists:sum(Multiple). (Assume the input parameter "Multiple" is a multiple). But need author's clarification.