Is there a way to know line separator for the current os?

I did not find a way to get line separator in the system independent
manner. Linux uses character 10 (lf), windowses (two characters: 10,
13), and mac uses 13(not sure?).
Is there already a way to do this? If not I can open a issue and propose a fix.

--
Panu

On Tue, Mar 29, 2011 at 09:08:33AM +0300, Panu Suominen wrote:
> I did not find a way to get line separator in the system independent
> manner. Linux uses character 10 (lf), windowses (two characters: 10,
> 13), and mac uses 13(not sure?).
> Is there already a way to do this? If not I can open a issue and propose a fix.

FileDirectory class>>slash

On Mar 29, 2011, at 8:08 AM, Panu Suominen wrote:

> I did not find a way to get line separator in the system independent
> manner. Linux uses character 10 (lf), windowses (two characters: 10,
> 13), and mac uses 13(not sure?).
> Is there already a way to do this? If not I can open a issue and propose a fix.


There is #guessDefaultLineEndConvention on MultiByteFileStream


--
Marcus Denker  -- http://www.marcusdenker.de
INRIA Lille -- Nord Europe. Team RMoD.

2011/3/30 Marcus Denker <[hidden email]>:
Thanx. This seems to do the trick.


--
Panu

I've spend the better time of the afternoon with trying metaprogramming
with Smalltalk I finally have come up with:

asciiNames := Array withAll: #(#nul #soh #stx #etx #eot #enq #ack #bel #bs #tab #lf).
        i := 1.
        asciiNames
                do: [ :each |
                        self class compile: each asString , '  ^ asciiField at: ', i asString.
                        i := i + 1 ].


which at least does  what I want given names to ASCI Codes. Now
I have to admit I dislike this stuff with putting a string togehter and
run compile: on this string.

I'm sure there is a more elegant way. Would you mind to lend me a hand?

Yes and because I think I will need much more code generation in the
near future. Does you have nice Tutorial, Book whatever to do such kind
of jobs in Smalltalk I've quite a lot of books about Smalltalk but none
has this as theme... So it's really a bit hard to get into it....

Regards
Friedrich


--
Q-Software Solutions GmbH; Sitz: Bruchsal; Registergericht: Mannheim
Registriernummer: HRB232138; Geschaeftsfuehrer: Friedrich Dominicus

Hi Friederich, here's two simple optimizations:

Hi Friedrich,

Am 31.03.2011 um 18:09 schrieb Friedrich Dominicus:

if you don't like string concatenation you may find this useful

#(#nul #soh #stx #etx #eot #enq #ack #bel #bs #tab #lf)
   withIndexDo:  [:each :index |
        self class compile: ('<1s><n><t>^ asciiField at: <2s>' expandMacrosWith: each with: index printString)
                classified: #(#accessing)]



Well, I think you have the perfect source for learning right in front of you. The code snippet above didn't come from my mind. I just remembered that if you are on a class you can open the menu, open 'Refactor Class' and select Accessors to have pharo generate getters and setters for you. So I wrote "Accessors" (with the double quotes) in the workspace selected it and did a string seach (shortcurt E). From the choice given 'ORCmdAccessorClassRefactoring' seemed to be the most feasible. Then I investigated the methods found some other reference and so on (I leave this as an exercise). Finally I found a snippet which I copied and altered for your needs. This way I learned something about OmniBrowser, Refactoring-Engine and that there is a method expandMacrosWith: :)
I think this is the way to go. Often it is hard if you first have to think about a problem that you can solve. But you have a problem so with the very good searching capabilities of pharo you can find such things easily.

Hope this helps,

Norbert

AxiNat <[hidden email]> writes:

> Hi Friederich, here's two simple optimizations:
>
> #(#nul #soh #stx #etx #eot #enq #ack #bel #bs #tab #lf)
>    withIndexDo:  [:each :index |
>       self class compile: each asString , '  ^ asciiField at: ', index
> asString.].
>
> Array withAll: #(1 2 3) is equivalent to #(1 2 3), and if you're not going to
> use the array again, you don't even need to assign it to a variable :)
Well I do not want the compile part with String

>
> Instead of manually using an index, it is more elegant to use
> withIndexDo:
Ok that is really  nice to know

Regards
Friedrich

--
Q-Software Solutions GmbH; Sitz: Bruchsal; Registergericht: Mannheim
Registriernummer: HRB232138; Geschaeftsfuehrer: Friedrich Dominicus

On Apr 1, 2011, at 10:10 AM, Friedrich Dominicus wrote:


So you could override #doesNotUnderstand: and there check if someone
sends a message #etx (and the others) and then return the value. (you then want to
store the value in a dictionary keyed by the the message name).

doesNotUnderstand: aMessage

        | dict |
        dict := (#(nul soh stx etx eot enq ack bel bs tab lf) collectWithIndex: [:elem :index |
                elem -> index]) asDictionary.
        ^dict at: aMessage selector ifAbsent: [^super doesNotUnderstand: aMessage].

But of course doesNotUnderstand: is always a bit magic... and should be used with care.

As for strings stiched together: yes, *horrible*. And we plan to move on to something
better soon.

In general, I like the idea to have a reflective model of  the structure of methods...

When talking about code, we should not use string... the Quasi-Quoting from
Helvetia is far better than string manipulation:

        http://scg.unibe.ch/research/helvetia/examples#168249369
 
That combined with a good AST should make these things much easier.

        Marcus

--
Marcus Denker  -- http://www.marcusdenker.de
INRIA Lille -- Nord Europe. Team RMoD.

Norbert Hartl <[hidden email]> writes:
Well I'm old fashioned, I like to look into books and see somethign
step-by-step. Unfortunatly the most Smalltalk books are lacking in that
area (for my taste)

Regards
Friedrich


--
Q-Software Solutions GmbH; Sitz: Bruchsal; Registergericht: Mannheim
Registriernummer: HRB232138; Geschaeftsfuehrer: Friedrich Dominicus