Joachim Durchholz wrote:
> Am Freitag, den 15.02.2008, 06:10 +0000 schrieb Jon Harrop:
>> Joachim Durchholz wrote:
>> > Just to complement Jon's view: I don't think there are any *serious*
>> > problems with recovering indented code from web pages or mail quotes.
>> > Just make sure that whitespace is properly handed through, that's
all,
>> > and it's not even difficult. Hey, Evolution will even strip the angle
>> > brackets when copying from quoted mail ;-)
>>
>> Ugh. I was bitten by this *again* yesterday, trying to cut and paste
some
>> F# code from a blog article and Windows removed all "\n"s.
>
> Strange - I wouldn't have expected to lose line breaks when copying from
> a web page. Possibly it was formatted using <br />, though even then,
> the line breaks should be preserved. This sounds like something awfully
> braindamaged happening in the database->blog->server->browser chain.
>
> If code is represented using <pre>, you should get everything (including
> indentation) faithfully preserved on copy&paste. I could imagine that
> tab-based indentation could still break in some cir***stances, but I
> wouldn't recommend using that anyway.
The code was color syntax highlighted:
http://jyliao.blogspot.com/2008/02/learning-wpf-with-f-toolbars-and-status.html
>> In the end I used
>> my Linux box to cut and paste into a text file that I loaded into the
>> Windows box using putty and then cut and pasted that (correctly
>> formatted) into Visual Studio...
>
> Sounds like brain damage in your browser.
MSIE 7.
> Copying from the source might have worked, too.
The source HTML looks like:
<span style="color:#008000;">//</span><br><span style="color:#008000;">//
Defining drop function...couldn't find a Seq equivalent</span><br><span
style="color:#008000;">// I see Seq.take and Seq.truncate but no
Seq.drop</span><br><span style="color:#0000FF;">let</span><span
style="color:#;"> </span><span style="color:#0000FF;">rec</span><span
style="color:#;"> seqdrop n (list:seq<'a>) =</span><br><span
style="color:#;"> </span><span style="color:#0000FF;">if</span><span
style="color:#;"> n<=0 </span><span
style="color:#0000FF;">then</span><span style="color:#;">
list</span><br><span style="color:#;"> </span><span
style="color:#0000FF;">else</span><br><span style="color:#;">
</span><span style="color:#0000FF;">match</span><span style="color:#;">
(List.of_seq list) </span><span style="color:#0000FF;">with</span><span
style="color:#;"> </span><br><span style="color:#;"> | _ :: []
</span>
....
>> > The argument for indentation is that you indent your code anyway, so
>> > why specify nesting with both keywords and indentation? That's just
>> > silly.
>>
>> Turn autoindenting on!
>
> I tend to never get the indenting style that I want. For some reason,
> most IDEs insist on stuff like
> if (...)
> {
> ...
> }
> where aesthetics wins over vertical space.
The OCaml autoindenter is great, IMHO. Hit ALT+Q and it renindents the
current block of code. I'd much rather have that than a whitespace-
sensitive syntax.
>> > Besides, keywords eat vertical screen estate (if only for the closing
>> > brace or the "end" keyword), and that's a real restriction: the more
>> > code you can see on a screenful, the easier it is to put its
>> > constituents into proper perspective.
>>
>> True. I think the IDE should compact those keywords though.
>
> Probably there should be alternate syntax with and without block-end
> tokens, with means to automatically convert between them (initially,
> that would probably be a function of the compiler, later, IDEs would
> catch on and convert on the fly at the flick of an option switch).
I like Torben's idea of making "end" optional. I think that would work
well...
>> Same goes
>> for "+." in OCaml, which should by typeset as "+" in an IDE.
>
> I don't think that should be done. I'd want to see what addition is
> meant, particularly if the code is polymorphic (unless it's impossible
> to have Integer/Real polymorphism in OCaml).
You can copy the ad-hoc polymorphism of SML but its useless because it
can't
be extended to float32, vector, matrix, quaternions etc.
Regardless, type throwback already solved the problem you're referring to.
F# provides all of these benefits and it works beautifully, a dream come
true!
--
Dr Jon D Harrop, Flying Frog Consultancy Ltd.
http://www.ffconsultancy.com/products/?u
|
