• Jason den Dulk (11/11) Aug 14 2013 Hello.
• anonymous (5/12) Aug 14 2013 I think you don't need put(char). put(char[]) or
• Jonathan M Davis (16/26) Aug 15 2013 All strings are treated as ranges of dchar when using the range APIs, so...
• monarch_dodra (6/9) Aug 15 2013 Unfortunately, right now, yes. "put" doesn't know how to convert

"Jason den Dulk" <public2 jasondendulk.com> writes:

Hello.

When working with my code I noticed that if I use front on a 
char[], it yields a dchar. Am I correct in concluding that it 
does a UTF-8 to UTF-32 conversion and popFont will skip the whole 
character, not just a code unit?

Also, does this mean that if I'm creating an output range for 
char[], will I need to implement a put(dchar) as well as a 
put(char)?

Thanks
Regards
Jason

"anonymous" <anonymous example.com> writes:

On Thursday, 15 August 2013 at 00:49:00 UTC, Jason den Dulk wrote:
 When working with my code I noticed that if I use front on a 
 char[], it yields a dchar. Am I correct in concluding that it 
 does a UTF-8 to UTF-32 conversion and popFont will skip the 
 whole character, not just a code unit?

yup

 Also, does this mean that if I'm creating an output range for 
 char[], will I need to implement a put(dchar) as well as a 
 put(char)?

I think you don't need put(char). put(char[]) or 
put(const(char)[]) could be worthwhile to prevent decoding. But 
put(dchar) alone would suffice.

Jonathan M Davis <jmdavisProg gmx.com> writes:

On Thursday, August 15, 2013 02:48:58 Jason den Dulk wrote:
 Hello.
 
 When working with my code I noticed that if I use front on a
 char[], it yields a dchar. Am I correct in concluding that it
 does a UTF-8 to UTF-32 conversion and popFont will skip the whole
 character, not just a code unit?
 
 Also, does this mean that if I'm creating an output range for
 char[], will I need to implement a put(dchar) as well as a
 put(char)?

All strings are treated as ranges of dchar when using the range APIs, so you 
pretty much don't do anything with char or wchar where ranges are concerned 
unless you're optimizing a particular function for narrow strings. There is no 
reason to implement put(char), just put(dchar). Range-based code shouldn't 
generally care what type of string it's dealing with, so you wouldn't normally 
be writing any range-based code that cared about char[] unless you're 
optimizing a particular function's implementation (in which case, all of that 
would be internal to the function and wouldn't affect its semantics).

Here are a couple of stackoverflow questions that discuss ranges and strings. 
Perhaps, you'll find them useful.

http://stackoverflow.com/questions/16590650/how-to-read-a-string-character-by-character-as-a-range-in-d

http://stackoverflow.com/questions/12288465/std-algorithm-joinerstring-string-why-result-elements-are-dchar-and-not-ch


- Jonathan M Davis

P.S. I really should finish writing the article that I started explaining 
ranges. So much to do, so little time.

"monarch_dodra" <monarchdodra gmail.com> writes:

On Thursday, 15 August 2013 at 00:49:00 UTC, Jason den Dulk wrote:
 Also, does this mean that if I'm creating an output range for 
 char[], will I need to implement a put(dchar) as well as a 
 put(char)?

Unfortunately, right now, yes. "put" doesn't know how to convert 
on the fly to the right type.

However, I have an open pull request so that anything that 
accepts some form of character, or character string, can be feed 
any form of character, or character stream.