Re: survey: string external representation

[Thien-Thi Nguyen <ttn@gnuvola.org>]

[-- Attachment #1: Type: text/plain, Size: 343 bytes --]

() Thien-Thi Nguyen <ttn@gnuvola.org>
() Fri, 27 Jan 2012 11:27:30 +0100

   The code assumes Guile 2 DTRT [...]

Well, further investigation raises new doubts.  The issue really
is in contiguous mixed raw and \x-escaped octets, and not just
single byte external representation, so here is a followup
experiment that addresses that directly:


[-- Attachment #2: xrep2.scm --]
[-- Type: text/x-scheme, Size: 558 bytes --]

(setlocale LC_ALL "")

(define (hmm symbol)
  (define (show x)
    (display x) (display "\t") (write x) (newline))
  (newline)
  (show symbol)
  (let ((string (symbol->string symbol)))
    (show string)
    (show (object->string string))))

(display "LANG: ") (write (getenv "LANG")) (newline)
(hmm 'foo)
(hmm '#{f\"o b\\r}#)
(hmm '⌬)                                ; U+232C (utf-8: E2 8C AC)
(hmm '䕫)                               ; U+2F9B2 (utf-8: F0 AF A6 B2)
(hmm '蜨)                               ; U+2F9BC (utf-8: F0 AF A6 BC)

[-- Attachment #3: Type: text/plain, Size: 150 bytes --]


Below are the output of two runs:

  guile -s xrep2.scm \
    | tee xrep2-$(guile --version | sed 's/.* //;q')-$LANG.out

What do other people see?


[-- Attachment #4: xrep2-1.4.1.124-it_IT.UTF-8.out --]
[-- Type: text/plain, Size: 276 bytes --]

LANG: "it_IT.UTF-8"

foo	foo
foo	"foo"
"foo"	"\"foo\""

#{f\"o\ b\\r}#	#{f\"o\ b\\r}#
f"o b\r	"f\"o b\\r"
"f\"o b\\r"	"\"f\\\"o b\\\\r\""

⌬	⌬
⌬	"⌬"
"⌬"	"\"⌬\""

䕫	䕫
䕫	"䕫"
"䕫"	"\"䕫\""

蜨	蜨
蜨	"蜨"
"蜨"	"\"蜨\""

[-- Attachment #5: xrep2-1.8.7-it_IT.UTF-8.out --]
[-- Type: text/plain, Size: 316 bytes --]

LANG: "it_IT.UTF-8"

foo	foo
foo	"foo"
"foo"	"\"foo\""

#{\f\\\"o\ b\\\\r}#	#{\f\\\"o\ b\\\\r}#
f\"o b\\r	"f\\\"o b\\\\r"
"f\\\"o b\\\\r"	"\"f\\\\\\\"o b\\\\\\\\r\""

âÐŒ	âÐŒ
âÐŒ	"â\x8c¬"
"â\x8c¬"	"\"â\\x8c¬\""

䕫	䕫
䕫	"䕫"
"䕫"	"\"䕫\""

蜨	蜨
蜨	"蜨"
"蜨"	"\"蜨\""