| 1 | This hack makes Guile default to UTF-8. This avoids calls to
|
| 2 | `iconv_open'; `iconv_open' tries to open shared objects that aren't
|
| 3 | available during bootstrap, so using UTF-8 avoids that (and UTF-8 has
|
| 4 | built-in conversions in glibc, too.)
|
| 5 |
|
| 6 | diff --git a/libguile/bytevectors.c b/libguile/bytevectors.c
|
| 7 | index cf41f2f..facfb91 100644
|
| 8 | |
| 9 | |
| 10 | @@ -1887,7 +1887,7 @@ utf_encoding_name (char *name, size_t utf_width, SCM endianness)
|
| 11 | if (scm_i_is_narrow_string (str)) \
|
| 12 | { \
|
| 13 | err = mem_iconveh (scm_i_string_chars (str), c_strlen, \
|
| 14 | - "ISO-8859-1", c_utf_name, \
|
| 15 | + "UTF-8", c_utf_name, \
|
| 16 | iconveh_question_mark, NULL, \
|
| 17 | &c_utf, &c_utf_len); \
|
| 18 | if (SCM_UNLIKELY (err)) \
|
| 19 | diff --git a/libguile/ports.c b/libguile/ports.c
|
| 20 | index 301bc44..b0ea2e6 100644
|
| 21 | |
| 22 | |
| 23 | @@ -1750,7 +1750,7 @@ scm_ungetc (scm_t_wchar c, SCM port)
|
| 24 | if (pt->encoding != NULL)
|
| 25 | encoding = pt->encoding;
|
| 26 | else
|
| 27 | - encoding = "ISO-8859-1";
|
| 28 | + encoding = "UTF-8";
|
| 29 |
|
| 30 | len = sizeof (result_buf);
|
| 31 | result = u32_conv_to_encoding (encoding,
|
| 32 | @@ -2212,7 +2212,7 @@ scm_i_set_port_encoding_x (SCM port, const char *encoding)
|
| 33 | pt = SCM_PTAB_ENTRY (port);
|
| 34 |
|
| 35 | if (encoding == NULL)
|
| 36 | - encoding = "ISO-8859-1";
|
| 37 | + encoding = "UTF-8";
|
| 38 |
|
| 39 | if (pt->encoding != encoding)
|
| 40 | pt->encoding = scm_gc_strdup (encoding, "port");
|
| 41 | diff --git a/libguile/posix.c b/libguile/posix.c
|
| 42 | index 4f8b8ac..fea7f74 100644
|
| 43 | |
| 44 | |
| 45 | @@ -1740,7 +1740,7 @@ SCM_DEFINE (scm_setlocale, "setlocale", 1, 1, 0,
|
| 46 | SCM_SYSERROR;
|
| 47 | }
|
| 48 |
|
| 49 | - enc = locale_charset ();
|
| 50 | + enc = "UTF-8";
|
| 51 |
|
| 52 | /* Set the default encoding for new ports. */
|
| 53 | scm_i_set_default_port_encoding (enc);
|
| 54 | diff --git a/libguile/script.c b/libguile/script.c
|
| 55 | index 83daf8a..083891e 100644
|
| 56 | |
| 57 | |
| 58 | @@ -387,7 +387,7 @@ locale_arguments_to_string_list (int argc, char **const argv)
|
| 59 | SCM lst;
|
| 60 | const char *encoding;
|
| 61 |
|
| 62 | - encoding = environ_locale_charset ();
|
| 63 | + encoding = "UTF-8";
|
| 64 | for (i = argc - 1, lst = SCM_EOL;
|
| 65 | i >= 0;
|
| 66 | i--)
|
| 67 | diff --git a/libguile/strings.c b/libguile/strings.c
|
| 68 | index 5d0db23..8266247 100644
|
| 69 | |
| 70 | |
| 71 | @@ -1576,7 +1576,7 @@ scm_from_locale_string (const char *str)
|
| 72 | SCM
|
| 73 | scm_from_locale_stringn (const char *str, size_t len)
|
| 74 | {
|
| 75 | - return scm_from_stringn (str, len, locale_charset (),
|
| 76 | + return scm_from_stringn (str, len, "UTF-8",
|
| 77 | scm_i_default_port_conversion_handler ());
|
| 78 | }
|
| 79 |
|
| 80 | @@ -1803,7 +1803,7 @@ char *
|
| 81 | scm_to_locale_stringn (SCM str, size_t *lenp)
|
| 82 | {
|
| 83 | return scm_to_stringn (str, lenp,
|
| 84 | - locale_charset (),
|
| 85 | + "UTF-8",
|
| 86 | scm_i_default_port_conversion_handler ());
|
| 87 | }
|
| 88 |
|
| 89 | @@ -2054,7 +2054,7 @@ scm_to_stringn (SCM str, size_t *lenp, const char *encoding,
|
| 90 | "string contains #\\nul character: ~S",
|
| 91 | scm_list_1 (str));
|
| 92 |
|
| 93 | - if (scm_i_is_narrow_string (str) && (encoding == NULL))
|
| 94 | + if (scm_i_is_narrow_string (str))
|
| 95 | {
|
| 96 | /* If using native Latin-1 encoding, just copy the string
|
| 97 | contents. */
|
| 98 | @@ -2079,11 +2079,11 @@ scm_to_stringn (SCM str, size_t *lenp, const char *encoding,
|
| 99 | len = 0;
|
| 100 | enc = encoding;
|
| 101 | if (enc == NULL)
|
| 102 | - enc = "ISO-8859-1";
|
| 103 | + enc = "UTF-8";
|
| 104 | if (scm_i_is_narrow_string (str))
|
| 105 | {
|
| 106 | ret = mem_iconveh (scm_i_string_chars (str), ilen,
|
| 107 | - "ISO-8859-1", enc,
|
| 108 | + "UTF-8", enc,
|
| 109 | (enum iconv_ilseq_handler) handler, NULL,
|
| 110 | &buf, &len);
|
| 111 |
|
| 112 | |
| 113 | |
| 114 | @@ -2512,7 +2512,7 @@ scm_i_port_iconv_descriptors (SCM port,
|
| 115 | const char *precise_encoding;
|
| 116 |
|
| 117 | if (!pt->encoding)
|
| 118 | - pt->encoding = "ISO-8859-1";
|
| 119 | + pt->encoding = "UTF-8";
|
| 120 |
|
| 121 | /* If the specified encoding is UTF-16 or UTF-32, then make
|
| 122 | that more precise by deciding what byte order to use. */
|