summaryrefslogtreecommitdiff
path: root/src/util/string.vala
diff options
context:
space:
mode:
authorJörg Frings-Fürst <debian@jff-webhosting.net>2017-01-21 19:17:10 +0100
committerJörg Frings-Fürst <debian@jff-webhosting.net>2017-01-21 19:17:10 +0100
commita7ff7c49085c320c1e0a2ffb66cc6d283c5acb8e (patch)
treed9f86def771c7845f241ec81219b243c70523c60 /src/util/string.vala
parent66f6a7bd3f7e00022191f16cb0e82a39e7333a76 (diff)
New upstream version 0.25.3upstream/0.25.3
Diffstat (limited to 'src/util/string.vala')
-rw-r--r--src/util/string.vala52
1 files changed, 47 insertions, 5 deletions
diff --git a/src/util/string.vala b/src/util/string.vala
index 7331780..bf7e605 100644
--- a/src/util/string.vala
+++ b/src/util/string.vala
@@ -91,15 +91,57 @@ public enum PrepareInputTextOptions {
DEFAULT = EMPTY_IS_NULL | VALIDATE | INVALID_IS_NULL | STRIP_CRLF | STRIP | NORMALIZE;
}
+private string? guess_convert(string text) {
+ string? output = null;
+ size_t bytes_read = 0;
+ unowned string charset = null;
+ debug ("CONVERT: Text did not validate as UTF-8, trying conversion");
+
+ // Try with locale
+ if (!GLib.get_charset(out charset)) {
+ output = text.locale_to_utf8(text.length, out bytes_read, null, null);
+ if (bytes_read == text.length) {
+ debug ("CONVERT: Locale is not UTF-8, convert from %s", charset);
+ return output;
+ }
+ }
+
+ try {
+ output = GLib.convert (text, text.length, "UTF-8", "WINDOWS-1252", out bytes_read);
+ charset = "WINDOWS-1252";
+ } catch (ConvertError error) {
+ if (error is ConvertError.NO_CONVERSION) {
+ try {
+ output = GLib.convert (text, text.length, "UTF-8", "ISO-8859-1", out bytes_read);
+ charset = "ISO-8859-1";
+ } catch (Error error) { /* do nothing */ }
+ }
+ }
+
+ if (bytes_read == text.length) {
+ debug ("CONVERT: Guessed conversion from %s", charset);
+
+ return output;
+ }
+
+ return null;
+}
+
public string? prepare_input_text(string? text, PrepareInputTextOptions options, int dest_length) {
if (text == null)
return null;
- if ((options & PrepareInputTextOptions.VALIDATE) != 0 && !text.validate())
- return (options & PrepareInputTextOptions.INVALID_IS_NULL) != 0 ? null : "";
-
- string prepped = text;
-
+ string? prepped = text;
+ if (PrepareInputTextOptions.VALIDATE in options) {
+ if (!text.validate()) {
+ prepped = guess_convert (text);
+
+ if (prepped == null) {
+ return (options & PrepareInputTextOptions.INVALID_IS_NULL) != 0 ? null : "";
+ }
+ }
+ }
+
// Using composed form rather than GLib's default (decomposed) as NFC is the preferred form in
// Linux and WWW. More importantly, Pango seems to have serious problems displaying decomposed
// forms of Korean language glyphs (and perhaps others). See: