Some distros have html2text patches that may generate non-ASCII output even when -ascii is used. This patch adds another case (seen in Fedora) where HTML entity   (non-breaking space) is converted into a multibyte whitespace. Also add a sanity check to make sure non-ASCII text is not introduced in lib/bluetooth.c. --- tools/update_compids.sh | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-) diff --git a/tools/update_compids.sh b/tools/update_compids.sh index 332fb16..38d1fff 100755 --- a/tools/update_compids.sh +++ b/tools/update_compids.sh @@ -21,13 +21,16 @@ cd $tmpdir path=en-us/specification/assigned-numbers/company-identifiers # Use "iconv -c" to strip unwanted unicode characters -# Also strip <input> tags of type checkbox because html2text generates UTF-8 -# for them in some distros even when using -ascii (e.g. Fedora 18) +# Fixups: +# - strip <input> tags of type "checkbox" because html2text generates UTF-8 for +# them in some distros even when using -ascii (e.g. Fedora) +# - replace " " (non-breaking space) with whitespace manually, because +# some versions incorrectly convert it into "\xC2\xA0" curl https://www.bluetooth.org/$path | iconv -c -f utf8 -t ascii | \ - sed '/<input.*type="checkbox"/d' | \ + sed '/<input.*type="checkbox"/d; s/ / /g' | \ html2text -ascii -o identifiers.txt >/dev/null -# Some versions of html2text do not replace & (e.g. Fedora 18) +# Some versions of html2text do not replace & (e.g. Fedora) sed -i 's/&/\&/g' identifiers.txt sed -n '/^const char \*bt_compidtostr(int compid)/,/^}/p' \ @@ -41,6 +44,12 @@ if ! grep -q "return \"" new.c; then echo "ERROR: could not parse company IDs from bluetooth.org" >&2 exit 1 fi +if [ -n "$(tr -d '[:print:]\t\n' < new.c)" ]; then + echo -n "ERROR: invalid non-ASCII characters found while parsing" >&2 + echo -n " company IDs. Please identify offending sequence and fix" >&2 + echo " tools/update_compids.sh accordingly." >&2 + exit 1 +fi echo -e '\tcase 65535:\n\t\treturn "internal use";' >> new.c echo -e '\tdefault:\n\t\treturn "not assigned";\n\t}\n}' >> new.c -- 1.7.9.5 -- To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
