Package: openthesaurus Severity: normal Version: 20110119-3 Tag: patch When building openthesaurus on a system with UTF-8 locale, I get the following error:
--- unzip -o ooo3/Deutscher-Thesaurus.oxt && \ /usr/share/mythes/th_gen_idx.pl -o th_de_DE_v2.idx \ < th_de_DE_v2.dat && \ LANG=C cat th_de_DE_v2.dat | sed -e 's/<DF>/ss/' > th_de_CH_v2.dat; \ /usr/share/mythes/th_gen_idx.pl -o th_de_CH_v2.idx \ < th_de_CH_v2.dat Archive: ooo3/Deutscher-Thesaurus.oxt inflating: th_de_DE_v2.idx inflating: th_de_DE_v2.dat inflating: README.txt inflating: description.xml inflating: Dictionaries.xcu inflating: META-INF/manifest.xml th_de_DE_v2.idx sed: -e expression #1, char 7: unterminated `s' command Use of uninitialized value $encoding in length at /usr/share/mythes/th_gen_idx.pl line 76. Use of uninitialized value $encoding in scalar chomp at /usr/share/mythes/th_gen_idx.pl line 77. Use of uninitialized value $encoding in concatenation (.) or string at /usr/share/mythes/th_gen_idx.pl line 103. --- This leads to a mythes-de-ch package with an empty swiss german thesaurus file. The reason seems to be the debian/rules file which contains an ISO-8859-1 encoded character (see the above sed expression). $ file -i openthesaurus-20110119.orig/debian/rules openthesaurus-20110119.orig/debian/rules: text/plain; charset=iso-8859-1 Escaping this character fixes this issue (see attached patch). Tom
diff -Naurp openthesaurus-20110119.orig/debian/rules openthesaurus-20110119/debian/rules --- openthesaurus-20110119.orig/debian/rules 2011-01-19 22:05:28.000000000 +0100 +++ openthesaurus-20110119/debian/rules 2011-09-23 10:25:01.704519908 +0200 @@ -22,7 +22,7 @@ build-indep-stamp: configure-stamp unzip -o ooo3/Deutscher-Thesaurus.oxt && \ /usr/share/mythes/th_gen_idx.pl -o th_de_DE_v2.idx \ < th_de_DE_v2.dat && \ - LANG=C cat th_de_DE_v2.dat | sed -e 's/ß/ss/' > th_de_CH_v2.dat; \ + LANG=C cat th_de_DE_v2.dat | sed -e 's/\ß/ss/' > th_de_CH_v2.dat; \ /usr/share/mythes/th_gen_idx.pl -o th_de_CH_v2.idx \ < th_de_CH_v2.dat