diff options
author | Chris Dyer <cdyer@cs.cmu.edu> | 2011-09-12 19:22:59 +0100 |
---|---|---|
committer | Chris Dyer <cdyer@cs.cmu.edu> | 2011-09-12 19:22:59 +0100 |
commit | f826e90538dc15fbde4152b0e2f7d30fd6e56784 (patch) | |
tree | 3bca97b634b637c79241466a39abcfea2ded22ea /utils | |
parent | 4b6222733c4f9dcb3516bcc64394fa8b4716ce48 (diff) |
source syntax features ~ blunsom emnlp 2008
Diffstat (limited to 'utils')
-rw-r--r-- | utils/stringlib.cc | 7 |
1 files changed, 6 insertions, 1 deletions
diff --git a/utils/stringlib.cc b/utils/stringlib.cc index 7aaee9f0..ade02ca9 100644 --- a/utils/stringlib.cc +++ b/utils/stringlib.cc @@ -32,7 +32,12 @@ void ParseTranslatorInput(const string& line, string* input, string* ref) { void ProcessAndStripSGML(string* pline, map<string, string>* out) { map<string, string>& meta = *out; string& line = *pline; - string lline = LowercaseString(line); + string lline = *pline; + if (lline.find("<SEG")==0 || lline.find("<Seg")==0) { + cerr << "Segment tags <seg> must be lowercase!\n"; + cerr << " " << *pline << endl; + abort(); + } if (lline.find("<seg")!=0) return; size_t close = lline.find(">"); if (close == string::npos) return; // error |