you're right. If you change it, please commit it to github. however, imo, getting everyone to use the function GetFactorDelimiter() is difficult to enforce. In the decoder, and impossible in the scripts.
The use of | is too entrenched. Therefore, I think it's better to delete that function, rather than let it continue to be a source of bugs. On 04/08/2012 20:53, Marcin Junczys-Dowmunt wrote: > Hi, > I just found a possible bug in moses/src/Word.cpp. > In the function Word::CreateFromString you have the line > > Tokenize(wordVec, str, "|"); > > shouldn't that be rather something like: > > const std::string& factorDelimiter = > StaticData::Instance().GetFactorDelimiter(); > TokenizeMultiCharSeparator(wordVec, str , factorDelimiter); > > The hardcoded delimiter caused me quite some headache. > > Or is it the other way round, that the phrase table should provide only > factors delimited with "|" ? > > Best, > Marcin > _______________________________________________ > Moses-support mailing list > [email protected] > http://mailman.mit.edu/mailman/listinfo/moses-support > _______________________________________________ Moses-support mailing list [email protected] http://mailman.mit.edu/mailman/listinfo/moses-support
