Richard Heck wrote: > OK. I did a good bit of testing earlier and saw no problems, but I > haven't tested it since updating it to current trunk. It's actually less > complex than it seems. The improvement lies in our storing the data for > the individual BibTeX fields in a std::map rather than one massive > string which we then have to parse over and over---this is all in the > parser in InsetBibtex.cpp. As a result, though, the signature of certain > function calls changes, so there are a lot of very minor changes to > function calls. And we can then simplify a lot else, removing the > complicated code that attempted to recover the year, author, etc, from > the massive string.
I saw it's an improvement. But it frighten me still a bit (for 1.5.1). > In any event, here's an up-to-date patch. Should I commit to trunk at > some point so it can be tested there? Once I've done a little more > testing here? I'd say so, yes. Jürgen
