On Thursday 26 April 2007 10:56:25 pm Uwe Stöhr wrote: > > Yes, it looks sane. I can make some remarks about the code but they are > > minor: > > > > - You initialise both h and k outside the loop, there is no need, you > > can remove that definition and the code will work the same. > > I removed these initialisations.
Good. > > - The code is inefficient, it will run over all lines even if you don't > > have any table in the document. Again this can be changed later if the > > profiler shows that there is a significant lag do to this code. > > How could this be done better? Instead of trying to search every line we are only interested to search inside the tabular inset. For that we need a way to determine each table begin and end, and then we run each search in this range. But as I said this is an improvement and it only worth if the code proves to be very slow. > > So regarding the lyx2lyx part the code is OK, so this can go in. > > It's in: > http://www.lyx.org/trac/changeset/18052 > > regards Uwe -- José Abílio