[PUSHED][3-5] Re: [REVIEW:3-5] fdo#49968 speed up RTF import of repeated character/paragraph properties

[Miklos Vajna]

On Mon, Jun 04, 2012 at 01:29:02PM +0200, Michael Stahl <mstahl at redhat.com> wrote:
> pushed to libreoffice-3-5
> 
> why does the tokenizer store the sprms in a vector anyway, would a map
> or hash_map perhaps be more suitable?

Check the ones, where the set() is called with false, i.e. not
overwriting previous sprms:

grep '\.set(.*false' source/rtftok/rtfdocumentimpl.cxx

Tables and numberings are the only ones, where repeated sprms are
necessary, I think. (Without them, a map would work, yes.)