On Mon, Jan 18, 2016 at 01:45:58PM -0800, Pavel Sanda wrote: > Scott Kostyshak wrote: > > > Scott, you might want to look over changes in LFUNs API in LyXAction.cpp > > > during 2.2dev cycle if we meant this conversion seriously. > > > > Which conversion? I don't understand. What specifically should I look > > at? > > I would say check git log -p LyXAction.cpp during 2.2 cycle, if there > are some changes which would screw up functionality of e.g. user bind files. > > It also used to be good habit to hint any changes in LFUNs in RELEASE_NOTES > (adding/removing/changing semantics, this check would tell you whether the > list is up to date.)
Ah yes this is a good idea. We do still document changes in LFUNs. I like the idea of looking at git log -p LyXAction.cpp. > > > That said I think bunch of LFUN changes is not easy or possible to convert > > > at all. > > > > Do you mean that if we have changes to LFUNs then we might consider > > converting e.g. a bind file to approximate the behavior before the LFUN > > change? If I understood correctly (which I doubt), I'm not sure we > > should do that. > > I think it is reasonably easy to make rename conversions (i.e. the lfun is > exactly the same, only the name changed), I wouldn't push it to the extent > that > we have to preserve changes in syntax (e.g. number of parameters and similar). > > In any way I am not pushing this forward; it's just that the only reason > why to keep LFUN format number is to have these kind of conversions working. > If we don't want to do those conversions we can drop the whole effort of > keeping LFUN_FORMAT in the code. Indeed, good point. I definitely agree with the rename conversions. Scott
signature.asc
Description: PGP signature
