[FYI] Adding support for Unicode normalisation conversion to lib/util/charset/

Douglas Bagnall douglas.bagnall at catalyst.net.nz
Wed Apr 10 10:55:45 UTC 2019

hi Ralph,

I am interested to see this because a few years ago I was looking at using
libicu for the sort module in ldb (our current algorithm is idiosyncratic)
but that work stopped when I realised Windows does something else again.

It is something I think we need to grapple with at some point, whether we
follow Windows or cleave away and join the rest of the world in the
Unicode standard with ICU.


On 10/04/19 4:50 AM, Ralph Böhme via samba-technical wrote:
> Hi!
> Fwiw, maybe this of interest to anyone: I have a small patchset that adds
> support for Unicode normalisation NFC/NFD conversion to lib/util/charset/
> using libicu.
> I need this to convert strings coming in over mdssvc RPC (aka Spotlight)
> from Mac clients.
> There's a more complete patchset running in this pipeline:
> https://gitlab.com/samba-team/devel/samba/pipelines/55974836
> This is on-top of the gitlab WIP patches for container image rebuild
> automation. I'm adding libicu-devel packages to the images here to ensure
> the torture test of the patchset is always run.
> I will submit a full patchset once the gitlab stuff is upstream. Hopefully
> soon! :)
> -slow

More information about the samba-technical mailing list