[gdal-dev] Motion: adopt RFC 87: Signed int8 data type for raster

[Even Rouault]

>
> So based on that RFC it is a breaking change when reading signed 
> images. I confess that this type of silent behavor change scares me.
It is documented in 
https://github.com/OSGeo/gdal/blob/master/MIGRATION_GUIDE.TXT which is 
always referenced in the release notes.
>
> Why dont make this an opt-in behavior instead of just breaking 
> application code silently?
Opt-in behaviour would mean clutering the GDAL code base even more, to 
implement both the legacy and the new behavior. One of the purposes of 
the RFC was to de-clutter the GDAL code base.
> Otherwise, I would rather have GDT_Byte removed so that application 
> code would not compile against new versions of GDAL. At least that 
> makes migration to new versions less risky.

Removing GDT_Byte would have enormous breaking implications! Signed Byte 
is I believe a marginal feature not commonly used by GDAL based 
applications, partly because the support for it before RFC87 was clunky 
and partial.

Even


-- 
http://www.spatialys.com
My software is free, but my time generally not.