This is a collection of libraries which offer data provenance facilities and then use that to provide a variety of marked data structures.
The data marks are intended to be similar to, though not entirely the same as
data structures seen in other systems as Span
or similar. Instead data marks
are a combination of that span-like data and also file sources. The intent here
being that combining data from multiple sources is a safe and sensible thing to
do and that after doing so, it is important to be able to discover where something
came from.
Currently we provide:
marked-yaml
: A deliberately simplified YAML subset which supports data marking
The marked-yaml
library offers a subset of YAML designed for configuration files
and similar use. The data it reads in is "marked" with its origin. If you want
to use marked-yaml
with your existing serde
applications, you can enable the
serde
feature, and if you want the errors produced by the marked-yaml
deserializer to include nice paths to any problem, along with ensuring the marker
for the problem area is populated in any errors, use the serde-path
feature.