[WIP] Implement initialisation pass of the schema to find all $id references - #8

Open

KayEss wants to merge 24 commits into

feature/two-pass

Commits on Jul 3, 2019

We will need a schema cache in each schema to handle their `$id` entries.
KayEss
committed
7f5f602
clang-format
KayEss
committed
f0fb3f9
It's very important that this does not end up as an alias for `nullptr`.
KayEss
committed
5f75ce4

Commits on Jul 4, 2019

Add an initial pass when a schema is first loaded to look for embedded `$id`s
KayEss
committed
03d1f52
Change the cache map to be keyed on actual URLs
KayEss
committed
74693f2
Preloading requires knowledge of the base URL so we can calculate the real one.
KayEss
committed
8f3fc4b
Deal better with URL fragments that aren't JSON pointers.
KayEss
committed
9fdda79
clang-format
KayEss
committed
926bfd4
Temporarily lock to and older version of the test suite.
KayEss
committed
1c64ba5
Merge branch 'temp-pin' into feature/two-pass
KayEss
committed
fae7cb7

Commits on Jul 8, 2019

Propogate URL type up call chain.
KayEss
committed
6ee7fd5
We must rely on the base URL set for the schema in its constructor
KayEss
committed
0ec72a6

Commits on Jul 9, 2019

Commits on Jul 10, 2019

Properly preserve the original URL.
KayEss
committed
d826c9a

Commits on Jul 16, 2019

Refactor the cache lookup to return the sub-schema part for json pointers.
KayEss
committed
e68ac41

Commits on Jul 18, 2019

Remove some memory allocations in regex pattern matching.
KayEss
committed
d02cb29

Commits on Jul 21, 2019

Make all URLs relative to the same base.
KayEss
committed
e44618c

Commits on Jul 22, 2019