summaryrefslogtreecommitdiff
path: root/jsonschema/_validators.py
Commit message (Collapse)AuthorAgeFilesLines
* Revert "Respect ECMA regex syntax for schemas."v3.1.1Julian Berman2019-10-091-3/+3
| | | | | This reverts commit 1a02f547fda27f2b267fcbd03711fe36e50b7026, reversing changes made to cbb222c145b71abb51e743ad48083c8ba5fa0cdb.
* Use JS regex syntaxZac Hatfield-Dodds2019-10-071-3/+3
| | | | Note that js_regex.compile has a cache, just like re.compile, so calling it in loops does a lot of dict lookups but very little work.
* And now fix enum.Julian Berman2019-08-011-1/+6
| | | | Closes: #575.
* Unskip and fix the const-related compare bug.Julian Berman2019-08-011-8/+15
| | | | Refs: #575
* Split out dependencies for draft 3.Julian Berman2019-01-141-14/+6
| | | | | | | | Draft 3 allowed simple string dependencies, but no later draft does. This also lets us use the same fix for boolean schema nonsense as we just did for items.
* Here's an easy way out of this one.Julian Berman2019-01-131-11/+5
|
* And tests for the same for if/then/else.Julian Berman2019-01-111-3/+2
|
* Fix the error properties for propertyNames.Julian Berman2019-01-111-1/+0
| | | | | | | | | | | | Note that propertyNames is another example (like #119) of something that needs an additional way to say it has errored on a *key* within the instance. So, that's still missing, but path is not the right thing, same as in the earlier instances. It's also becoming somewhat clear that compound validators should possibly grow a way to become lazy...
* Get these out of the way.Julian Berman2019-01-111-157/+0
|
* Dead.Julian Berman2018-10-011-12/+2
|
* If/Then/Else.draft7Julian Berman2018-09-301-0/+13
|
* More suffix killin'.Julian Berman2018-09-291-2/+2
|
* Make draft3's any type not be special cased.Julian Berman2018-09-031-2/+0
| | | | | | | Previously, is_any was uncovered/dead code. As a crazy 'side effect', you can redefine the any type now. Go wild.
* These are here to stay.Julian Berman2018-08-191-5/+5
| | | | Might as well give 'em names that say that.
* Nope, seems unlikely.Julian Berman2018-08-011-0/+12
| | | | | | | items, which has this same singleton checking, fails its tests when this is removed, so probably these are just missing boolean tests. Back in they go.
* Handled a level up now.Julian Berman2018-08-011-12/+0
|
* Fix the contains error message.v3.0.0a1Julian Berman2018-06-031-1/+1
|
* Fix the const error message.Julian Berman2018-05-151-3/+1
|
* Pass the propertyNames test properly now.Julian Berman2017-10-291-0/+3
|
* FineJulian Berman2017-10-291-6/+6
|
* propertyNamesJulian Berman2017-10-291-0/+10
|
* oneOfJulian Berman2017-10-291-0/+28
|
* minPropertiesJulian Berman2017-10-291-1/+1
|
* maxPropertiesJulian Berman2017-10-291-1/+1
|
* itemsJulian Berman2017-10-291-0/+21
|
* dependenciesJulian Berman2017-10-291-0/+5
|
* containsJulian Berman2017-10-291-0/+10
|
* ConstJulian Berman2017-10-291-0/+7
|
* bignumJulian Berman2017-10-291-2/+46
|
* anyOfJulian Berman2017-10-291-0/+18
|
* allOfJulian Berman2017-10-291-4/+14
|
* And the additionalProperties tests.Julian Berman2017-10-291-1/+1
|
* Pass the additionalItems tests.Julian Berman2017-10-291-1/+1
|
* Address review comments.Itamar Turner-Trauring2017-01-261-3/+6
|
* Better error message for additionalProperties: false when there are ↵Itamar Turner-Trauring2017-01-261-2/+12
| | | | patternProperties.
* More style fun.Julian Berman2016-11-271-1/+1
|
* Revert "Made required_draft4 more robust when used with Draft3"Julian Berman2016-10-091-5/+0
| | | | This reverts commit 41f2f8d9658606876786de2564d07fbb5bc619f8.
* Made required_draft4 more robust when used with Draft3William Speirs2016-09-271-0/+5
|
* Fix formatting error for unique items if instance is tuple.Igor Davydenko2015-05-081-1/+1
|
* Need to preserve backwards compat for RefResolvers without the new methods.Julian Berman2015-03-211-7/+13
|
* Remove context manager from ref() validation.Daniel Nephin2015-03-011-1/+6
|
* Improve mutlipleOf validator with floating point value.Vincent Pelletier2014-12-291-5/+2
| | | | | Rely on system's built-in float tolerance by comparing quotient with cast quotient instead of checking distance to remainder.
* Closes maximum issue which wasn't an issue.Sebastian Mika2014-09-231-4/+4
| | | | Remove float conversion on minimum validator to make it work for very negative numbers
* All tests pass when adding `float()` conversion on `instance` and `maximum` ↵Sebastian Mika2014-09-231-4/+4
| | | | and `minimum` in respective validators
* Add in float(instance) in maximum validatorSebastian Mika2014-09-231-2/+2
|
* No idea why this call to float was here.Julian Berman2014-06-011-2/+2
| | | | | | Fix large precision comparisions (even for ints). Closes #166
* Don't floatify the instance in minimum/maximum messages.Julian Berman2013-12-261-6/+4
|
* Minor formatting.Julian Berman2013-12-221-3/+3
|
* Format can apply to instances of any type.Julian Berman2013-10-281-4/+1
| | | | Closes #125
* This allows defining custom typesJohn Anderson2013-10-171-1/+1
|
View Lorry Controller Status