|
Attachment karaf.log has been added with description: karaf log of integration while mounting the device
|
|
Input YANG file contains invalid regular expression, hence the error. Note that YANG follows XSD regular expressions, as defined in https://www.w3.org/TR/2004/REC-xmlschema-2-20041028/#regexs. These do not allow ^/$ anchors (as anchoring is implicit) – which are obviously present.
Error on invalid input, downgrading to improvement/normal, targetting to SR4. What yangtools can do is catch the case when an expression starts/ends with an anchor, complain loudly and ignore the anchor points.
|
|
Also, a YANG snippet defining the leaf would be nice.
|
|
Thanks for your support Robert.
I'll look into my yang files and try to resolve the anchors which are not allowed.
|
|
I am not sure, but I think one problem with regex matching may be that it does not interpret $ correctly, we'll continue to investigate.
|
|
fix pushed: https://git.opendaylight.org/gerrit/#/c/35399/
stable/lithium: https://git.opendaylight.org/gerrit/#/c/35407/
|
|
THe agreed plan of action is to import the XSD RegEx engine from Xerces, under third-party/xsd-regex. Note the import should include only required classes and those need to be imported in the org.opendaylight.yangtools.xsd.regex.
These will then need to be modified, such that RegularExpression has a:
public java.util.Pattern toPattern() throws PatternSyntaxException;
method, which will produce a compiled Pattern. If an expression cannot be converted to a Pattern, PatternSyntaxException should be reported.
This pattern will then used in Java to validate inputs.
|
|
import of Xerces regex engine: https://git.opendaylight.org/gerrit/39168
cleanup to comply with ODL coding standards: https://git.opendaylight.org/gerrit/39169
|
|
master: https://git.opendaylight.org/gerrit/45145
|
|
While using Xerces' RegularExpression works, it has a very undesirable side-effect of expanding regexes – which renders debugging and error reporting a horror.
Hence I think it will be more feasible to create a simple XSD RegEx tokenizer, which will emit minimally-modified strings.
|
|
Alternate approach: https://git.opendaylight.org/gerrit/46990
|
|
RCO delivery
|
|
(In reply to Robert Varga from comment #1)
> What yangtools can do is catch the case when an expression starts/ends
> with an anchor, complain loudly and ignore the anchor points.
I am not sure whether we are really able to do that, because for an example consider the following XSD pattern "x$". Are both chars '' and '$' added mistakenly as anchor points and should therefore be omitted or are they added intentionally in order to match string "^x$" ?
|
|
Adapted XSD regex grammar does not work for ANTLR correctly. I spent some time trying to fix it, but there are more issues in the grammar like ambiguity of rules, order of rules, etc. Therefore, I have proposed the following patch based on iteration over regex chars: https://git.opendaylight.org/gerrit/#/c/52999/
I think it is sufficient for our needs...
|
|
Patch for xsd-regex removal from Yangtools:
https://git.opendaylight.org/gerrit/#/c/53000/
It needs to be fixed (yangtools-carbon-distribution-check job fails). Resolving in progress ...
|
|
We should actually backport this to Boron
|
|
Reopening as problem persists in Boron-SR3. The comments below suggests, that if a regexp contains anchors, yangtools should ignore them. Instead of that, yangtools is escaping them and adding one more set of anchors, generating a compilable but unusable regexp pattern:
Yang (https://github.com/openconfig/public/blob/master/release/models/types/openconfig-yang-types.yang) :
typedef dotted-quad {
type string {
pattern '^(([0-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|' +
'25[0-5])\.)
{3}([0-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4]' +
'[0-9]|25[0-5])$';
}
}
- note all openconfig regexps contain anchors ^ and $
Generated java regexp:
public static final List<String> PATTERN_CONSTANTS = ImmutableList.of("^\\^(([0-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-5])
.){3}
([0-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-5])
$$");
Test :
@Test
public void testDottedQuad() {
new DottedQuad("127.0.0.1");
}
Output :
(([0-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-5])\.)
{3}([0-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-5])
Exception in thread "main" java.lang.IllegalArgumentException: Supplied value "127.0.0.1" does not match required pattern "^^(([0-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-5])\.){3}
([0-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-5])\$$"
at com.google.common.base.Preconditions.checkArgument(Preconditions.java:145)
at org.opendaylight.yang.gen.v1.http.openconfig.net.yang.types.yang.rev170403.DottedQuad.<init>(DottedQuad.java:52)
|
|
https://tools.ietf.org/html/rfc6020#section-9.4.6:
The "pattern" statement, which is an optional substatement to the
"type" statement, takes as an argument a regular expression string,
as defined in [XSD-TYPES].
https://www.w3.org/TR/2004/REC-xmlschema-2-20041028/#regexs:
Note: Unlike some popular regular expression languages (including those defined by Perl and standard Unix utilities), the regular expression language defined here implicitly anchors all regular expressions at the head and tail, as the most common use of regular expressions in ·pattern· is to match entire literals.
That means that the models are wrong and they should not contain anchors.
|
Generated at Wed Feb 07 20:53:41 UTC 2024 using Jira 8.20.10#820010-sha1:ace47f9899e9ee25d7157d59aa17ab06aee30d3d.
karaf.log