Closed phdowling closed 1 month ago
It looks to me like the pattern in the generated code is wrong. Pasting the restriction source into an html decoder gives the following output
<xs:restriction base="xs:string">
<xs:pattern value="( |'|[,-\.]|[A-Z]|[`-z]|~|¨|´|·|[À-Ö]|[Ø-ö]|[ø-ž]|[Ƈ-ƈ]|Ə|Ɨ|[Ơ-ơ]|[Ư-ư]|Ʒ|[Ǎ-ǜ]|[Ǟ-ǟ]|[Ǣ-ǰ]|[Ǵ-ǵ]|[Ǹ-ǿ]|[Ȓ-ȓ]|[Ș-ț]|[Ȟ-ȟ]|[ȧ-ȳ]|ə|ɨ|ʒ|[ʹ-ʺ]|[ʾ-ʿ]|ˈ|ˌ|[Ḃ-ḃ]|[Ḇ-ḇ]|[Ḋ-ḑ]|ḗ|[Ḝ-ḫ]|[ḯ-ḷ]|[Ḻ-ḻ]|[Ṁ-ṉ]|[Ṓ-ṛ]|[Ṟ-ṣ]|[Ṫ-ṯ]|[Ẁ-ẇ]|[Ẍ-ẗ]|ẞ|[Ạ-ỹ]|’|‡|A̋|C(̀|̄|̆|̈|̕|̣|̦|̨̆)|D̂|F(̀|̄)|G̀|H(̄|̦|̱)|J(́|̌)|K(̀|̂|̄|̇|̕|̛|̦|͟H|͟h)|L(̂|̥|̥̄|̦)|M(̀|̂|̆|̐)|N(̂|̄|̆|̦)|P(̀|̄|̕|̣)|R(̆|̥|̥̄)|S(̀|̄|̛̄|̱)|T(̀|̄|̈|̕|̛)|U̇|Z(̀|̄|̆|̈|̧)|a̋|c(̀|̄|̆|̈|̕|̣|̦|̨̆)|d̂|f(̀|̄)|g̀|h(̄|̦)|j́|k(̀|̂|̄|̇|̕|̛|̦|͟h)|l(̂|̥|̥̄|̦)|m(̀|̂|̆|̐)|n(̂|̄|̆|̦)|p(̀|̄|̕|̣)|r(̆|̥|̥̄)|s(̀|̄|̛̄|̱)|t(̀|̄|̕|̛)|u̇|z(̀|̄|̆|̈|̧)|Ç̆|Û̄|ç̆|û̄|ÿ́|Č(̕|̣)|č(̕|̣)|ē̍|Ī́|ī́|ō̍|Ž(̦|̧)|ž(̦|̧)|Ḳ̄|ḳ̄|Ṣ̄|ṣ̄|Ṭ̄|ṭ̄|Ạ̈|ạ̈|Ọ̈|ọ̈|Ụ(̄|̈)|ụ(̄|̈))*"/>
Looks like something is converted wrong and generates a newline character, which causes ruff to complain on the unterminated string.
Thanks for reporting @rbruenig the fix is on main.
The pattern is not used during parsing, it's rendered only as information for the developer, at some point I am hoping I can translate xsd patterns to python regex.
You can also disable rendering them completely, see --ignore-patterns
Apologies in advance for exposing you to the convolution that is the German XJustiz data format. Nevertheless, I ran into this issue today: the model generation fails for all of the schemas here (select any version and click on XSD Schemata to download one, for example this one).
For example, when I run
xsdata xjustiz_0400_register_3_3.xsd
this happens:From a quick look around, I would guess the issue has to do with the long sequences of special characters in
din-norm-91379-datatypes.xsd
, e.g. starting on line 29:Any help would be greatly appreciated.