inspect:inspect-module($module) break on '@', '%', etc

[duncdrum]

What is the problem

xqDoc annotations that contain an @ are not processed correctly by inspect:inspect-module the fault is somewhere here i guess exist/src/org/exist/xquery/functions/inspect/InspectModule.java

What did you expect

@see or @returns comments can contain @ as part of the annotation, (email address and attributes) but the function stops at the character immediately before the @

Describe how to reproduce or add a test

add xqDoc comment to function and run inspect:inspect-module on the module, e.g.:

(:~
: this …
: @return taxonomy[@type = "reign"]
:)

will become

 <returns type="item()" cardinality="zero or more"> taxonomy[</returns>

instead of:

 <returns type="item()" cardinality="zero or more">taxonomy[@type = "reign"]</returns>

Context information

Please always add the following information

• eXist-db version 3.1.1
• Java8u121
• MacOs 10.12.4
• 64 bit

[duncdrum]

see also this tread from exist-open there are more problems with xqdoc annotations and special characters

[kohsah]

Same issue applies to % directives in the functions, where it simply fails.

I think the lexer needs to be updated here : https://github.com/eXist-db/exist/tree/develop/src/org/exist/xquery/xqdoc/parser

[adamretter]

@kohsah Okay you are right the Lexer does need to be updated. I had a brief attempt this morning and failed. Sadly I am not very familiar with Antlr2, perhaps @wolfgangmm or @ljo could make the small necessary changes?

@duncdrum I just want to point out something that may be confusing. If you are using inspect:inspect-module this uses eXist's own XQDocParser as found here: https://github.com/eXist-db/exist/tree/develop/src/org/exist/xquery/xqdoc/parser

However if you are using the XQDoc extension module, i.e. xqdm.scan then that uses the XQDocParser from the xqdoc.org project, which is only designed for XQuery 1.0.

I would also note that the output documents of inspect:inspect-module and xqdm:scan are different which likely further causes confusion. I personally think that both should be standardised on the updated Schema provided by @ChristianGruen here http://xqdoc.org/xqdoc-1.0.01132014.xsd

[duncdrum]

@lcahlander since your recent work in this area, tagging you here for future reference. Any comments welcome.