Propsal for Annotations about tikxml HOT 2 CLOSED

Regarding the points of inconsistency:

1. maybe another solution is to expect, that xml tags are unique (e.g. should always be mapped to a java class Book). if so you don´t have to specify all the TypMatcher and can search for a class with this name or annotated with this name. and if you have an xml document where should be mapped to another class you could specify this with the TypMatcher as you already could with converter

2. if possible you should only have one annotation @ChildElement. the problem is here the inlineproperty. you could simple ignore it if the annotated field is not a collection or write an extra lint check not to use this property on other fields which are not a collection. but if this is not possible you have to use two annotation - simple xml is also using @Element and `@ElementList``

3. first of all @Pathshould work with both annotations. you could avoid @TextContent and use @PropertyElementwith an inline property but then only one @PropertyElement can be used with inline=true. so i think it would be ok to use both annotations also because of a better readability

from tikxml.

1. We need to know how to map from xml element name to java class, hence an additional annotation like @TypeMatcher is required. Furthermore, if you have classes annotated in a library and in the app itself, annotation processing runs during compiling the library. Then you include the compiled library into your app and annotation processing runs only on your android apps annotated classes. So this dynamic mapping from element name is not possible at runtime. The only way to do that at runtime is by using reflections, but its super slow to scan all classes to determine the xml element name.
2. Ok, we use only @Element for both, simple child elements and list of elements. Additionally we introduce @InlineList for inlining list of elements.
3. @TextContent applies on the current xml element whereas @PropertyElement applies on chile elements of the current element. Hence both are used for different use case and needed.

Additionally we also introduce a "mode" to write less annotations: @Xml( mode = Mode.ANNOTATIONS_ONLY or Mode.COMMON_CASE) where COMMON CASE means, read primitive java types as xml attributes (corresponds to annotating the field with @Attribute) and object as xml child elements (corresponds to annotating the field with @Element). So for the common case less annotations are required. However, in the common case we need a way to tell the parser to ignore a field. We could use java transient keyword. However, that might cause conflict with other tools. Therefore we will introduce @IgnoreXml.

from tikxml.