[
https://issues.apache.org/jira/browse/UIMA-239?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13181421#comment-13181421
]
Marshall Schor commented on UIMA-239:
-------------------------------------
After some more investigation, there's quite a difference between the
JAXB-generated classes and our current "models"; it would be a very big piece
of work to switch, I think.
So, I'm now thinking of this alternative:
1) modify the existing parser to optionally (additionally) save the DOM (that
is already being produced, with the ignorable-whitespace and comments nodes.
2) add a method for serializing to XML that takes this saved DOM, and then does
a fairly naive adding back of the comment and whitespace nodes while
serializing, from the saved DOM, where it can find a "match". This ought to
save most of the comments, except of course, if whole nodes have been deleted
or added. A general solution might be quite complex, but I suspect we can
cover the major use cases quite simply, since the xml at each level is
"ordered" (mostly).
This would have a very minimal impact on the existing code :-). And the
additional information (including comment nodes themselves) would only be
collected if requested. (I suspect only the CDE will use this for now).
> Tools like the CDE which use XML parsing should preserve XML comments;
> parsers should provide access to XML comments.
> ---------------------------------------------------------------------------------------------------------------------
>
> Key: UIMA-239
> URL: https://issues.apache.org/jira/browse/UIMA-239
> Project: UIMA
> Issue Type: Improvement
> Components: Core Java Framework, Tools
> Reporter: Marshall Schor
>
> An example of a problem which happens if this is not done: the descriptors
> that come with the UIMA toolkit have a comment that specifies the license
> information for that file. If you edit a descriptor containing such a comment
> with the CDE, it strips out all comments.
> An alternative for license info: add a new, optional element to the
> descriptor for specifying the license (but this might be ambigious as to
> what is being licensed - the descriptor file or what it refers to or ??? )
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
