Fix inefficiency of the jiraissuesmacro

The jiraissues macro currently blindly accepts the xml query specified via the "url" parameter to the macro. This results in the fetching of all fields of the issue and all of the comments even though the vast majority of this information will be thrown away by Confluence when the macro is rendered. This is extremely inefficient and results in very long page loading times when more than a handful of issues are displayed or if some fields have large quantities of information.

The workaround is to specify "field=" parameters for only the necessary fields that the macro will display as defined by the "columns" parameter of the macro. Unfortunately this is not straightforward because the "columns" parameter of the macro uses actual column names but the "field=" parameters to the xml search URL uses database column names. Also, the use must manually add "field=link" in order for the issue to be clickable.

All of this should be automatic and the fields fetched from JIRA by Confluence should be limited to only the fields specified via the "columns" parameter of the jiraissues macro.