Trying to connect some restricted Redmine instance to our Eclipse Mylyn environment it worked in the beginning, but the re-imports did not with some error "Failed to parse RSS feed".
I stumbled across this #246440 Eclipse Mylyn ticket where some workaround was to recreate the Task Repository including the Task List Queries by hand.
But this is not a nice solution.
So I played around a bit more and found the following that solved our import issues:
most likely for your needs: remove the key value (or other security-relevant data) from the exported <task list query>.xml.zip / tasklist.xml since the queries contain some user-dependent authentication API (e.g. if shared with other users)
Task Repository for all dependent queries and will be re-imported automatically on later importmake sure that (e.g. through some used formatter, CTRL + F or manual formatting) there are no whitespaces in text-value XML nodes, because thus the queries may stop working after import:
e.g.
<Attribute Key="Regexp">^({Id}\d+);({Type}[^;]*);...$
</Attribute>
should be:
<Attribute Key="Regexp">^({Id}\d+);({Type}[^;]*);...$</Attribute>
go on Task List -> <your imported query> -> right click -> Properties -> Finish so some internal magic "fixes" your query
Another debugging hint: you can always check the retrieved files (and Query Pattern regexp using the Preview button) using the <your query -> Properties -> Advanced Configuration -> Open button, which should put the unparsed query result in e.g. c:\Users\<loginname>\AppData\Local\Temp\mylyn-web-connector4155864524987884464.html.
By the way: (If you are at the above point it may likely be useful for you or your team ...) Using the web connector I found the integration via the API key in combination with the .../issues.csv... format much more useful and configurable than the .../issues.xml... variant.
We used something like this for parsing the CSV (and generated the params, their order etc. via normal filter dialogs): ^({Id}\d+);({Type}[^;]*);({Status}[^;]*);"?({Owner}[^";]*)"?;({Description}[^;]*)$.
Advantages are: easier regexp, concatenatable data for Description via column-ordering and fetching of all data without paging (=> we could skip page, per_page, limit, offset).