And Chameleon does not workaround by lowercasing them automatically.

Revert "Test deferred style with converting to a format"

This reverts commit cbfa8d56.
-> pushed by mistake

Revert "Set cloudoo on preferences in ERP5TypeTestCase"

This reverts commit 28824962.
-> pushed by mistake

Revert "Listbox from Folder_viewWorkflowActionDialog has hidden columns"

This reverts commit 802d3f53.
-> a test is failing

Adjust test accordingly and fix bad test.

- getting document original content if format is empty
- conversion raster image -> svg

cause running out of memory so enforce some restrictions by default which will prevent this.

Also simplify sample HTML in test_safeHTML_conversion

portal_caches tool is expected to allow only Cache Factory instances inside.
Fix tests to use Cache Factory instead of Cache Bag.

Fix tests (DMS) to use by default Cache Bags.
Fix some dead code and clean up.
Move cache initialization away from DMS into Cache Factory.

We cannot wait for the document to be created to introspect its properties,
because we need to pass the input parameters to newContent. Instead use a fixed
list of common parameters

This is a followup to commit 8995b645
("always start transactions with begin() and make tic() commit automatically"):
there is now no need to use self.stepTic() instead of self.tic()

In poppler 0.18+, pdftohtml generates " " instead of " ", so
make sure both are accepted.

Fix url calculation.
Pass only applicable arguments to conversion API.

Add test to show a failure ot web page to image when more than one external image is used in web page body.

Increase time for convertion due to change to cloudoo (slow but more stable oood) and parallel test execution.

Followup of 00bdb44f. This fixes
test_parallel_conversion from testDms.TestDocument.

This fixes test_base_convertable_behaviour_with_successive_updates from
testDmsWithPreConversion.TestDocumentWithPreConversion which was previously
working properly with files only.

Also, split up the test for deleting data as it's not implemented yet.

Start preconversion in a separate serialize able activity.

Make testDms use by default RAM based cache storage as some tests require it. (testDmsWithFlare just changes this storage to a persisten one).

If pre_converted_only is True system will return desired conversion only if it's already cached.
Extend testDms to cover it.

This patch will always perform conversion against given encoding, in order to check if this codec is valid or not.

Check that declared encoding is supported by python before using it.

If Image is a Text, then conversion to image format is performed

The change from ocropus to tesseract as the conversor application
responsable  to convert image to text, made this tinny difference
between the outputs.

The difference was manually tested by Rafael to verify that it was caused
by the change of conversor application.