getTreeIdList is still required by Folder_reindexAll.
getTreeIdList is not immediate anymore but the reindexing of a folder should
anyway be reimplemented to not depend on this method: only one tree is indexed
at a time, which is unefficient when they're small.

_getOb and similar methods are reimplemented in a faster and safer way.
It now checks it is only used to return leafs.

Similarly, _delOb now refuses to delete trees at the root.

__getattr__ wrongly returned wrapped results (__of__).

Since commit 055d0a69
("HBTreeFolder2: make object{Ids,Values,Items} really lazy"),
_fixCount() does nothing because objectIds() was optimized in a way that
len(self.objectIds()) already returns self._count()

The line added in commit e42b2816
is not valid XML.

1.8.3 is now required to use profiling.
It provides two important features:
- SQL query tracing (which queries were executed, how many times,
how long did they take to run)
- ZODB __setstate__ per-oid statistics

- contentValues(filter) is really slow
- portal should be explicitly retrieved before using one of its methods

* for bulk substitution, re.sub is much faster than join of dict
  lookup result per byte.

Check that if only the code of a PythonScript (_body) is different,
it will be identified as modified.

now it is like "<title>public - Revions 12345: /path</title>".

For scheduled workflow transition. Use in your workflow transition as:
%(content_url)s/Base_viewScheduleWorkflowActionDialog?workflow_action=%(transition_id)s&next_workflow_action=SOME_ACTION
(replacing SOME_ACTION with some appropriate value).

Delayed workflow transition works with 3 states:
- initial state: where we are
- final state: where we want to go
- waiting state: where we wait for the right time to transition to final state
The main advantages of the waiting state are:
- prevents multiple transitions from being scheduled
- allows customising object security (ex: preventing unwanted edits)
- allows storing intended execution date in a transition comment

Do not require caller to cast between these types when they can all make
sense to appear in the same list.

in most cases, where ASCII-range characters are less than 80% of input
non-utf8 string, the result is same but we can reduce needless string
replacement calls. in rare case, for example wrongly iso-8859-1
encoded properties or PDF data, the result becomes different, i.e.
before: ASCII-range characters + escaped characters => after: base64
encoding, but the meaning is still same.

* stop using deprecated string functions.
* reduce lambda creation inside functions.
* move import statements outside of function.

* remove needless copy of sub objects that are already recursively
  copied by their top most object.
* remove needless manage_beforeDelete() calls when copying objects.

- Document_viewSVGEditor/my_data changes web illustration content type
- WebIllustration_viewSVGEditor/my_text_content does not change it

Installation of business template may change the encoding of a document to utf-8.
Nonetheless, export should preserve the original encoding.
This test creates a ZopePageTemplate with output_encoding of 'iso-8859-15'
and checks that while installed version becomes 'utf-8', when exported it
keeps the original one.

Only xml metadata should be exported in this case.
In the previous version the check of hasattr could pass
if attribute is defined as an attribute of the class and not of the instance,
causing also AttributeError later when the code tries to invoke delattr.

Also, restructure export. First try to delete property and then
export as separate file.

Test also added.

also make it work for Image portal types

each view has both editable and view mode, they can be accessed and switched
Editable button bring to editable mode
View button select view instead of Tabs which is meaningless

/reviewed-on nexedi/erp5!108

About getSingleCategoryAcquiredMembershipList, getPhysicalPath is so slow that
that cache was useless.