getToolByName() vs. others

Question

What's the recommended method of accessing Plone persistent utilities and why?

• getToolByName(context, "portal_url")

• direct acquisition: context.portal_url

• ITools interface

etc.

Answer 1

I think that the answer is in the code of getToolByName itself ( http://svn.zope.org/Products.CMFCore/trunk/Products/CMFCore/utils.py?view=markup ). That method does this:

• as a first attempt it tries to get the desired tool by looking up for the interface with getUtility and including it in a context (that seems a good thing to me)
• as a fallback it tries to directly acquire the tool from the source context

So the only method "to rule them all" seems to be: getToolByName But, as @keul said, there's also the caching involved, and looking here ( http://collective-docs.readthedocs.org/en/latest/misc/context.html#itools-interface ) it is clear that using the ITools interface, when it doesn't fail due to tools that don't yet implement that interface, it's faster.

For the reasons above, in the end I'd suggest:

1. ITools (faster)
2. getToolByName (safer)

(as the direct acquisition is already tried by the getToolByName, if it fails, you won't have more fortune)

Answer 2

The use of ITools utilities must be the best method, as they are cached. However this sometimes gave me problems (related to portal_membership tool, but I don't remember details) so in that rare cases I switch to getToolByName. The last one is the direct acquisition.