using AtomicString will mess up "<" and ">" characters

[sander76]

for example:
div.text = AtomicString("< test >")

will come out as:
< test >

using python 3.5

[waylan]

I expect this is probably related to mkdocs/mkdocs#751.

Thanks for the report.

[waylan]

I've taken a closer look and this is not related to the MkDocs issue. That is specifically a bug in MkDocs.

Actually, now that I think about it, this behavior is expected. I supose you were thinking that because it was an AtomicString, that no additional processing should be done on it. That is actually what happens. The "parser" does no additional processing. However, the serializer does sanitation/escaping on any input it receives, including AtomicStrings. In fact, the serializer has no knowledge of AtomicString. It simply sees a normal string and treats it the same.

If you want to bypass the serializer, then you need to not have your output (presumably from and extension) go through the serializer. Note that some extensions (codehilite and fenced_code_blocks) do this by storing their output in the HtmlStash for later retrieval after the serializer has run. Others may operate as a postprocessor completely for the same reason.

In any event, there is no bug here, unless you can show different behavior in different versions of Python (which is what I thought you were reporting at first--turns out the behavior is the same across versions in my testing).

[sander76]

Ok thanks for diving into this.
However,
div.text = "< test >"
comes out as:
< test >

This is what I use now as it gives me the result I want, but I feel I should have been using AtomicString as I want my string not to be processed anymore..

[waylan]

It sounds like what you want is something that will be coming in version 3.0. We will be introducing a new type: RawText which will get ignored by both the parser and the seriallizer. You will have to wait for 3.0 to get that feature though. See the Roadmap for more info.

In the meantime, I'll label this as 3.0 and leave it open.