wrong language parsing case

If you have an header with just the language and no spaces or other arguments, like #+begin_src python, when point is inside a polymode managed buffer (so a src block) org parses the language wrongly, it includes the subsequent line...the regex used by org to parse the language uses \S so any includes non white-space...and indeed \n is not a whitespace..but why does it happen only inside polymode? Parsing the point is at the header line works fine...

My current workaround is to just append a space to the end of the line if there isn't one

[vspinu]

org parses the language wrongly, i

I don't quite follow. Can you provide a simple org mode example with the location of the pointer and explain what the above "parses wrongly" means?

Thanks!

We have:

#+begin_src python
import os
#+end_src

If (point) is on #+begin_src or #+end_src, I get:

(print (first (org-babel-get-src-block-info))

output is:
"python"

If (point) is inside...that is

#+begin_src python
import os # <-- HERE
#+end_src

I get...

(print (first (org-babel-get-src-block-info))

output is:

"python
import"

Note the new line, and it doesn't have os because there is a space between import and os,
indeed the function org-element-src-block-parser defined in org-element.el matches against \S that is any non whitespace; since the org regex is loose about new lines I guess this is a problem with narrowing? since probably that org function only matches against 1 line, but when inside an innermode it matches against multi lines.