Bastien, When there is no property drawer before the code block, C-c C-x p affects the code block and either doesn't create a property drawer or it leaves the 'real' property drawer unaffected.
An example for the second case, where org-set-property matches the code block, despite an existing property drawer that happens to be below it. The wrong "NAME" is matched, and "DEPENDS" isn't matched at all. This could throw off for example org-dotemacs, where it would tangle code blocks in the wrong order without the DEPENDS matching. * stuff for bug report #+BEGIN_SRC emacs-lisp (setq org-capture-templates (quote ( ("c" "Contacts" entry (file+headline "~/my-stuff/file.org" "Contacts") "* %^{Name: } :PROPERTIES: :EMAIL: %^{Email} :PHONE: %^{Phone number} :NAME: %^{Full Name} :END: %? ")))) #+END_SRC :PROPERTIES: :NAME: org-capture-templates :DEPENDS: org :END: Thanks, Tod b...@gnu.org writes: > Hi Tod, > > Tod Middlebrook <todmiddlebr...@gmail.com> writes: > >> The bug below prevents me from easily using dependencies in org-dotemacs. >> >> To reproduce, >> start with this entry: >> >> *** stuff for bug report >> #+BEGIN_SRC emacs-lisp >> (setq org-capture-templates >> (quote >> ( >> ("c" "Contacts" entry (file+headline "~/my-stuff/file.org" >> "Contacts") >> "* %^{Name: } >> :PROPERTIES: >> :EMAIL: %^{Email} >> :PHONE: %^{Phone number} >> :END: >> %? >> ")))) >> #+END_SRC >> >> Then do C-c C-x p EMAIL [RET] TestValue, and get the same block, with >> the properties drawer folded. When expanded, there is: >> >> *** stuff for bug report >> #+BEGIN_SRC emacs-lisp >> (setq org-capture-templates >> (quote >> ( >> ("c" "Contacts" entry (file+headline "~/my-stuff/file.org" >> "Contacts") >> "* %^{Name: } >> :PROPERTIES: >> :EMAIL: TestValue >> :PHONE: %^{Phone number} >> :END: >> %? >> ")))) >> #+END_SRC > > I'm not sure to understand what the problem is exactly: if the problem > is that `C-x C-c p' works in the context of source code blocks, we can > easily fix it. If the problems is that such properties are matched in > contexts where they should not, we need more information about when > you observe the wrong behavior, i.e. in what context do you see the > properties taken into account while you expect them to be ignored? > > Thanks in advance for further details,