Eric Schulte <schulte.e...@gmail.com> writes:
> Eric Abrahamsen <e...@ericabrahamsen.net> writes:
>
>> Gary Oberbrunner <ga...@oberbrunner.com> writes:
>>
>>> It seems like you can define "procedures" in org-mode and call them
>>> from elsewhere, with args.
>>> But I'm not sure how well-defined that process is; the documentation
>>> is not completely perfect yet I think. Here's one thing I'm trying
>>> that seems not to work.
>>>
>>> I define a "procedure" as a named ref called recorddate with two args,
>>> ver and order. The idea is I could later call that with different
>>> values of those args.
>>>
>>> #+NAME: recorddate(order="desc")
>>> #+BEGIN_SRC sql :exports none :colnames no :results scalar
>>> select Event.CreatedAt from Event join MachineInfo as MI on
>>> Event.MachineInfoId=MI.Id
>>> where Event.CreatedAt is not NULL order by CreatedAt $order limit 1;
>>> #+END_SRC sql
>>>
>>> (BTW, I really like how $ vars are substituted into SQL. Nice.) But
>>> when I try to call it like this:
>>>
>>> * earliest record is call_recorddate(ver="'.'", order="asc")
>>> or like this:
>>> #+CALL: recorddate(ver="'.'", order="asc")
>>>
>>> and I try to export as LaTeX (or anything), I get
>>> org-babel-ref-resolve: Reference 'recorddate' not found in this buffer
>>>
>>> Is this supposed to work?
>>
>
> Yes, your example should work. From the "Evaluating code blocks"
> section of the Org-mode manual.
>
> ,----
> | It is also possible to evaluate named code blocks from anywhere in an
> | Org mode buffer or an Org mode table. Live code blocks located in the
> | current Org mode buffer or in the "Library of Babel" (see *note Library
> | of Babel::) can be executed. Named code blocks can be executed with a
> | separate '#+CALL:' line or inline within a block of text.
> `----
>
> There is no need to load code blocks in the same buffer into the library
> of babel.
I think what he (and I, when I tried his version) ran into is that
without a ":var" attribute on the named code block, you get
the "not found" error.
Ie, simply declaring #+name: recorddate(order="desc") isn't enough to
let the block know to expect the var.
Presumably that *should* be an error, as the variable should be declared
before being used, but the actual error message is a little misleading.
Eric
> This example works for me evaluating code blocks in the same buffer
> using call lines.
>
> #+Title: Call Example
>
> #+name: example-block
> #+begin_src sh :var input=""
> echo "input is $input"
> #+end_src
>
> Here's a simple call using a named argument.
> #+call: example-block(input="foo")
>
> #+RESULTS: example-block(input="foo")
> : input is foo
>
> It also works with a positional argument.
> #+call: example-block("bar")
>
> #+RESULTS: example-block("bar")
> : input is bar
>
> When I export this to e.g., html I get the following.
>
> Call Example
>
> Call Example
>
> echo "input is $input"
>
> Here's a simple call using a named argument.
>
>
> input is foo
>
> It also works with a positional argument.
>
>
> input is bar
>
> Date: 2013-04-01T19:39-0600
>
> Author:
>
> Org version 7.9.3f with Emacs version 24
>
> Validate XHTML 1.0
>
>
> The call lines are replaced with their results as part of the export
> process.
>
> If the above doesn't work for you, then I imagine something is wrong
> with your install.
