The original text:
,----
| We define refs the same way we define atoms, but the two are used rather
differently.
| Let's take a quick look at how they work below.
|
| (def names (ref []))
|
| (dosync
| (ref-set names ["John"])
| (alter names #(if (not-empty %)
| (conj % "Jane") %)))
|
| In this code, we define a ref called names, then open a transaction using the
dosync statement. Inside the transaction we set names to a vector with the
value "John". Next, we call alter to check if names is not empty and add "Jane"
to the vector of the names if that's the case.
|
| Note that since this is happening inside a transaction, the check for
emptiness depends on the existing state along with any state built up within
the same transaction. If we tried to add or remove a name in a different
transaction, it would have no visible effect on ours. In case of a collision,
one of the transactions would end up being retried.
`----
Then I use region select code from (def to ))), then press [C-c C-, s].
It has weird offset like this:
,----
| #+begin_src
| (def names (ref []))
|
| (dosync
| (ref-set names ["John"])
| (alter names #(if (not-empty %)
| (conj % "Jane") %)))
| #+end_src
`----
But when I region select by backward from ))) to (def, it works fine. no offset.
--
[ stardiviner ] don't need to convince with trends.
Blog: https://stardiviner.github.io/
IRC(freenode): stardiviner
GPG: F09F650D7D674819892591401B5DF1C95AE89AC3
