ref: 866176be971c460d05a7b3b09b1034eb9a25ec56
parent: 7705e32ec26b4b6602ca66152fa86cb35a04123f
author: Steve Moser <[email protected]>
date: Thu Mar 3 17:16:49 EST 2016
Docs: fix missing words, tense in datadrivencontent.md
--- a/docs/content/extras/datadrivencontent.md
+++ b/docs/content/extras/datadrivencontent.md
@@ -1,7 +1,7 @@
---
aliases:
- /doc/datadrivencontent/
-lastmod: 2015-07-23
+lastmod: 2016-03-03
date: 2015-02-14
menu:
main:
@@ -117,7 +117,7 @@
To load local files with the two functions `getJSON` and `getCSV`, the
source files must reside within Hugo's working directory. The file
-extension does not matter but the content.
+extension does not matter but the content does.
It applies the same output logic as in the topic: *Calling the functions with an URL*.
@@ -125,14 +125,14 @@
There is no chance to trigger a [LiveReload](/extras/livereload/) when
the content of an URL changes. However, when a local JSON/CSV file changes,
-then a LiveReload will be triggered of course. Symlinks not supported.
+then a LiveReload will be triggered of course. Symlinks are not supported.
**URLs and LiveReload**: If you change any local file and the LiveReload
-got triggered, Hugo will either read the URL content from the cache or, if
+is triggered, Hugo will either read the URL content from the cache or, if
you have disabled the cache, Hugo will re-download the content.
This can create huge traffic and you may also reach API limits quickly.
-As downloading of content takes a while, Hugo stops with processing
+As downloading of content takes a while, Hugo stops processing
your Markdown files until the content has been downloaded.
## Examples