Quick and dirty fix to test whether scrapy crawl fails.

`scrapy crawl` command does not return any exit status depending on the
errors from the spider. I have to simulate this by checking whether
there is any instance of 'log_count/ERROR' in the log.
This commit will deliberately fail in Travis

.travis.yml

@@ -2,4 +2,4 @@ language: python
 python:
   - "2.7"
 install: pip install scrapy py-dateutil
-script: scrapy crawl diavgeia_spider
+script: scrapy crawl diavgeia_spider --logfile=test && [[ $(grep -c 'log_count/ERROR' test) == 0 ]] && cat test && rm test

diavgeia/settings.py

@@ -10,7 +10,7 @@
 SPIDER_MODULES = ['diavgeia.spiders']
 NEWSPIDER_MODULE = 'diavgeia.spiders'
 ITEM_PIPELINES = { 
-    'diavgeia.pipelines.DownloaderPipeline': 100
+#    'diavgeia.pipelines.DownloaderPipeline': 100
 }
 
 # DownloaderPipeline settings