From 2ffc0e8d1ba6085056ab8b5e1c403102b9d3736e Mon Sep 17 00:00:00 2001
From: Jules Dejaeghere <dej.j@live.be>
Date: Fri, 27 Nov 2020 18:21:17 +0100
Subject: [PATCH] Catch parsing error when caching a calendar

---
 app/tools/caching.py | 29 +++++++++++++++++++----------
 requirements.txt     |  3 ++-
 2 files changed, 21 insertions(+), 11 deletions(-)

diff --git a/app/tools/caching.py b/app/tools/caching.py
index 8c7502a..1f01a89 100644
--- a/app/tools/caching.py
+++ b/app/tools/caching.py
@@ -5,9 +5,11 @@ import threading
 import time
 from hashlib import sha256
 
+import traceback
 import arrow
 import requests
 from ics import Calendar
+from tatsu.exceptions import FailedParse
 
 
 def cache(entry: dict, scheduler: sched.scheduler = None) -> None:
@@ -27,18 +29,15 @@ def cache(entry: dict, scheduler: sched.scheduler = None) -> None:
     :type scheduler: sched.scheduler
     """
 
-    if not os.path.isdir('app/cache'):
-        os.mkdir('app/cache')
-
-    url = entry['url']
-    path = "app/cache/" + sha256(url.encode()).hexdigest() + ".ics"
-
     try:
+        if not os.path.isdir('app/cache'):
+            os.mkdir('app/cache')
+
+        url = entry['url']
+        path = "app/cache/" + sha256(url.encode()).hexdigest() + ".ics"
+
         r = requests.get(entry["url"], allow_redirects=True)
-    except Exception as e:
-        print(arrow.now().format("YYYY-MM-DD HH:mm:ss"), "Could not cache", entry)
-        print(e)
-    else:
+
         if "encoding" in entry:
             cal = Calendar(imports=r.content.decode(encoding=entry["encoding"]))
         else:
@@ -47,6 +46,16 @@ def cache(entry: dict, scheduler: sched.scheduler = None) -> None:
         cal = horodate(cal, 'Cached at')
         open(path, 'w').writelines(cal)
         print(arrow.now().format("YYYY-MM-DD HH:mm:ss"), "Cached", entry['name'])
+
+    except FailedParse:
+        print("Could not parse", entry['name'])
+
+    # Save stack trace when an unknown error occurs
+    except Exception as e:
+        with open("error " + arrow.now().format("YYYY-MM-DD HH:mm:ss")+".txt", 'w') as file:
+            file.write(arrow.now().format("YYYY-MM-DD HH:mm:ss") + "\nCould not cache : " + str(entry))
+            file.write(str(e))
+            file.write(str(traceback.format_exc()))
     finally:
         if scheduler is not None:
             delay = entry['cache'] if entry['cache'] > 0 else 10
diff --git a/requirements.txt b/requirements.txt
index a143602..c3b9d5a 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -2,4 +2,5 @@ requests~=2.22.0
 ics~=0.7
 pathvalidate~=2.3.0
 flask~=1.1.1
-arrow~=0.14.7
\ No newline at end of file
+arrow~=0.14.7
+tatsu~=4.4.0
\ No newline at end of file