[tor-commits] [arm/master] Rewriting state file parsing

[atagar at torproject.org]

commit 8e75f349eee89a9565020e66af4d2f59bd1d7582
Author: Damian Johnson <atagar at torproject.org>
Date:   Wed Sep 10 09:35:58 2014 -0700

    Rewriting state file parsing
    
    Revising the part of prepopulate_from_state() that parses our state file. Most
    interesting bit is that we were doing a dumb method of converting a 'UTC
    timestamp => unix time'. We used mktime() which converted to localtime, then
    converted back. The calendar.timegm() function, however, does exactly what we
    want, as described in the table at the top of...
    
      https://docs.python.org/3.3/library/time.html
---
 arm/graphing/bandwidth_stats.py |   64 +++++++++++++++------------------------
 1 file changed, 24 insertions(+), 40 deletions(-)

diff --git a/arm/graphing/bandwidth_stats.py b/arm/graphing/bandwidth_stats.py
index d2a0b33..a766435 100644
--- a/arm/graphing/bandwidth_stats.py
+++ b/arm/graphing/bandwidth_stats.py
@@ -3,6 +3,8 @@ Tracks bandwidth usage of the tor process, expanding to include accounting
 stats if they're set.
 """
 
+import calendar
+import os
 import time
 import curses
 
@@ -152,57 +154,44 @@ class BandwidthStats(graph_panel.GraphStats):
     elif uptime < (24 * 60 * 60):
       raise ValueError("insufficient uptime, tor must've been running for at least a day")
 
-    # get the user's data directory (usually '~/.tor')
+    # read the user's state file in their data directory (usually '~/.tor')
 
     data_dir = controller.get_conf('DataDirectory', None)
 
     if not data_dir:
       raise ValueError("unable to determine tor's data directory")
 
-    # attempt to open the state file
-
-    state_file_path = '%s%s/state' % (CONFIG['tor.chroot'], data_dir)
+    state_path = os.path.join(CONFIG['tor.chroot'] + data_dir, 'state')
 
     try:
-      with open(state_file_path) as state_file:
-        state_file_content = state_file.readlines()
+      with open(state_path) as state_file:
+        state_content = state_file.readlines()
     except IOError as exc:
-      raise ValueError('unable to read the state file at %s, %s' % (state_file_path, exc))
+      raise ValueError('unable to read the state file at %s, %s' % (state_path, exc))
 
-    # get the BWHistory entries (ordered oldest to newest) and number of
-    # intervals since last recorded
+    # We're interested in two types of entries from our state file...
+    #
+    # * BWHistory*Values - Comma separated list of bytes we read or wrote
+    #   during each fifteen minute period. The last value is an incremental
+    #   counter for our current period, so ignoring that.
+    #
+    # * BWHistory*Ends - When our last sampling was recorded, in UTC.
 
     bw_read_entries, bw_write_entries = None, None
     missing_read_entries, missing_write_entries = None, None
 
-    # converts from gmt to local with respect to DST
-
-    tz_offset = time.altzone if time.localtime()[8] else time.timezone
-
-    for line in state_file_content:
+    for line in state_content:
       line = line.strip()
 
-      # According to the rep_hist_update_state() function the BWHistory*Ends
-      # correspond to the start of the following sampling period. Also, the
-      # most recent values of BWHistory*Values appear to be an incremental
-      # counter for the current sampling period. Hence, offsets are added to
-      # account for both.
-
-      if line.startswith('BWHistoryReadValues'):
-        bw_read_entries = line[20:].split(',')
-        bw_read_entries = [int(entry) / 1024.0 / 900 for entry in bw_read_entries]
-        bw_read_entries.pop()
-      elif line.startswith('BWHistoryWriteValues'):
-        bw_write_entries = line[21:].split(',')
-        bw_write_entries = [int(entry) / 1024.0 / 900 for entry in bw_write_entries]
-        bw_write_entries.pop()
-      elif line.startswith('BWHistoryReadEnds'):
-        last_read_time = time.mktime(time.strptime(line[18:], '%Y-%m-%d %H:%M:%S')) - tz_offset
-        last_read_time -= 900
+      if line.startswith('BWHistoryReadValues '):
+        bw_read_entries = [int(entry) / 1024.0 / 900 for entry in line[20:].split(',')[:-1]]
+      elif line.startswith('BWHistoryWriteValues '):
+        bw_write_entries = [int(entry) / 1024.0 / 900 for entry in line[21:].split(',')[:-1]]
+      elif line.startswith('BWHistoryReadEnds '):
+        last_read_time = calendar.timegm(time.strptime(line[18:], '%Y-%m-%d %H:%M:%S')) - 900
         missing_read_entries = int((time.time() - last_read_time) / 900)
-      elif line.startswith('BWHistoryWriteEnds'):
-        last_write_time = time.mktime(time.strptime(line[19:], '%Y-%m-%d %H:%M:%S')) - tz_offset
-        last_write_time -= 900
+      elif line.startswith('BWHistoryWriteEnds '):
+        last_write_time = calendar.timegm(time.strptime(line[19:], '%Y-%m-%d %H:%M:%S')) - 900
         missing_write_entries = int((time.time() - last_write_time) / 900)
 
     if not bw_read_entries or not bw_write_entries or not last_read_time or not last_write_time:
@@ -445,12 +434,7 @@ class BandwidthStats(graph_panel.GraphStats):
     if end_interval:
       # converts from gmt to local with respect to DST
 
-      if time.localtime()[8]:
-        tz_offset = time.altzone
-      else:
-        tz_offset = time.timezone
-
-      sec = time.mktime(time.strptime(end_interval, '%Y-%m-%d %H:%M:%S')) - time.time() - tz_offset
+      sec = calendar.timegm(time.strptime(end_interval, '%Y-%m-%d %H:%M:%S')) - time.time()
 
       if CONFIG['features.graph.bw.accounting.isTimeLong']:
         queried['reset_time'] = ', '.join(str_tools.time_labels(sec, True))