Split up parsing and filtering for --keep-within (#2726)

Split up parsing and filtering for --keep-within

Fixes #2610

Parse --keep-within argument early, via new method within_range passed
to argparse type=, so that better error messages can be given.

Also swallows ValueError stacktrace per the comment in the old code that
including it wasn't desirable.

borg/archiver.py

@@ -21,7 +21,7 @@ import collections
 from . import __version__
 from .helpers import Error, location_validator, archivename_validator, format_line, format_time, format_file_size, \
     parse_pattern, PathPrefixPattern, to_localtime, timestamp, safe_timestamp, bin_to_hex, \
-    get_cache_dir, prune_within, prune_split, check_python, \
+    get_cache_dir, interval, prune_within, prune_split, check_python, \
     Manifest, NoManifestError, remove_surrogates, update_excludes, format_archive, check_extension_modules, Statistics, \
     dir_is_tagged, bigint_to_int, ChunkerParams, CompressionSpec, PrefixSpec, is_slow_msgpack, yes, sysinfo, \
     EXIT_SUCCESS, EXIT_WARNING, EXIT_ERROR, log_multi, PatternMatcher, ErrorIgnoringTextIOWrapper, set_ec, \
@@ -1717,7 +1717,7 @@ class Archiver:
         subparser.add_argument('--list', dest='output_list',
                                action='store_true', default=False,
                                help='output verbose list of archives it keeps/prunes. Requires -v/--verbose.')
-        subparser.add_argument('--keep-within', dest='within', type=str, metavar='WITHIN',
+        subparser.add_argument('--keep-within', dest='within', type=interval, metavar='INTERVAL',
                                help='keep all archives within this time interval')
         subparser.add_argument('-H', '--keep-hourly', dest='hourly', type=int, default=0,
                                help='number of hourly archives to keep')

borg/helpers.py

@@ -279,15 +279,32 @@ class Manifest:
         return archives
 
 
-def prune_within(archives, within):
+def interval(s):
+    """Convert a string representing a valid interval to a number of hours."""
     multiplier = {'H': 1, 'd': 24, 'w': 24 * 7, 'm': 24 * 31, 'y': 24 * 365}
+
+    if s.endswith(tuple(multiplier.keys())):
+        number = s[:-1]
+        suffix = s[-1]
+    else:
+        # range suffixes in ascending multiplier order
+        ranges = [k for k, v in sorted(multiplier.items(), key=lambda t: t[1])]
+        raise argparse.ArgumentTypeError(
+            'Unexpected interval time unit "%s": expected one of %r' % (s[-1], ranges))
+
     try:
-        hours = int(within[:-1]) * multiplier[within[-1]]
-    except (KeyError, ValueError):
-        # I don't like how this displays the original exception too:
-        raise argparse.ArgumentTypeError('Unable to parse --keep-within option: "%s"' % within)
+        hours = int(number) * multiplier[suffix]
+    except ValueError:
+        hours = -1
+
     if hours <= 0:
-        raise argparse.ArgumentTypeError('Number specified using --keep-within option must be positive')
+        raise argparse.ArgumentTypeError(
+            'Unexpected interval number "%s": expected an integer greater than 0' % number)
+
+    return hours
+
+
+def prune_within(archives, hours):
     target = datetime.now(timezone.utc) - timedelta(seconds=hours * 3600)
     return [a for a in archives if a.ts > target]
 

borg/testsuite/helpers.py

@@ -1,4 +1,5 @@
 import hashlib
+from argparse import ArgumentTypeError
 from time import mktime, strptime
 from datetime import datetime, timezone, timedelta
 from io import StringIO
@@ -11,7 +12,7 @@ import msgpack.fallback
 import time
 
 from ..helpers import Location, format_file_size, format_timedelta, format_line, PlaceholderError, make_path_safe, \
-    prune_within, prune_split, get_cache_dir, get_keys_dir, get_security_dir, Statistics, is_slow_msgpack, \
+    interval, prune_within, prune_split, get_cache_dir, get_keys_dir, get_security_dir, Statistics, is_slow_msgpack, \
     yes, TRUISH, FALSISH, DEFAULTISH, \
     StableDict, int_to_bigint, bigint_to_int, parse_timestamp, CompressionSpec, ChunkerParams, \
     ProgressIndicatorPercent, ProgressIndicatorEndless, load_excludes, parse_pattern, \
@@ -664,16 +665,48 @@ class PruneSplitTestCase(BaseTestCase):
         dotest(test_archives, 0, [], [])
 
 
-class PruneWithinTestCase(BaseTestCase):
+class IntervalTestCase(BaseTestCase):
+    def test_interval(self):
+        self.assert_equal(interval('1H'), 1)
+        self.assert_equal(interval('1d'), 24)
+        self.assert_equal(interval('1w'), 168)
+        self.assert_equal(interval('1m'), 744)
+        self.assert_equal(interval('1y'), 8760)
 
-    def test(self):
+    def test_interval_time_unit(self):
+        with pytest.raises(ArgumentTypeError) as exc:
+            interval('H')
+        self.assert_equal(
+            exc.value.args,
+            ('Unexpected interval number "": expected an integer greater than 0',))
+        with pytest.raises(ArgumentTypeError) as exc:
+            interval('-1d')
+        self.assert_equal(
+            exc.value.args,
+            ('Unexpected interval number "-1": expected an integer greater than 0',))
+        with pytest.raises(ArgumentTypeError) as exc:
+            interval('food')
+        self.assert_equal(
+            exc.value.args,
+            ('Unexpected interval number "foo": expected an integer greater than 0',))
+
+    def test_interval_number(self):
+        with pytest.raises(ArgumentTypeError) as exc:
+            interval('5')
+        self.assert_equal(
+            exc.value.args,
+            ("Unexpected interval time unit \"5\": expected one of ['H', 'd', 'w', 'm', 'y']",))
+
+
+class PruneWithinTestCase(BaseTestCase):
+    def test_prune_within(self):
 
         def subset(lst, indices):
             return {lst[i] for i in indices}
 
         def dotest(test_archives, within, indices):
             for ta in test_archives, reversed(test_archives):
-                self.assert_equal(set(prune_within(ta, within)),
+                self.assert_equal(set(prune_within(ta, interval(within))),
                                   subset(test_archives, indices))
 
         # 1 minute, 1.5 hours, 2.5 hours, 3.5 hours, 25 hours, 49 hours