[yelp-tools] Adding --skip option to list filenames to skip

[Shaun McCance <shaunm src gnome org>]

commit 3f0c73b8215753fbb578f870644517a9e22c3d63
Author: Shaun McCance <shaunm redhat com>
Date:   Sun Feb 7 19:50:24 2021 -0500

    Adding --skip option to list filenames to skip
    
    This is mostly useful in config files, where you have specific pages that
    you might want exempted from some custom checker, but you don't want to
    force people to explicitly list which filenames to test.

 tools/yelp-check.in | 39 ++++++++++++++++++++++++++++-----------
 1 file changed, 28 insertions(+), 11 deletions(-)
---
diff --git a/tools/yelp-check.in b/tools/yelp-check.in
index 67153e56..02f17771 100755
--- a/tools/yelp-check.in
+++ b/tools/yelp-check.in
@@ -86,6 +86,7 @@ class Checker:
         self.options = {}
         self.fileargs = []
         self.tmpdir = None
+        self.skipfiles = []
 
     def __del__(self):
         if self.tmpdir is not None:
@@ -140,6 +141,7 @@ class Checker:
             except Exception as e:
                 print(e, file=sys.stderr)
                 sys.exit(1)
+        self.skipfiles = self.get_option_list('skip') or []
         return 0
 
     def get_option_bool(self, arg):
@@ -197,12 +199,16 @@ class Checker:
         if len(self.fileargs) == 0:
             self.fileargs.append('.')
         for filearg in self.fileargs:
+            if filearg in self.skipfiles:
+                continue
             if os.path.isdir(filearg):
                 if issite:
                     for infile in self.iter_site(filearg, '/'):
                         yield infile
                 else:
                     for fname in os.listdir(filearg):
+                        if fname in self.skipfiles:
+                            continue
                         if fname.endswith('.page'):
                             yield InputFile(filearg, fname)
             else:
@@ -275,7 +281,8 @@ class HrefsChecker (Checker):
     arguments = [
         ('help', '-h', None, 'Show this help and exit'),
         ('site', '-s', None, 'Treat pages as belonging to a Mallard site'),
-        ('allow', None, 'URL', 'Allow URL or list of URLs without checking')
+        ('allow', None, 'URL', 'Allow URL or list of URLs without checking'),
+        ('skip', None, 'FILES', 'List of files to skip')
     ]
     postblurb = 'URL may be a comma- and/or space-separated list, or specified\nmultiple times.'
 
@@ -333,7 +340,8 @@ class IdsChecker (Checker):
     formats = ['mallard']
     arguments = [
         ('help', '-h', None, 'Show this help and exit'),
-        ('site', '-s', None, 'Treat pages as belonging to a Mallard site')
+        ('site', '-s', None, 'Treat pages as belonging to a Mallard site'),
+        ('skip', None, 'FILES', 'List of files to skip')
     ]
 
     def main(self, args):
@@ -372,7 +380,8 @@ class LinksChecker (Checker):
         ('help', '-h', None, 'Show this help and exit'),
         ('site', '-s', None, 'Treat pages as belonging to a Mallard site'),
         ('cache', '-c', 'CACHE', 'Use the existing Mallard cache CACHE'),
-        ('ignore', '-i', None, 'Ignore xrefs where href is present')
+        ('ignore', '-i', None, 'Ignore xrefs where href is present'),
+        ('skip', None, 'FILES', 'List of files to skip')
     ]
 
     def __init__(self, yelpcheck):
@@ -477,7 +486,8 @@ class MediaChecker (Checker):
     formats = ['docbook4', 'docbook5', 'mallard']
     arguments = [
         ('help', '-h', None, 'Show this help and exit'),
-        ('site', '-s', None, 'Treat pages as belonging to a Mallard site')
+        ('site', '-s', None, 'Treat pages as belonging to a Mallard site'),
+        ('skip', None, 'FILES', 'List of files to skip')
     ]
 
     def main(self, args):
@@ -524,7 +534,8 @@ class OrphansChecker (Checker):
     arguments = [
         ('help', '-h', None, 'Show this help and exit'),
         ('site', '-s', None, 'Treat pages as belonging to a Mallard site'),
-        ('cache', '-c', 'CACHE', 'Use the existing Mallard cache CACHE')
+        ('cache', '-c', 'CACHE', 'Use the existing Mallard cache CACHE'),
+        ('skip', None, 'FILES', 'List of files to skip')
     ]
 
     def __init__(self, yelpcheck):
@@ -656,7 +667,8 @@ class ValidateChecker (Checker):
         ('site', '-s', None, 'Treat pages as belonging to a Mallard site'),
         ('strict', None, None, 'Disallow unknown namespaces'),
         ('allow', None, 'NS', 'Explicitly allow namespace NS in strict mode'),
-        ('jing', None, None, 'Use jing instead of xmllint for RNG validation')
+        ('jing', None, None, 'Use jing instead of xmllint for RNG validation'),
+        ('skip', None, 'FILES', 'List of files to skip')
     ]
     postblurb = 'NS may be a comma- and/or space-separated list, or specified\nmultiple times.'
 
@@ -774,7 +786,8 @@ class CommentsChecker (Checker):
     formats = ['docbook4', 'docbook5', 'mallard']
     arguments = [
         ('help', '-h', None, 'Show this help and exit'),
-        ('site', '-s', None, 'Treat pages as belonging to a Mallard site')
+        ('site', '-s', None, 'Treat pages as belonging to a Mallard site'),
+        ('skip', None, 'FILES', 'List of files to skip')
     ]
 
     def main(self, args):
@@ -866,7 +879,8 @@ class LicenseChecker (Checker):
         ('site', '-s', None, 'Treat pages as belonging to a Mallard site'),
         ('only', None, 'LICENSES', 'Only show pages whose license is in LICENSES'),
         ('except', None, 'LICENSES', 'Exclude pages whose license is in LICENSES'),
-        ('totals', None, None, 'Show total counts for each license')
+        ('totals', None, None, 'Show total counts for each license'),
+        ('skip', None, 'FILES', 'List of files to skip')
     ]
     postblurb = 'LICENSES may be a comma- and/or space-separated list, or specified\nmultiple times.'
 
@@ -948,7 +962,8 @@ class StatusChecker (Checker):
         ('newer',  None, 'DATE', 'Only show pages newer than DATE'),
         ('only',   None, 'STATUSES', 'Only show pages whose status is in STATUSES'),
         ('except', None, 'STATUSES', 'Exclude pages whose status is in STATUSES'),
-        ('totals', None, None, 'Show total counts for each status')
+        ('totals', None, None, 'Show total counts for each status'),
+        ('skip', None, 'FILES', 'List of files to skip')
     ]
     postblurb = 'VER and STATUSES may be comma- and/or space-separated lists, or specified\nmultiple times.'
 
@@ -1052,7 +1067,8 @@ class StyleChecker (Checker):
         ('site', '-s', None, 'Treat pages as belonging to a Mallard site'),
         ('only',   None, 'STYLES', 'Only show pages whose style is in STATUSES'),
         ('except', None, 'STYLES', 'Exclude pages whose style is in STATUSES'),
-        ('totals', None, None, 'Show total counts for each style')
+        ('totals', None, None, 'Show total counts for each style'),
+        ('skip', None, 'FILES', 'List of files to skip')
     ]
     postblurb = 'STYLES may be comma- and/or space-separated lists, or specified\nmultiple times.'
 
@@ -1113,7 +1129,8 @@ class CustomChecker(Checker):
     formats = ['docbook4', 'docbook5', 'mallard']
     arguments = [
         ('help', '-h', None, 'Show this help and exit'),
-        ('site', '-s', None, 'Treat pages as belonging to a Mallard site')
+        ('site', '-s', None, 'Treat pages as belonging to a Mallard site'),
+        ('skip', None, 'FILES', 'List of files to skip')
     ]
 
     def __init__(self, name, yelpcheck):