find_compatibles_dts: Convert to a python class to increase performances

Signed-off-by: Maxime Chretien <maxime.chretien@bootlin.com>

find_compatible_dts.py

@@ -18,77 +18,48 @@
 #  You should have received a copy of the GNU Affero General Public License
 #  along with Elixir.  If not, see <http://www.gnu.org/licenses/>.
 
-from sys import argv
 import re
-import os
 from urllib import parse
-import query
+from query import decode
 
-usage_message = ("USAGE: find_compatible_dts.py <file> <family>\n"
+class FindCompatibleDTS:
-                 "file : The file you want to search in\n"
+    def __init__(self):
-                 "family : The type of file (C for .c/.cpp/.h/...\n"
+        # Compile regexes
-                 "                           D for .dts/.dtsi\n"
+        self.regex_c = re.compile("\s*{*\s*\.compatible\s*=\s*\"(.+?)\"")
-                 "                           B for bindings docs files)")
+        self.regex_dts1 = re.compile("\s*compatible")
+        self.regex_dts2 = re.compile("\"(.+?)\"")
+        self.regex_bindings = re.compile("([\w-]+,?[\w-]+)")
 
-ident_list = ""
+    def parse_c(self, content):
+        return self.regex_c.findall(content)
 
-# Compile regexes
+    def parse_dts(self, content):
-regex_c = re.compile("\s*{*\s*\.compatible\s*=\s*\"(.+?)\"")
+        ret = []
-regex_dts1 = re.compile("\s*compatible")
+        if self.regex_dts1.match(content) != None:
-regex_dts2 = re.compile("\"(.+?)\"")
+            ret = self.regex_dts2.findall(content)
-regex_bindings = re.compile("([\w-]+,?[\w-]+)")
+        return ret
 
+    def parse_bindings(self, content):
+        # There are a lot of wrong results
+        # but we don't apply that to a lot of files
+        # so it should be fine
+        return self.regex_bindings.findall(content)
 
-def parse_c(content):
+    def run(self, file_lines, family):
-    return regex_c.findall(content)
+        ident_list = ""
 
-def parse_dts(content):
+        # Iterate though lines and search for idents
-    ret = []
+        for num, line in enumerate(file_lines, 1):
-    if regex_dts1.match(content) != None:
+            line = query.decode(line)
-        ret = regex_dts2.findall(content)
+            if family == 'C':
-    return ret
+                ret = self.parse_c(line)
+            elif family == 'D':
+                ret = self.parse_dts(line)
+            elif family == 'B':
+                ret = self.parse_bindings(line)
 
-def parse_bindings(content):
+            for i in range(len(ret)):
-    # There are a lot of wrong results
+                ident_list += str(parse.quote(ret[i])) + ' ' + str(num) + '\n'
-    # but we don't apply that to a lot of files
-    # so it should be fine
-    return regex_bindings.findall(content)
 
+        return ident_list
 
-# Main
-# Test and get args
-if len(argv) < 3:
-    print("ERROR: Missing arguments !\n" + usage_message)
-    exit(1)
-
-filename = argv[1]
-family = argv[2]
-
-# Make sure it's an accepted family
-if not family in ['C', 'D', 'B']:
-    print("ERROR: Unknown family !\n" + usage_message)
-    exit(1)
-
-# Make sure file exists
-try:
-    f = open(filename, 'rb')
-except IOError:
-    print("ERROR: File doesn't exist !\n" + usage_message)
-    exit(1)
-
-# Iterate though lines and search for idents
-for num, line in enumerate(f, 1):
-    line = query.decode(line)
-    if family == 'C':
-        ret = parse_c(line)
-    elif family == 'D':
-        ret = parse_dts(line)
-    elif family == 'B':
-        ret = parse_bindings(line)
-
-    for i in range(len(ret)):
-        ident_list += str(parse.quote(ret[i])) + ' ' + str(num) + '\n'
-
-# Print the list and exit
-print(ident_list, end='')
-exit(0)

script.sh

@@ -208,16 +208,6 @@ parse_docs()
     rm -rf "$tmpfile"
 }
 
-parse_comps()
-{
-    tmpfile=`mktemp`
-
-    git cat-file blob "$opt1" > "$tmpfile"
-    "$script_dir/find_compatible_dts.py" "$tmpfile" "$opt3" || exit "$?"
-
-    rm -rf "$tmpfile"
-}
-
 dts_comp()
 {
     echo $dts_comp_support
@@ -296,10 +286,6 @@ case $cmd in
         parse_docs
         ;;
 
-    parse-comps)
-        parse_comps
-        ;;
-
     dts-comp)
         dts_comp
         ;;

update.py

@@ -23,17 +23,21 @@
 # This is different from that blob's Git hash.
 
 from sys import argv
-from lib import script, scriptLines
-import lib
-import data
 import os
-from data import PathList
 from threading import Thread, Lock, Event, Condition
 
+import lib
+from lib import script, scriptLines
+import data
+from data import PathList
+from find_compatible_dts import FindCompatibleDTS
+
 verbose = False
 
 dts_comp_support = int(script('dts-comp'))
 
+compatibles_parser = FindCompatibleDTS()
+
 db = data.DB(lib.getDataDir(), readonly=False, dtscomp=dts_comp_support)
 
 # Number of cpu threads (+2 for version indexing)
@@ -407,7 +411,7 @@ class UpdateComps(Thread):
             family = lib.getFileFamily(filename)
             if family in [None, 'K']: continue
 
-            lines = scriptLines('parse-comps', hash, filename, family)
+            lines = compatibles_parser.run(scriptLines('get-blob', hash), family)
             comps = {}
             for l in lines:
                 ident, line = l.split(b' ')
@@ -472,7 +476,7 @@ class UpdateCompsDocs(Thread):
                 filename = db.file.get(idx)
 
             family = 'B'
-            lines = scriptLines('parse-comps', hash, filename, family)
+            lines = compatibles_parser.run(scriptLines('get-blob', hash), family)
             comps_docs = {}
             with comps_lock:
                 for l in lines: