]> git.decadent.org.uk Git - dak.git/blobdiff - dak/generate_index_diffs.py
Merge remote-tracking branch 'nthykier/auto-decruft'
[dak.git] / dak / generate_index_diffs.py
index 992edf9c65314b088d4e6e3675cff22df8f11621..6fe9436a72a75c2a8344063c06863466a7a07210 100755 (executable)
@@ -1,7 +1,8 @@
 #!/usr/bin/env python
 
+""" generates partial package updates list"""
+
 ###########################################################
-# generates partial package updates list
 
 # idea and basic implementation by Anthony, some changes by Andreas
 # parts are stolen from 'dak generate-releases'
 
 ################################################################################
 
-import sys, os, tempfile
+import sys
+import os
+import tempfile
+import time
 import apt_pkg
-import daklib.utils
+import glob
+
+from daklib import utils
+from daklib.dbconn import Archive, Component, DBConn, Suite, get_suite, get_suite_architectures
+#from daklib.regexes import re_includeinpdiff
+import re
+re_includeinpdiff = re.compile(r"(Translation-[a-zA-Z_]+\.(?:bz2|xz))")
 
 ################################################################################
 
-projectB = None
 Cnf = None
 Logger = None
 Options = None
@@ -48,13 +57,15 @@ def usage (exit_code=0):
 Write out ed-style diffs to Packages/Source lists
 
   -h, --help            show this help and exit
+  -a <archive>          generate diffs for suites in <archive>
   -c                    give the canonical path of the file
   -p                    name for the patch (defaults to current time)
+  -d                    name for the hardlink farm for status
+  -m                    how many diffs to generate
   -n                    take no action
     """
     sys.exit(exit_code)
 
-
 def tryunlink(file):
     try:
         os.unlink(file)
@@ -76,7 +87,7 @@ def smartlink(f, t):
         os.system("bzip2 -d < %s.bz2 > %s" % (f, t))
     else:
         print "missing: %s" % (f)
-        raise IOError, f
+        raise IOError(f)
 
 def smartopen(file):
     if os.path.isfile(file):
@@ -98,7 +109,7 @@ def pipe_file(f, t):
     t.close()
 
 class Updates:
-    def __init__(self, readpath = None, max = 14):
+    def __init__(self, readpath = None, max = 56):
         self.can_path = None
         self.history = {}
         self.history_order = []
@@ -107,51 +118,51 @@ class Updates:
         self.filesizesha1 = None
 
         if readpath:
-          try:
-            f = open(readpath + "/Index")
-            x = f.readline()
+            try:
+                f = open(readpath + "/Index")
+                x = f.readline()
 
-            def read_hashs(ind, f, self, x=x):
-                while 1:
-                    x = f.readline()
-                    if not x or x[0] != " ": break
+                def read_hashs(ind, f, self, x=x):
+                    while 1:
+                        x = f.readline()
+                        if not x or x[0] != " ": break
+                        l = x.split()
+                        if not self.history.has_key(l[2]):
+                            self.history[l[2]] = [None,None]
+                            self.history_order.append(l[2])
+                        self.history[l[2]][ind] = (l[0], int(l[1]))
+                    return x
+
+                while x:
                     l = x.split()
-                    if not self.history.has_key(l[2]):
-                        self.history[l[2]] = [None,None]
-                       self.history_order.append(l[2])
-                    self.history[l[2]][ind] = (l[0], int(l[1]))
-                return x
-
-            while x:
-                l = x.split()
 
-                if len(l) == 0:
-                    x = f.readline()
-                    continue
+                    if len(l) == 0:
+                        x = f.readline()
+                        continue
 
-                if l[0] == "SHA1-History:":
-                    x = read_hashs(0,f,self)
-                    continue
+                    if l[0] == "SHA1-History:":
+                        x = read_hashs(0,f,self)
+                        continue
 
-                if l[0] == "SHA1-Patches:":
-                    x = read_hashs(1,f,self)
-                    continue
+                    if l[0] == "SHA1-Patches:":
+                        x = read_hashs(1,f,self)
+                        continue
 
-                if l[0] == "Canonical-Name:" or l[0]=="Canonical-Path:":
-                    self.can_path = l[1]
+                    if l[0] == "Canonical-Name:" or l[0]=="Canonical-Path:":
+                        self.can_path = l[1]
 
-                if l[0] == "SHA1-Current:" and len(l) == 3:
-                    self.filesizesha1 = (l[1], int(l[2]))
+                    if l[0] == "SHA1-Current:" and len(l) == 3:
+                        self.filesizesha1 = (l[1], int(l[2]))
 
-                x = f.readline()
+                    x = f.readline()
 
-          except IOError:
-            0
+            except IOError:
+                0
 
     def dump(self, out=sys.stdout):
         if self.can_path:
             out.write("Canonical-Path: %s\n" % (self.can_path))
-        
+
         if self.filesizesha1:
             out.write("SHA1-Current: %s %7d\n" % (self.filesizesha1))
 
@@ -164,7 +175,7 @@ class Updates:
                 tryunlink("%s/%s.gz" % (self.readpath, h))
                 del hs[h]
             l = l[cnt-self.max:]
-           self.history_order = l[:]
+            self.history_order = l[:]
 
         out.write("SHA1-History:\n")
         for h in l:
@@ -191,8 +202,9 @@ def sizesha1(f):
     sha1sum = apt_pkg.sha1sum(f)
     return (sha1sum, size)
 
-def genchanges(Options, outdir, oldfile, origfile, maxdiffs = 14):
-    if Options.has_key("NoAct"): 
+def genchanges(Options, outdir, oldfile, origfile, maxdiffs = 56):
+    if Options.has_key("NoAct"):
+        print "Not acting on: od: %s, oldf: %s, origf: %s, md: %s" % (outdir, oldfile, origfile, maxdiffs)
         return
 
     patchname = Options["PatchName"]
@@ -220,7 +232,7 @@ def genchanges(Options, outdir, oldfile, origfile, maxdiffs = 14):
         return
 
     if oldstat[1:3] == origstat[1:3]:
-        print "%s: hardlink unbroken, assuming unchanged" % (origfile)
+        #print "%s: hardlink unbroken, assuming unchanged" % (origfile)
         return
 
     oldf = smartopen(oldfile)
@@ -229,20 +241,9 @@ def genchanges(Options, outdir, oldfile, origfile, maxdiffs = 14):
     # should probably early exit if either of these checks fail
     # alternatively (optionally?) could just trim the patch history
 
-    if upd.filesizesha1:
-        if upd.filesizesha1 != oldsizesha1:
-            print "warning: old file seems to have changed! %s %s => %s %s" % (upd.filesizesha1 + oldsizesha1)
-
-    # XXX this should be usable now
-    #
-    #for d in upd.history.keys():
-    #    df = smartopen("%s/%s" % (outdir,d))
-    #    act_sha1size = sizesha1(df)
-    #    df.close()
-    #    exp_sha1size = upd.history[d][1]
-    #    if act_sha1size != exp_sha1size:
-    #        print "patch file %s seems to have changed! %s %s => %s %s" % \
-    #            (d,) + exp_sha1size + act_sha1size
+    #if upd.filesizesha1:
+    #    if upd.filesizesha1 != oldsizesha1:
+    #        print "info: old file " + oldfile + " changed! %s %s => %s %s" % (upd.filesizesha1 + oldsizesha1)
 
     if Options.has_key("CanonicalPath"): upd.can_path=Options["CanonicalPath"]
 
@@ -255,11 +256,13 @@ def genchanges(Options, outdir, oldfile, origfile, maxdiffs = 14):
     if newsizesha1 == oldsizesha1:
         os.unlink(newfile)
         oldf.close()
-        print "%s: unchanged" % (origfile)
+        #print "%s: unchanged" % (origfile)
     else:
-        if not os.path.isdir(outdir): os.mkdir(outdir)
-        w = os.popen("diff --ed - %s | gzip -c -9 > %s.gz" % 
-                         (newfile, difffile), "w")
+        if not os.path.isdir(outdir):
+            os.mkdir(outdir)
+
+        w = os.popen("diff --ed - %s | gzip --rsyncable -c -9 > %s.gz" %
+                     (newfile, difffile), "w")
         pipe_file(oldf, w)
         oldf.close()
 
@@ -284,92 +287,85 @@ def genchanges(Options, outdir, oldfile, origfile, maxdiffs = 14):
 def main():
     global Cnf, Options, Logger
 
-    os.umask(0002)
+    os.umask(0o002)
 
-    Cnf = daklib.utils.get_conf()
+    Cnf = utils.get_conf()
     Arguments = [ ('h', "help", "Generate-Index-Diffs::Options::Help"),
+                  ('a', 'archive', 'Generate-Index-Diffs::Options::Archive', 'hasArg'),
                   ('c', None, "Generate-Index-Diffs::Options::CanonicalPath", "hasArg"),
                   ('p', "patchname", "Generate-Index-Diffs::Options::PatchName", "hasArg"),
-                  ('r', "rootdir", "Generate-Index-Diffs::Options::RootDir", "hasArg"),
                   ('d', "tmpdir", "Generate-Index-Diffs::Options::TempDir", "hasArg"),
                   ('m', "maxdiffs", "Generate-Index-Diffs::Options::MaxDiffs", "hasArg"),
-                 ('n', "n-act", "Generate-Index-Diffs::Options::NoAct"),
+                  ('n', "n-act", "Generate-Index-Diffs::Options::NoAct"),
                 ]
-    suites = apt_pkg.ParseCommandLine(Cnf,Arguments,sys.argv)
-    Options = Cnf.SubTree("Generate-Index-Diffs::Options")
+    suites = apt_pkg.parse_commandline(Cnf,Arguments,sys.argv)
+    Options = Cnf.subtree("Generate-Index-Diffs::Options")
     if Options.has_key("Help"): usage()
 
-    maxdiffs = Options.get("MaxDiffs::Default", "14")
+    maxdiffs = Options.get("MaxDiffs::Default", "56")
     maxpackages = Options.get("MaxDiffs::Packages", maxdiffs)
     maxcontents = Options.get("MaxDiffs::Contents", maxdiffs)
     maxsources = Options.get("MaxDiffs::Sources", maxdiffs)
 
     if not Options.has_key("PatchName"):
         format = "%Y-%m-%d-%H%M.%S"
-        i,o = os.popen2("date +%s" % (format))
-        i.close()
-        Options["PatchName"] = o.readline()[:-1]
-        o.close()
-
-    AptCnf = apt_pkg.newConfiguration()
-    apt_pkg.ReadConfigFileISC(AptCnf,daklib.utils.which_apt_conf_file())
+        Options["PatchName"] = time.strftime( format )
 
-    if Options.has_key("RootDir"): Cnf["Dir::Root"] = Options["RootDir"]
+    session = DBConn().session()
 
     if not suites:
-        suites = Cnf.SubTree("Suite").List()
+        query = session.query(Suite.suite_name)
+        if Options.get('Archive'):
+            query = query.join(Suite.archive).filter(Archive.archive_name == Options['Archive'])
+        suites = [ s.suite_name for s in query ]
 
-    for suite in suites:
-        print "Processing: " + suite
-        SuiteBlock = Cnf.SubTree("Suite::" + suite)
+    for suitename in suites:
+        print "Processing: " + suitename
 
-        if SuiteBlock.has_key("Untouchable"):
-            print "Skipping: " + suite + " (untouchable)"
-            continue
+        suiteobj = get_suite(suitename.lower(), session=session)
 
-        suite = suite.lower()
+        # Use the canonical version of the suite name
+        suite = suiteobj.suite_name
 
-        architectures = SuiteBlock.ValueList("Architectures")
+        if suiteobj.untouchable:
+            print "Skipping: " + suite + " (untouchable)"
+            continue
 
-        if SuiteBlock.has_key("Components"):
-            components = SuiteBlock.ValueList("Components")
-        else:
-            components = []
+        architectures = get_suite_architectures(suite, skipall=True, session=session)
+        components = [ c.component_name for c in session.query(Component.component_name) ]
 
-        suite_suffix = Cnf.Find("Dinstall::SuiteSuffix")
+        suite_suffix = Cnf.find("Dinstall::SuiteSuffix")
         if components and suite_suffix:
             longsuite = suite + "/" + suite_suffix
         else:
             longsuite = suite
 
-        tree = SuiteBlock.get("Tree", "dists/%s" % (longsuite))
-
-        if AptCnf.has_key("tree::%s" % (tree)):
-            sections = AptCnf["tree::%s::Sections" % (tree)].split()
-        elif AptCnf.has_key("bindirectory::%s" % (tree)):
-            sections = AptCnf["bindirectory::%s::Sections" % (tree)].split()
-        else:
-            aptcnf_filename = os.path.basename(daklib.utils.which_apt_conf_file())
-            print "ALERT: suite %s not in %s, nor untouchable!" % (suite, aptcnf_filename)
-            continue
-
-        for architecture in architectures:
-            if architecture == "all":
-                continue
-
-            if architecture != "source":
-                # Process Contents
-                file = "%s/Contents-%s" % (Cnf["Dir::Root"] + tree,
-                        architecture)
-                storename = "%s/%s_contents_%s" % (Options["TempDir"], suite, architecture)
-                genchanges(Options, file + ".diff", storename, file, \
-                  Cnf.get("Suite::%s::Generate-Index-Diffs::MaxDiffs::Contents" % (suite), maxcontents))
-
-            # use sections instead of components since dak.conf
-            # treats "foo/bar main" as suite "foo", suitesuffix "bar" and
-            # component "bar/main". suck.
-
-            for component in sections:
+        tree = os.path.join(suiteobj.archive.path, 'dists', longsuite)
+
+        # See if there are Translations which might need a new pdiff
+        cwd = os.getcwd()
+        for component in components:
+            #print "DEBUG: Working on %s" % (component)
+            workpath=os.path.join(tree, component, "i18n")
+            if os.path.isdir(workpath):
+                os.chdir(workpath)
+                for dirpath, dirnames, filenames in os.walk(".", followlinks=True, topdown=True):
+                    for entry in filenames:
+                        if not re_includeinpdiff.match(entry):
+                            #print "EXCLUDING %s" % (entry)
+                            continue
+                        (fname, fext) = os.path.splitext(entry)
+                        processfile=os.path.join(workpath, fname)
+                        #print "Working: %s" % (processfile)
+                        storename="%s/%s_%s_%s" % (Options["TempDir"], suite, component, fname)
+                        #print "Storefile: %s" % (storename)
+                        genchanges(Options, processfile + ".diff", storename, processfile, maxdiffs)
+        os.chdir(cwd)
+
+        for archobj in architectures:
+            architecture = archobj.arch_string
+
+            for component in components:
                 if architecture == "source":
                     longarch = architecture
                     packages = "Sources"
@@ -378,12 +374,14 @@ def main():
                     longarch = "binary-%s"% (architecture)
                     packages = "Packages"
                     maxsuite = maxpackages
+                    # Process Contents
+                    file = "%s/%s/Contents-%s" % (tree, component, architecture)
+                    storename = "%s/%s_%s_contents_%s" % (Options["TempDir"], suite, component, architecture)
+                    genchanges(Options, file + ".diff", storename, file, maxcontents)
 
-                file = "%s/%s/%s/%s" % (Cnf["Dir::Root"] + tree,
-                           component, longarch, packages)
+                file = "%s/%s/%s/%s" % (tree, component, longarch, packages)
                 storename = "%s/%s_%s_%s" % (Options["TempDir"], suite, component, architecture)
-                genchanges(Options, file + ".diff", storename, file, \
-                  Cnf.get("Suite::%s::Generate-Index-Diffs::MaxDiffs::%s" % (suite, packages), maxsuite))
+                genchanges(Options, file + ".diff", storename, file, maxsuite)
 
 ################################################################################