Python handlers: factorise tmp dir code

src/filters/rclexecm.py

@@ -265,19 +265,30 @@ def execPythonScript(icmd):
     
 # Temp dir helper
 class SafeTmpDir:
-    def __init__(self, em):
+    def __init__(self, tag, em=None):
+        self.tag = tag
         self.em = em
-        self.toptmp = ""
+        self.toptmp = None
-        self.tmpdir = ""
+        self.tmpdir = None
 
     def __del__(self):
-        try:
+        if self.toptmp:
-            if self.toptmp:
+            try:
-                shutil.rmtree(self.tmpdir, True)
+                if self.tmpdir:
+                    shutil.rmtree(self.tmpdir, True)
                 os.rmdir(self.toptmp)
-        except Exception as err:
+            except Exception as err:
-            self.em.rclog("delete dir failed for " + self.toptmp)
+                if self.em:
+                    self.em.rclog("delete dir failed for " + self.toptmp)
 
+    def vacuumdir(self):
+        if self.tmpdir:
+            for fn in os.listdir(self.tmpdir):
+                path = os.path.join(self.tmpdir, fn)
+                if os.path.isfile(path):
+                    os.unlink(path)
+        return True
+    
     def getpath(self):
         if not self.tmpdir:
             envrcltmp = os.getenv('RECOLL_TMPDIR')
@@ -286,7 +297,7 @@ class SafeTmpDir:
             else:
                 self.toptmp = tempfile.mkdtemp(prefix='rcltmp')
 
-            self.tmpdir = os.path.join(self.toptmp, 'rclsofftmp')
+            self.tmpdir = os.path.join(self.toptmp, self.tag)
             os.makedirs(self.tmpdir)
 
         return self.tmpdir

src/filters/rclocrtesseract.py

@@ -42,33 +42,25 @@ pdftoppmcmd = None
 def _deb(s):
     rclexecm.logmsg("rclocrtesseract: %s" % s)
 
-def vacuumdir(dir):
-    if dir:
-        for fn in os.listdir(dir):
-            path = os.path.join(dir, fn)
-            if os.path.isfile(path):
-                os.unlink(path)
-    return True
-
-
 tmpdir = None
+
 def _maybemaketmpdir():
     global tmpdir
     if tmpdir:
-        if not vacuumdir(tmpdir):
+        if not tmpdir.vacuumdir():
-            _deb("openfile: vacuumdir %s failed" % tmpdir)
+            _deb("openfile: vacuumdir %s failed" % tmpdir.getpath())
             return False
     else:
-        tmpdir = tempfile.mkdtemp(prefix='rclocrtmp')
+        tmpdir = rclexecm.SafeTmpDir("rclocrtesseract")
 
 
 def cleanocr():
     global tmpdir
     if tmpdir:
-        vacuumdir(tmpdir)
+        del tmpdir
-        os.rmdir(tmpdir)
         tmpdir = None
-        
+
+
 # Return true if tesseract and the appropriate conversion program for
 # the file type (e.g. pdftoppt for pdf) appear to be available
 def ocrpossible(config, path):
@@ -165,12 +157,12 @@ def _pdftesseract(config, path):
 
     tesseractlang = _guesstesseractlang(config, path)
 
-    #tesserrorfile = os.path.join(tmpdir, "tesserrorfile")
+    #tesserrorfile = os.path.join(tmpdir.getpath(), "tesserrorfile")
-    tmpfile = os.path.join(tmpdir, "ocrXXXXXX")
+    tmpfile = os.path.join(tmpdir.getpath(), "ocrXXXXXX")
 
     # Split pdf pages
     try:
-        vacuumdir(tmpdir)
+        tmpdir.vacuumdir()
         cmd = [pdftoppmcmd, "-r", "300", path, tmpfile]
         #_deb("Executing %s" % cmd)
         subprocess.check_call(cmd)

src/filters/rclpdf.py

@@ -69,8 +69,7 @@ _htmlsuffix = b'''</pre></body></html>'''
 def finalcleanup():
     global tmpdir
     if tmpdir:
-        vacuumdir(tmpdir)
+        del tmpdir
-        os.rmdir(tmpdir)
         tmpdir = None
 
 ocrproc = None
@@ -93,14 +92,6 @@ except: pass
 try: signal.signal(signal.SIGTERM, signal_handler)
 except: pass
 
-def vacuumdir(dir):
-    if dir:
-        for fn in os.listdir(dir):
-            path = os.path.join(dir, fn)
-            if os.path.isfile(path):
-                os.unlink(path)
-    return True
-
 class PDFExtractor:
     def __init__(self, em):
         self.currentindex = 0
@@ -221,7 +212,7 @@ class PDFExtractor:
             # no big deal
             return True
         try:
-            vacuumdir(tmpdir)
+            tmpdir.vacuumdir()
             # Note: the java version of pdftk sometimes/often fails
             # here with writing to stdout:
             #    Error occurred during initialization of VM
@@ -231,9 +222,9 @@ class PDFExtractor:
             # output, until we fix the error or preferably find a way
             # to do it with poppler...
             subprocess.check_call(
-                [self.pdftk, self.filename, "unpack_files", "output",
+                [self.pdftk, self.filename, "unpack_files", "output", tmpdir.getpath()],
-                 tmpdir], stdout=sys.stderr)
+                stdout=sys.stderr)
-            self.attachlist = sorted(os.listdir(tmpdir))
+            self.attachlist = sorted(os.listdir(tmpdir.getpath()))
             return True
         except Exception as e:
             self.em.rclog("extractAttach: failed: %s" % e)
@@ -407,11 +398,12 @@ class PDFExtractor:
     def maybemaketmpdir(self):
         global tmpdir
         if tmpdir:
-            if not vacuumdir(tmpdir):
+            if not tmpdir.vacuumdir():
-                self.em.rclog("openfile: vacuumdir %s failed" % tmpdir)
+                self.em.rclog("openfile: vacuumdir %s failed" % tmpdir.getpath())
                 return False
         else:
-            tmpdir = tempfile.mkdtemp(prefix='rclmpdf')
+            tmpdir = rclexecm.SafeTmpDir("rclpdf", self.em)
+            #self.em.rclog("Using temporary directory %s" % tmpdir.getpath())
             if self.pdftk and re.match("/snap/", self.pdftk):
                 # We know this is Unix (Ubuntu actually). Check that tmpdir
                 # belongs to the user as snap commands can't use /tmp to share
@@ -423,9 +415,7 @@ class PDFExtractor:
                     if st.st_uid == os.getuid():
                         ok = True
                 if not ok:
-                    self.em.rclog(
+                    self.em.rclog("pdftk is a snap command and needs TMPDIR to be owned by you")
-                        "pdftk is a snap command and needs TMPDIR to be "
-                        "a directory you own")
 
     def _process_annotations(self, html):
         doc = Poppler.Document.new_from_file(
@@ -530,7 +520,7 @@ class PDFExtractor:
         if not self.attextractdone:
             if not self.extractAttach():
                 return (False, "", "", rclexecm.RclExecM.eofnow)
-        path = os.path.join(tmpdir, ipath)
+        path = os.path.join(tmpdir.getpath(), ipath)
         if os.path.isfile(path):
             f = open(path, "rb")
             docdata = f.read();