fix a crash in opnk when cached but not downloaded

As initially identified by Paul Wise in [Debian Bug#1064209], opnk.py
experiences the following warning when running under python3.12:

	$ python3.12 opnk.py gemini://ploum.net >/dev/null
	/home/emollier/debian/forward-upstream/offpunk/opnk.py:52: SyntaxWarning: invalid escape sequence '\%'
	  less_prompt = "page %%d/%%D- lines %%lb/%%L - %%Pb\%%"

This is due to the interpretation of escape sequences being less
relaxed in the new Python interpreter version.  Doubling the backslash
is one way to resolve this issue.

[Debian Bug#1064209]: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1064209

Signed-off-by: Étienne Mollier <emollier@debian.org>

CHANGELOG

@@ -1,5 +1,12 @@
 # Offpunk History
 
+## 2.3 - Unreleased
+- offpunk/netcache: fix IPv6 as an URL (bug #40)
+- ansicat: display empty files (instead of opening them with xdg-open)
+- fix escape sequence warning in python 3.12 (by Étienne Mollier) (Debian #1064209)
+- ansicat : fix crash when feedparser is crashing on bad RSS
+- netcache: fix spartan protocol error
+- opnk: fix a crash when caching returns None
 
 ## 2.2 - February 13th 2023
 - cache folder is now configurable through $OFFPUNK_CACHE_PATH environment variable (by prx)

ansicat.py

@@ -727,6 +727,13 @@ class GemtextRenderer(AbstractRenderer):
         links += hidden_links
         return r.get_final(), links
 
+class EmptyRenderer(GemtextRenderer):
+    def get_mime(self):
+        return "text/empty"
+    def prepare(self,body,mode=None):
+        text= "(empty file)"
+        return [[text, "GemtextRenderer"]]
+
 class GopherRenderer(AbstractRenderer):
     def get_mime(self):
         return "text/gopher"
@@ -870,10 +877,15 @@ class FeedRenderer(GemtextRenderer):
         return "application/rss+xml"
     def is_valid(self):
         if _DO_FEED:
-            parsed = feedparser.parse(self.body)
+            try:
+                parsed = feedparser.parse(self.body)
+            except:
+                parsed = False
         else:
             return False
-        if parsed.bozo:
+        if not parsed:
+            return False
+        elif parsed.bozo:
             return False
         else:
             #If no content, then fallback to HTML
@@ -1312,11 +1324,16 @@ _FORMAT_RENDERERS = {
     "text/gopher": GopherRenderer,
     "image/*": ImageRenderer,
     "application/javascript": HtmlRenderer,
+    "application/json": HtmlRenderer,
+    "text/empty": EmptyRenderer,
 }
 def get_mime(path,url=None):
     #Beware, this one is really a shaddy ad-hoc function
     if not path:
         return None
+    #If the file is empty, simply returns it
+    elif os.path.exists(path) and os.stat(path).st_size == 0:
+        return "text/empty"
     elif url and url.startswith("gopher://"):
         #special case for gopher
         #code copy/pasted from netcache

netcache.py

@@ -656,6 +656,9 @@ def _fetch_gemini(url,timeout=DEFAULT_TIMEOUT,interactive=True,accept_bad_ssl_ce
     # Send request and wrap response in a file descriptor
     url = urllib.parse.urlparse(url)
     new_netloc = host
+    #Handle IPV6 hostname
+    if ":" in new_netloc:
+        new_netloc = "[" + new_netloc + "]"
     if port != standard_ports["gemini"]:
         new_netloc += ":" + str(port)
     url = urllib.parse.urlunparse(url._replace(netloc=new_netloc))
@@ -799,8 +802,10 @@ def fetch(url,offline=False,download_image_first=True,images_mode="readable",val
                 path=_fetch_finger(url,**kwargs)
             elif scheme == "gemini":
                 path,newurl=_fetch_gemini(url,**kwargs)
+            elif scheme == "spartan":
+                path,newurl=_fetch_spartan(url,**kwargs)
             else:
-                print("scheme %s not implemented yet")
+                print("scheme %s not implemented yet"%scheme)
         except UserAbortException:
             return None, newurl
         except Exception as err:

offutils.py

@@ -99,7 +99,7 @@ def fix_ipv6_url(url):
         netloc, rest = schemaless.split("/",1)
         if netloc.count(":") > 2 and "[" not in netloc and "]" not in netloc:
             schemaless = "[" + netloc + "]" + "/" + rest
-    elif schemaless.count(":") > 2:
+    elif schemaless.count(":") > 2 and "[" not in schemaless and "]" not in schemaless:
         schemaless = "[" + schemaless + "]/"
     if schema:
         return schema + "://" + schemaless
@@ -121,7 +121,16 @@ def looks_like_url(word):
         if mailto:
             return "@" in word
         elif not local:
-            return start and ("." in word or "localhost" in word)
+            if start:
+                #IPv4
+                if "." in word or "localhost" in word:
+                    return True
+                #IPv6
+                elif "[" in word and ":" in word and "]" in word:
+                    return True
+                else: return False
+            else:   return False
+            return start and ("." in word or "localhost" in word or ":" in word)
         else:
             return "/" in word
     except ValueError:

opnk.py

@@ -49,7 +49,7 @@ else:
 # are there on purpose (surch in asciiart)
 #--incsearch : incremental search starting rev581
 def less_cmd(file, histfile=None,cat=False,grep=None):
-    less_prompt = "page %%d/%%D- lines %%lb/%%L - %%Pb\%%"
+    less_prompt = "page %%d/%%D- lines %%lb/%%L - %%Pb\\%%"
     if less_version >= 581:
         less_base = "less --incsearch --save-marks -~ -XRfWiS -P \"%s\""%less_prompt
     elif less_version >= 572:
@@ -159,7 +159,10 @@ class opencache():
                 if inpath in self.renderer_time.keys():
                     last_downloaded = netcache.cache_last_modified(inpath)
                     last_cached = self.renderer_time[inpath]
-                    usecache = last_cached > last_downloaded
+                    if last_cached and last_downloaded:
+                        usecache = last_cached > last_downloaded
+                    else:
+                        usecache = False
                 else:
                     usecache = False
             if not usecache: