From a2a4b506cb2222b634ada0fe1afca46526879e14 Mon Sep 17 00:00:00 2001
From: udf <tabhooked@gmail.com>
Date: Sun, 10 Jun 2018 20:40:15 +0200
Subject: [PATCH 1/2] Skip existing entities

It's a very hacky PoC, but it fixes things like urls being incorrectly parsed
---
 stdplugins/markdown.py | 31 +++++++++++++++++++++----------
 1 file changed, 21 insertions(+), 10 deletions(-)

diff --git a/stdplugins/markdown.py b/stdplugins/markdown.py
index ce08713..50d69f9 100644
--- a/stdplugins/markdown.py
+++ b/stdplugins/markdown.py
@@ -70,12 +70,17 @@ MATCHERS = [
 ]
 
 
-def parse(message):
+def parse(message, old_entities=[]):
     entities = []
+    old_entities = {e.offset: e for e in old_entities}
 
     i = 0
     message = _add_surrogate(message)
     while i < len(message):
+        # skip already existing entities if we're at one
+        if i in old_entities:
+            i += old_entities[i].length
+
         # find the first pattern that matches
         for pattern, parser in MATCHERS:
             match = pattern.match(message, pos=i)
@@ -86,6 +91,16 @@ def parse(message):
             continue
 
         text, entity = parser(match)
+
+        # shift old entities after our current position (so they stay in place)
+        shift = len(text) - len(message[match.start():match.end()])
+        if shift:
+            old_entities = old_entities.values()
+            for entity in old_entities:
+                if entity.offset >= i:
+                    entity.offset += shift
+            old_entities = {e.offset: e for e in old_entities}
+
         # replace whole match with text from parser
         message = ''.join((
             message[:match.start()],
@@ -100,20 +115,16 @@ def parse(message):
         # skip past the match
         i += len(text)
 
-    return _del_surrogate(message), entities
+    return _del_surrogate(message), entities + list(old_entities.values())
 
 
 @borg.on(events.MessageEdited(outgoing=True))
 @borg.on(events.NewMessage(outgoing=True))
 async def reparse(event):
-    message, msg_entities = await borg._parse_message_text(event.text, parse)
-    # filter out entities that we don't generate
-    old_entities = []
-    for entity in event.message.entities or []:
-        if isinstance(entity, PARSED_ENTITIES):
-            old_entities.append(entity)
-
-    if len(old_entities) == len(msg_entities) and event.raw_text == message:
+    old_entities = event.message.entities or []
+    parser = partial(parse, old_entities=old_entities)
+    message, msg_entities = await borg._parse_message_text(event.raw_text, parser)
+    if len(old_entities) >= len(msg_entities) and event.raw_text == message:
         return
 
     await borg(EditMessageRequest(

From a07a352d1ca37bdf0f52f4547ad2a67be23f4eef Mon Sep 17 00:00:00 2001
From: udf <tabhooked@gmail.com>
Date: Sun, 10 Jun 2018 21:10:48 +0200
Subject: [PATCH 2/2] use None as default parameter instead of empty list

---
 stdplugins/markdown.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/stdplugins/markdown.py b/stdplugins/markdown.py
index 50d69f9..976c936 100644
--- a/stdplugins/markdown.py
+++ b/stdplugins/markdown.py
@@ -70,9 +70,9 @@ MATCHERS = [
 ]
 
 
-def parse(message, old_entities=[]):
+def parse(message, old_entities=None):
     entities = []
-    old_entities = {e.offset: e for e in old_entities}
+    old_entities = {e.offset: e for e in old_entities or []}
 
     i = 0
     message = _add_surrogate(message)