slack_regex: Fix overlapping capture group in Slack regex.

The Slack text regexes match specific characters before and after the
formatted string to prevent cases like reformatting already formatted
text and test like `invlaid_bold_test` in
`slack_message_conversion.json`.

However, because the capture groups consume each matching character, two
formatted strings separated by a single matching character result in one
string not being matched, as the character is used to match the other
string.

e.g., (`*abc*b*def*` → **abc**b*def*) in this case, "b" is consumed to
match *abc*.

This prep commit modifies the Slack text regex in
`slack_message_conversion.py` to use a non-greedy quantifier for the
last capture group.

zerver/data_import/slack_message_conversion.py

@@ -52,21 +52,21 @@ SLACK_STRIKETHROUGH_REGEX = r"""
                              (\~)                                  # followed by an ~
                                 ([^~]*)            # any character except ~
                              (\~)                                  # followed by an ~
-                             (\n|$|[ -']|[+-/]|[:-?]|\*|\_|\}|\)|\]|\||\^)  # ends with specified characters
+                             (?=\n|$|[ -']|[+-/]|[:-?]|\*|\_|\}|\)|\]|\||\^)  # ends with specified characters
                              """
 SLACK_ITALIC_REGEX = r"""
                       (\n|^|[ -*]|[+-/]|[:-?]|\{|\[|\||\^|~)
                       (\_)
                           ([^_]*)                # any character except _
                       (\_)
-                      (\n|$|[ -']|[+-/]|[:-?]|\}|\)|\]|\*|\||\^|~)
+                      (?=\n|$|[ -']|[+-/]|[:-?]|\}|\)|\]|\*|\||\^|~)
                       """
 SLACK_BOLD_REGEX = r"""
                     (\n|^|[ -(]|[+-/]|[:-?]|\{|\[|\_|\||\^|~)
                     (\*)
                         ([^*]*)                 # any character except *
                     (\*)
-                    (\n|$|[ -']|[+-/]|[:-?]|\}|\)|\]|\_|\||\^|~)
+                    (?=\n|$|[ -']|[+-/]|[:-?]|\}|\)|\]|\_|\||\^|~)
                     """
 
 
@@ -139,9 +139,7 @@ def convert_markdown_syntax(text: str, regex: str, zulip_keyword: str) -> str:
     3. For italic formatting: This maps Slack's '_italic_' to Zulip's '*italic*'
     """
     for match in re.finditer(regex, text, re.VERBOSE):
-        converted_token = (
+        converted_token = match.group(1) + zulip_keyword + match.group(3) + zulip_keyword
-            match.group(1) + zulip_keyword + match.group(3) + zulip_keyword + match.group(5)
-        )
         text = text.replace(match.group(0), converted_token)
     return text
 

zerver/tests/fixtures/slack_message_conversion.json

@@ -94,6 +94,11 @@
       "name": "format_emoji_test",
       "input": "*1️⃣ bold* some _1️⃣ italic_ word ~1️⃣ strike~",
       "conversion_output": "**1️⃣ bold** some *1️⃣ italic* word ~~1️⃣ strike~~"
+    },
+    {
+      "name": "overlapping_capture_group_test",
+      "input": "*abc*\n*def*\n*ghi*\n*jkl*\n\n*ABC ABC*",
+      "conversion_output": "**abc**\n**def**\n**ghi**\n**jkl**\n\n**ABC ABC**"
     }
   ]
 }