From da056917bfa4109972d9a38c09136cc760e5afd5 Mon Sep 17 00:00:00 2001
From: Brett Smith <brettcsmith@brettcsmith.org>
Date: Sat, 5 Sep 2020 14:59:13 -0400
Subject: [PATCH] meta_repo_links: Check all non-RT links. RT#12517

This is less "future-proof," but the thing is, it's premature to try
to anticipate what other link formats will be in the future. See
discussion in comments.
---
 .../plugin/meta_repo_links.py                  | 18 +++++++++++++++---
 setup.py                                       |  2 +-
 tests/test_meta_repo_links.py                  |  6 ++++++
 3 files changed, 22 insertions(+), 4 deletions(-)

diff --git a/conservancy_beancount/plugin/meta_repo_links.py b/conservancy_beancount/plugin/meta_repo_links.py
index fdf1fc2..b2cac41 100644
--- a/conservancy_beancount/plugin/meta_repo_links.py
+++ b/conservancy_beancount/plugin/meta_repo_links.py
@@ -20,6 +20,7 @@ from . import core
 from .. import config as configmod
 from .. import data
 from .. import errors as errormod
+from .. import rtutil
 from ..beancount_types import (
     MetaKey,
     MetaValue,
@@ -35,7 +36,6 @@ from typing import (
 class MetaRepoLinks(core.TransactionHook):
     HOOK_GROUPS = frozenset(['linkcheck'])
     LINK_METADATA = data.LINK_METADATA.difference('rt-id')
-    PATH_PUNCT_RE = re.compile(r'[:/]')
     SKIP_FLAGS = '!'
 
     def __init__(self, config: configmod.Config) -> None:
@@ -57,8 +57,20 @@ class MetaRepoLinks(core.TransactionHook):
                 yield errormod.InvalidMetadataError(txn, key, meta[key], post)
             else:
                 for link in links:
-                    match = self.PATH_PUNCT_RE.search(link)
-                    if match and match.group(0) == ':':
+                    # FUTURE DEVELOPMENT NOTE: As of this writing we have two
+                    # link checkers. Right now the division of reporting
+                    # responsibility is: MetaRTLinks reports problems with any
+                    # link that starts with `rt:`, while this checker reports
+                    # problems with anything else.
+                    # If we add more link checkers in the future, we might need
+                    # to give more thought about which checker is responsible
+                    # for flagging links in "unknown formats." But today, I
+                    # can't write better code that anticipates that need
+                    # without having any idea of what future link formats will
+                    # look like. Today, asking the RT parser "can you handle
+                    # this?" is the safest way to make sure all bad links get
+                    # reported.
+                    if rtutil.RT.parse(link) is not None:
                         pass
                     elif not (self.repo_path / link).exists():
                         yield errormod.BrokenLinkError(txn, key, link)
diff --git a/setup.py b/setup.py
index a590f6a..97fb540 100755
--- a/setup.py
+++ b/setup.py
@@ -5,7 +5,7 @@ from setuptools import setup
 setup(
     name='conservancy_beancount',
     description="Plugin, library, and reports for reading Conservancy's books",
-    version='1.9.6',
+    version='1.9.7',
     author='Software Freedom Conservancy',
     author_email='info@sfconservancy.org',
     license='GNU AGPLv3+',
diff --git a/tests/test_meta_repo_links.py b/tests/test_meta_repo_links.py
index 68c2c2b..fbaaf81 100644
--- a/tests/test_meta_repo_links.py
+++ b/tests/test_meta_repo_links.py
@@ -45,6 +45,12 @@ GOOD_LINKS = [Path(s) for s in [
 BAD_LINKS = [Path(s) for s in [
     'NonexistentDirectory/NonexistentFile1.txt',
     'NonexistentDirectory/NonexistentFile2.txt',
+    'egproto:',
+    'egproto:123',
+    'egproto:123/456',
+    'egproto:foo'
+    'egproto:/foo/bar',
+    ';egproto::',
 ]]
 
 NOT_FOUND_MSG = '{} not found in repository: {}'.format