From 01bfe4de97d2d09b216aec19d0c819d70a13fb3c Mon Sep 17 00:00:00 2001
From: Roland <roland@learningselfreliance.com>
Date: Wed, 2 Feb 2022 20:48:43 -0700
Subject: [PATCH] Limiting maximum title length in Archive files.

---
 modules/archive/lib.py           | 3 +++
 modules/archive/test/test_lib.py | 7 ++++---
 2 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/modules/archive/lib.py b/modules/archive/lib.py
index ee17d29d..85a4a3d3 100644
--- a/modules/archive/lib.py
+++ b/modules/archive/lib.py
@@ -78,6 +78,7 @@ def get_new_archive_files(url: str, title: Optional[str]) -> ArchiveFiles:
     dt = archive_strftime(now())
 
     title = escape_file_name(title or 'NA')
+    title = title[:50]
     prefix = f'{dt}_{title}'
     singlefile_path = directory / f'{prefix}.html'
     readability_path = directory / f'{prefix}.readability.html'
@@ -407,6 +408,8 @@ def migrate_archive_files():
             with archive_files.readability_json.open() as fh:
                 title = escape_file_name(json.load(fh).get('title') or 'NA')
 
+            title = title[:50]
+
             dt = archive_strftime(dt)
             prefix = f'{dt}_{title}'
             singlefile_path = domain_directory / f'{prefix}.html'
diff --git a/modules/archive/test/test_lib.py b/modules/archive/test/test_lib.py
index b226abf0..a6402d81 100644
--- a/modules/archive/test/test_lib.py
+++ b/modules/archive/test/test_lib.py
@@ -490,7 +490,7 @@ def test_migrate_archive_files(test_session, archive_directory, archive_factory)
     # Archive6
     (example_org_directory / '2000-01-01 00:00:00.000000.html').touch()
     path = example_org_directory / '2000-01-01 00:00:00.000000.readability.json'
-    path.write_text(json.dumps({'title': 'invalid#*: title'}))
+    path.write_text(json.dumps({'title': 'A' * 500}))  # Really long file name.
 
     # Migration forms a plan, then performs it.
     plan = lib.migrate_archive_files()
@@ -510,9 +510,10 @@ def test_migrate_archive_files(test_session, archive_directory, archive_factory)
         (pathlib.Path('example.com/2021-10-05 16:20:10.346823.readability.json'),
          pathlib.Path('example.com/2021-10-05-16-20-10_my title.readability.json')),
         (pathlib.Path('example.org/2000-01-01 00:00:00.000000.html'),
-         pathlib.Path('example.org/2000-01-01-00-00-00_invalid# title.html')),
+         pathlib.Path('example.org/2000-01-01-00-00-00_AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA.html')),
         (pathlib.Path('example.org/2000-01-01 00:00:00.000000.readability.json'),
-         pathlib.Path('example.org/2000-01-01-00-00-00_invalid# title.readability.json')),
+         pathlib.Path(
+             'example.org/2000-01-01-00-00-00_AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA.readability.json')),
     ]
     # New Archive files still exist.
     assert all(i.is_file() for i in archive1.my_paths())