PyPtt 2.3.17__tar.gz → 2.3.18__tar.gz

{pyptt-2.3.17 → pyptt-2.3.18}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: PyPtt
-Version: 2.3.17
+Version: 2.3.18
 Summary: PyPtt github: https://github.com/PyPtt/PyPtt
 Author-email: CodingMan <pttcodingman@gmail.com>
 License: GNU Lesser General Public License v3 (LGPLv3)

{pyptt-2.3.17 → pyptt-2.3.18}/PyPtt/__init__.py

@@ -1,4 +1,4 @@
-__version__ = '2.3.17'
+__version__ = '2.3.18'
 __author__ = 'CodingMan'
 __email__ = 'pttcodingman@gmail.com'
 

{pyptt-2.3.17 → pyptt-2.3.18}/PyPtt/_api_get_post_list.py

@@ -11,6 +11,24 @@ from . import log
 from . import screens
 
 
+def _cell_width(char: str) -> int:
+    # 終端機上全形字佔 2 格 (例如推文數的「爆」)，用 big5uao 的位元組數當作顯示寬度
+    return len(char.encode('big5uao', 'replace'))
+
+
+def _split_columns(line: str, bounds: List[int]) -> List[str]:
+    """依顯示欄位 (不是字元位置) 切開列表的一行。bounds 是各欄的起始欄位。"""
+    fields = [''] * len(bounds)
+    column = 0
+    for char in line:
+        for i in range(len(bounds) - 1, -1, -1):
+            if column >= bounds[i]:
+                fields[i] += char
+                break
+        column += _cell_width(char)
+    return fields
+
+
 def get_post_list(api, board: str, limit: int = 20, offset: int = 0) -> List[Dict]:
     _api_util.one_thread(api)
 
@@ -84,18 +102,21 @@ def get_post_list(api, board: str, limit: int = 20, offset: int = 0) -> List[Dic
                     # offsets here were tuned to that padded layout.
                     line = f' {line}'
 
+                # 欄位是固定的顯示寬度；推文數「爆」是全形字，用字元位置切會讓後面每一欄都偏一格
+                index_field, status, comment, list_date, _, author, title = _split_columns(
+                    line, [0, 8, 10, 12, 17, 18, 31])
+
                 try:
-                    cur_index = int(line[:8].strip())
+                    cur_index = int(index_field.strip())
                 except ValueError:
                     # Pinned/announcement posts use '★' instead of a numeric
                     # index — they have no sequential index and cannot be
                     # fetched by index, so skip them.
                     continue
-                status = line[8:10].strip()
-                comment = line[10:12].strip()
-                list_date = line[12:17].strip()
-                author = line[18:31].strip()
-                title = line[31:]
+                status = status.strip()
+                comment = comment.strip()
+                list_date = list_date.strip()
+                author = author.strip()
 
                 cur_post = {
                     data_type.PostField.board: board,

{pyptt-2.3.17 → pyptt-2.3.18}/PyPtt/screens.py

@@ -356,7 +356,8 @@ xy_pattern_h = re.compile(r'^=ESC=\[[\d]+;[\d]+H')
 xy_pattern_s = re.compile(r'^=ESC=\[[\d]+;[\d]+s')
 # 依 ECMA-48 剝掉引擎不處理的 CSI (顏色、ESC[?2026h、ESC[6n…, PttCurrent M.1789835163.A.633),
 # 以及被下一個 ESC 打斷的孤立 ESC (主選單動畫尾端的 '\x1b\x1b[?2026h'); 留著會讓引擎凍住畫面
-_color_sgr = re.compile(r'\x1B(?=\x1B)|\x1B\[(?!(?:\d+;\d+)?[Hs]|K|2J)[0-?]*[ -/]*[@-~]')
+# OSC (ESC ] ... BEL 或 ESC \\) 也一併剝掉, 例如內文網址的 OSC 8 超連結 '\x1b]8;;URL\x1b\\'
+_color_sgr = re.compile(r'\x1B\][^\x07\x1B]*(?:\x07|\x1B\\)|\x1B(?=\x1B)|\x1B\[(?!(?:\d+;\d+)?[Hs]|K|2J)[0-?]*[ -/]*[@-~]')
 
 
 def _preprocess(data: str) -> str:
@@ -541,6 +542,8 @@ def _escape_complete(seq: str) -> bool:
     engine will bail on it, same as the batch parser)."""
     if len(seq) < 2:
         return False            # lone ESC — wait for more
+    if seq[1] == ']':
+        return '\x07' in seq or '\x1b\\' in seq   # OSC 要等到 BEL 或 ESC \\ 結尾
     if seq[1] != '[':
         return True             # ESC + non-'[': engine bails on it
     for ch in seq[2:]:
@@ -581,6 +584,9 @@ class IncrementalScreen(_VT100Engine):
         while cut > 0 and s[cut - 1] in ' \x08':
             cut -= 1
         esc = s.rfind('\x1b')
+        osc = s.rfind('\x1b]')
+        if osc != -1 and not _escape_complete(s[osc:]):
+            esc = osc           # OSC 的 ESC \\ 結尾可能還沒到, 從 OSC 開頭整段留著
         if esc != -1 and not _escape_complete(s[esc:]):
             while esc > 0 and s[esc - 1] == '\x1b':
                 esc -= 1        # 孤立 ESC 要跟後面的 ESC 一起送, _preprocess 才認得

{pyptt-2.3.17 → pyptt-2.3.18}/PyPtt.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: PyPtt
-Version: 2.3.17
+Version: 2.3.18
 Summary: PyPtt github: https://github.com/PyPtt/PyPtt
 Author-email: CodingMan <pttcodingman@gmail.com>
 License: GNU Lesser General Public License v3 (LGPLv3)

{pyptt-2.3.17 → pyptt-2.3.18}/tests/test_get_post_list_parser.py

@@ -118,6 +118,16 @@ class TestLineLayouts:
         assert posts[0][data_type.PostField.push_number] == ''
         assert posts[0][data_type.PostField.list_date] == '4/17'
 
+    def test_full_width_push_count(self):
+        # 「爆」是全形字，終端機上佔 2 格；用字元位置切的話後面每一欄都會左移一格
+        cursor = '>  9446 +爆10/15 someone      □ [問安] 1015'
+        api = _build_api(_build_screen(cursor, []), newest=9446)
+        posts = _api_get_post_list.get_post_list(api, 'Python', limit=1)
+        assert posts[0][data_type.PostField.push_number] == '爆'
+        assert posts[0][data_type.PostField.list_date] == '10/15'
+        assert posts[0][data_type.PostField.author] == "someone"
+        assert posts[0][data_type.PostField.title].strip() == '□ [問安] 1015'
+
     def test_reply_marker_R_in_title(self):
         # Title section includes the kind marker ('R:', '□', '轉', '鎖').
         cursor = '>  9440   4 1/15 ddavid       R: [問題] @property 真正的運用是啥'

{pyptt-2.3.17 → pyptt-2.3.18}/tests/test_incremental_parity.py

@@ -54,6 +54,8 @@ CRAFTED = [
     b'  \x08\x08\xe2\x97\x8e tail',                     # ◎ via reservation pair
     b'\x1b[10;1H  \x08\x08\xe2\x80\xbb\x1b[10;3H \xe7\x99\xbc\xe4\xbf\xa1\xe7\xab\x99: x',
     b'\x1b[6nHELLO',                                    # unknown escape -> bail
+    b'a\x1b]8;;https://x.y/z\x1b\\link\x1b]8;;\x1b\\b',     # OSC 8 hyperlink (ST)
+    b'a\x1b]8;;https://x.y/z\x07link\x1b]8;;\x07b',         # OSC 8 hyperlink (BEL)
     b'line1\r\nline2',                                  # CRLF
     b'helloworld\x1b[1;2HXY',                           # replace mode
     b'\x1b[100;1HX',                                    # row clamp

{pyptt-2.3.17 → pyptt-2.3.18}/tests/test_vt100_parser.py

@@ -481,3 +481,26 @@ def test_stray_esc_split_across_chunks():
         s.feed(data[:cut])
         s.feed(data[cut:])
         assert s.screen == VT100Parser(data, 'utf-8').screen, cut
+
+
+# 內文網址被包成 OSC 8 超連結 (ESC ] 8 ; ; URL ST 文字 ESC ] 8 ; ; ST); 沒被吃掉時引擎凍住, get_content 讀不到頁尾
+OSC8_POST = (b'\x1b[2J\x1b[3;1H\xe7\xb6\xb2\xe5\x9d\x80 \x1b]8;;https://example.com/a?b=1\x1b\\https://example.com/a?b=1'
+             b'\x1b]8;;\x1b\\ end\x1b[4;1H\x1b]8;;https://ptt.cc\x07ptt\x1b]8;;\x07\x1b[24;1Hfooter')
+
+
+def test_osc8_hyperlink_is_stripped():
+    lines = parse(OSC8_POST)
+    assert lines[2] == '網址 https://example.com/a?b=1 end'
+    assert lines[3] == 'ptt'
+    assert lines[23] == 'footer'
+
+
+def test_osc8_hyperlink_split_across_chunks():
+    from PyPtt.screens import IncrementalScreen
+    expected = VT100Parser(OSC8_POST, 'utf-8').screen
+    for cut in range(1, len(OSC8_POST)):
+        s = IncrementalScreen('utf-8')
+        s.feed(OSC8_POST[:cut])
+        s.feed(OSC8_POST[cut:])
+        assert not s._bailed, cut
+        assert s.screen == expected, cut