Jules

scripts/test.py

@@ -74,6 +74,8 @@
         If true (the default on non-OpenBSD systems), we let pip create and use
         its own new venv to build PyMuPDF. Otherwise we force pip to use the
         current venv.
+    --build-mupdf 0|1
+        Whether to rebuild mupdf when we build PyMuPDF. Default is 1.
     --gdb 0|1
         Run tests under gdb.
     --timeout <seconds>
@@ -118,7 +120,9 @@ def main(argv):
 
     build_isolation = None
     valgrind = False
+    s = True
     build_type = None
+    build_mupdf = True
     gdb = False
     implementations = None
     test_names = list()
@@ -164,6 +168,8 @@ def main(argv):
             timeout = float(next(args))
         elif arg == '-v':
             venv_quick = True
+        elif arg == '--build-mupdf':
+            build_mupdf = int(next(args))
         elif arg == '--gdb':
             gdb = int(next(args))
         elif arg == '--valgrind':
@@ -198,6 +204,7 @@ def do_build():
                 build_type=build_type,
                 build_isolation=build_isolation,
                 venv_quick=venv_quick,
+                build_mupdf=build_mupdf,
                 )
     def do_test():
         test(
@@ -272,7 +279,13 @@ def venv_info(pytest_args=None):
     return ret
 
 
-def build(implementations=None, build_type=None, build_isolation=None, venv_quick=False):
+def build(
+        implementations=None,
+        build_type=None,
+        build_isolation=None,
+        venv_quick=False,
+        build_mupdf=True,
+        ):
     '''
     Args:
         build_type:
@@ -281,7 +294,8 @@ def build(implementations=None, build_type=None, build_isolation=None, venv_quic
             See top-level option `--build-isolation`.
         venv_quick:
             See top-level option `-v`.
-
+        build_mupdf:
+            See top-level option `build-mupdf`
     '''
     print(f'{build_type=}')
     print(f'{build_isolation=}')
@@ -320,6 +334,8 @@ def build(implementations=None, build_type=None, build_isolation=None, venv_quic
         if 'r' in implementations or 'R' in implementations:
             v += 'b'
         env_extra['PYMUPDF_SETUP_IMPLEMENTATIONS'] = v
+    if not build_mupdf:
+        env_extra['PYMUPDF_SETUP_MUPDF_REBUILD'] = '0'
     if build_type:
         env_extra['PYMUPDF_SETUP_MUPDF_BUILD_TYPE'] = build_type
     gh_release.run(f'pip install{build_isolation_text} -vv {pymupdf_dir}', env_extra=env_extra)

src/__init__.py

@@ -17435,11 +17435,7 @@ def JM_set_ocg_arrays_imp(arr, list_):
     Works with dict like {"basestate":name, "on":list, "off":list, "rbg":list}
     '''
     pdf = mupdf.pdf_get_bound_document(arr)
-    for i, item in enumerate(list_):
-        xref = 0
-        if JM_INT_ITEM(list_, i)[0] == 1:
-            # Not found.
-            continue
+    for xref in list_:
         obj = mupdf.pdf_new_indirect(pdf, xref, 0)
         mupdf.pdf_array_push(arr, obj)
 

tests/test_textextract.py

@@ -3,6 +3,7 @@
 No checks performed - just contribute to code coverage.
 """
 import os
+import sys
 
 import fitz
 
@@ -239,3 +240,30 @@ def test_3186():
         t = page.get_text()
         texts.append(t)
     assert texts == texts_expected, f'Unexpected output: {texts=}'
+
+
+def test_3197():
+    '''
+    MuPDF's ActualText support fixes handling of test_3197.pdf.
+    '''
+    if fitz.mupdf_version_tuple < (1, 24):
+        print(f'Not running on {fitz.mupdf_version_tuple=}.')
+        return
+    path = os.path.abspath(f'{__file__}/../../tests/resources/test_3197.pdf')
+
+    text_utf8_expected = [
+            b'NYSE - Nasdaq Real Time Price \xe2\x80\xa2 USD\nFord Motor Company (F)\n12.14 -0.11 (-0.90%)\nAt close: 4:00 PM EST\nAfter hours: 7:43 PM EST\nAll numbers in thousands\nAnnual\nQuarterly\nDownload\nSummary\nNews\nChart\nConversations\nStatistics\nHistorical Data\nProfile\nFinancials\nAnalysis\nOptions\nHolders\nSustainability\nInsights\nFollow\n12.15 +0.01 (+0.08%)\nIncome Statement\nBalance Sheet\nCash Flow\nSearch for news, symbols or companies\nNews\nFinance\nSports\nSign in\nMy Portfolio\nNews\nMarkets\nSectors\nScreeners\nPersonal Finance\nVideos\nFinance Plus\nBack to classic\nMore\n',
+            b'Related Tickers\nTTM\n12/31/2023\n12/31/2022\n12/31/2021\n12/31/2020\n14,918,000\n14,918,000\n6,853,000\n15,787,000\n24,269,000\n-17,628,000\n-17,628,000\n-4,347,000\n2,745,000\n-18,615,000\n2,584,000\n2,584,000\n2,511,000\n-23,498,000\n2,315,000\n25,110,000\n25,110,000\n25,340,000\n20,737,000\n25,935,000\n-8,236,000\n-8,236,000\n-6,866,000\n-6,227,000\n-5,742,000\n51,659,000\n51,659,000\n45,470,000\n27,901,000\n65,900,000\n-41,965,000\n-41,965,000\n-45,655,000\n-54,164,000\n-60,514,000\n-335,000\n-335,000\n-484,000\n--\n--\n6,682,000\n6,682,000\n-13,000\n9,560,000\n18,527,000\n \nYahoo Finance Plus Essential\naccess required.\nUnlock Access\nBreakdown\nOperating Cash\nFlow\nInvesting Cash\nFlow\nFinancing Cash\nFlow\nEnd Cash Position\nCapital Expenditure\nIssuance of Debt\nRepayment of Debt\nRepurchase of\nCapital Stock\nFree Cash Flow\n12/31/2020 - 6/1/1972\nGM\nGeneral Motors Compa\xe2\x80\xa6\n39.49 +1.23%\n\xc2\xa0\nRIVN\nRivian Automotive, Inc.\n15.39 -3.15%\n\xc2\xa0\nNIO\nNIO Inc.\n5.97 +0.17%\n\xc2\xa0\nSTLA\nStellantis N.V.\n25.63 +0.91%\n\xc2\xa0\nLCID\nLucid Group, Inc.\n3.7000 +0.54%\n\xc2\xa0\nTSLA\nTesla, Inc.\n194.77 +0.52%\n\xc2\xa0\nTM\nToyota Motor Corporati\xe2\x80\xa6\n227.09 +0.14%\n\xc2\xa0\nXPEV\nXPeng Inc.\n9.08 +0.89%\n\xc2\xa0\nFSR\nFisker Inc.\n0.5579 -11.46%\n\xc2\xa0\nCopyright \xc2\xa9 2024 Yahoo.\nAll rights reserved.\nPOPULAR QUOTES\nTesla\nDAX Index\nKOSPI\nDow Jones\nS&P BSE SENSEX\nSPDR S&P 500 ETF Trust\nEXPLORE MORE\nCredit Score Management\nHousing Market\nActive vs. Passive Investing\nShort Selling\nToday\xe2\x80\x99s Mortgage Rates\nHow Much Mortgage Can You Afford\nABOUT\nData Disclaimer\nHelp\nSuggestions\nSitemap\n',
+            ]
+
+    with fitz.open(path) as document:
+        for i, page in enumerate(document):
+            text = page.get_text()
+            #print(f'{i=}:')
+            text_utf8 = text.encode('utf8')
+            #print(f'                {text_utf8=}')
+            #print(f'    {text_utf8_expected[i]=}')
+            if fitz.mupdf_version_tuple >= (1, 24):
+                assert text_utf8 == text_utf8_expected[i]
+            else:
+                assert text_utf8 != text_utf8_expected[i]