Used qpdf in all tests to compare resulting PDFs + improved development doc

docs/Development.md

@@ -64,22 +64,30 @@ in `.git/hooks/pre-commit` in order to always invoke `black` & `pylint`
 before every commit:
 
 ```shell
-#!/bin/sh
-modified_py_files=$(git diff --cached --name-only --diff-filter=ACM | grep '.py$')
+#!/bin/bash
+git_cached_names() { git diff --cached --name-only --diff-filter=ACM; }
+modified_py_files=$(git_cached_names | grep '.py$')
+modified_fpdf_files=$(git_cached_names | grep ^fpdf | grep '.py$')
+# if python files modified, format
 if [ -n "$modified_py_files" ]; then
     if ! black --check $modified_py_files; then
         black $modified_py_files
         exit 1
     fi
-    pylint $modified_py_files
+    # if core files modified, lint
+    [[ $modified_fpdf_files == "" ]] || pylint $modified_fpdf_files
 fi
+unset git_cached_names
 ```
 
 It will abort the commit if `pylint` found issues
 or `black` detect non-properly formatted code.
 In the later case though, it will auto-format your code
 and you will just have to run `git commit -a` again.
 
+Be sure to check on what actions will actually be run in the CI workflow in
+`.github/workflows/continuous-integration-workflow.yml`.
+
 ## Testing ##
 
 Testing is done with [Tox](https://tox.readthedocs.io/en/latest/), and is

test/cell_multi_cell_refactor/cell.py

@@ -57,5 +57,6 @@ def test_ln_positioning_and_page_breaking_for_cell(self):
             self, doc, "test_ln_positioning_and_page_breaking_for_cell.pdf"
         )
 
+
 if __name__ == "__main__":
     unittest.main()

test/cell_multi_cell_refactor/multi_cell.py

@@ -68,5 +68,6 @@ def test_ln_positioning_and_page_breaking_for_multicell(self):
             self, doc, "test_ln_positioning_and_page_breaking_for_multicell.pdf"
         )
 
+
 if __name__ == "__main__":
     unittest.main()

test/utilities.py

@@ -1,12 +1,13 @@
 import hashlib
 import inspect
 import os
+import pathlib
 import shutil
 import sys
+import tempfile
 import warnings
 from datetime import datetime
 from subprocess import check_call, check_output
-from tempfile import NamedTemporaryFile
 
 from fpdf.template import Template
 
@@ -18,6 +19,62 @@
     )
 
 
+class TempFile:
+    """docstring for TempFile"""
+
+    def __init__(self, name, handle):
+        super(TempFile, self).__init__()
+        self.name = name
+        self.handle = handle
+
+    def read_whole_file(self):
+        while True:
+            if not self.handle.read(10):
+                break
+        return self.handle.read()
+
+    def rwf(self):
+        return self.read_whole_file()
+
+    def close(self):
+        self.handle.close()
+
+
+class TempFileCtx:
+    """docstring for TempFileCtxs"""
+
+    def __init__(self, prefix, delete, suffix):
+        super(TempFileCtx, self).__init__()
+        self.prefix = prefix
+        self.delete = delete
+        self.suffix = suffix
+
+    def random(self):
+        return os.urandom(24).hex()
+
+    def get_os_tmp(self):
+        return tempfile.gettempdir()
+
+    def get_tmp_file_name_base(self):
+        return self.prefix + self.random() + self.suffix
+
+    def get_tmp_file_name(self):
+        name = self.get_tmp_file_name_base()
+        return os.path.join(self.get_os_tmp(), name)
+
+    def __enter__(self):
+        self.abs_name = self.get_tmp_file_name()
+        pathlib.Path(self.abs_name).touch()
+        tmp_file = TempFile(self.abs_name, open(self.abs_name, "rb"))
+        self.tmp_file = tmp_file
+        return tmp_file
+
+    def __exit__(self, type, value, traceback):
+        self.tmp_file.close()
+        if self.delete:
+            os.remove(self.abs_name)
+
+
 def assert_pdf_equal(test, pdf_or_tmpl, rel_expected_pdf_filepath, delete=True):
     """
     Args:
@@ -33,16 +90,16 @@ def assert_pdf_equal(test, pdf_or_tmpl, rel_expected_pdf_filepath, delete=True):
         pdf = pdf_or_tmpl
     set_doc_date_0(pdf)  # Ensure PDFs CreationDate is always the same
     expected_pdf_filepath = relative_path_to(rel_expected_pdf_filepath, depth=2)
-    with NamedTemporaryFile(
+    with TempFileCtx(
         prefix="pyfpdf-test-", delete=delete, suffix="-actual.pdf"
     ) as actual_pdf_file:
         pdf.output(actual_pdf_file.name, "F")
         if not delete:
             print("Temporary file will not be deleted:", actual_pdf_file.name)
         if QPDF_AVAILABLE:  # Favor qpdf-based comparison, as it helps a lot debugging:
-            with NamedTemporaryFile(
+            with TempFileCtx(
                 prefix="pyfpdf-test-", delete=delete, suffix="-actual-qpdf.pdf"
-            ) as actual_qpdf_file, NamedTemporaryFile(
+            ) as actual_qpdf_file, TempFileCtx(
                 prefix="pyfpdf-test-", delete=delete, suffix="-expected-qpdf.pdf"
             ) as expected_qpdf_file:
                 _qpdf(actual_pdf_file.name, actual_qpdf_file.name)
@@ -54,7 +111,7 @@ def assert_pdf_equal(test, pdf_or_tmpl, rel_expected_pdf_filepath, delete=True):
                         expected_qpdf_file.name,
                     )
                 test.assertSequenceEqual(
-                    actual_qpdf_file.readlines(), expected_qpdf_file.readlines()
+                    actual_qpdf_file.rwf(), expected_qpdf_file.rwf()
                 )
         else:  # Fallback to hash comparison
             actual_hash = calculate_hash_of_file(actual_pdf_file.name)