{"thread":{"id":"57091","subject":"[PATCH v2 0/5] git-p4: fix RCS keyword processing encoding errors","startedAt":"2021-12-16T13:47:22Z","lastAt":"2021-12-16T13:47:33Z","messageCount":6,"participants":["Joel Holdsworth"],"isPatch":true,"patchVersion":2,"patchTotal":5},"messages":[{"id":"444219","messageId":"20211216134619.2048348-1-jholdsworth@nvidia.com","threadId":"57091","inReplyTo":null,"subject":"[PATCH v2 0/5] git-p4: fix RCS keyword processing encoding errors","fromName":"Joel Holdsworth","fromEmail":"jholdsworth@nvidia.com","sentAt":"2021-12-16T13:46:14Z","receivedAt":"2021-12-16T13:47:22Z","isPatch":true,"sender":{"key":"jholdsworth@nvidia.com","avatar":"https://avatars.githubusercontent.com/u/1449493?v=4"},"body":"This patch-set fixes a family of issues with git-p4's handling of\nincoming text data that contains RCS keywords, when those files contain\nbytes which are invalid UTF-8 codes.\n\nAmong the patches is a fix for the issue, as well as some peripheral\ntidy-ups and improvements to the existing code.\n\nThis patch-set is compatible and has been tested with both Python 2 and\n3, and includes a test.\n\nThis v2 version of the patch-set seperates the read/write mode fix (4/5)\nout from the RCS byte processing patch (5/5)\n\nJoel Holdsworth (5):\n git-p4: use with statements to close files after use in\n patchRCSKeywords\n git-p4: pre-compile RCS keyword regexes\n git-p4: add raw option to read_pipelines\n git-p4: open temporary patch file for write only\n git-p4: resolve RCS keywords in bytes not utf-8\n\n git-p4.py | 66 ++++++++++++++++++-------------------------\n t/t9810-git-p4-rcs.sh | 15 ++++++++++\n 2 files changed, 42 insertions(+), 39 deletions(-)\n\n-- \n2.34.1\n\n"},{"id":"444220","messageId":"20211216134619.2048348-2-jholdsworth@nvidia.com","threadId":"57091","inReplyTo":"20211216134619.2048348-1-jholdsworth@nvidia.com","subject":"[PATCH v2 1/5] git-p4: use with statements to close files after use in patchRCSKeywords","fromName":"Joel Holdsworth","fromEmail":"jholdsworth@nvidia.com","sentAt":"2021-12-16T13:46:15Z","receivedAt":"2021-12-16T13:47:23Z","isPatch":true,"sender":{"key":"jholdsworth@nvidia.com","avatar":"https://avatars.githubusercontent.com/u/1449493?v=4"},"body":"Python with statements are used to wrap the execution of a block of code\nso that an object can be safely released when execution leaves the\nscope.\n\nThey are desirable for improving code tidyness, and to ensure that\nobjects are properly destroyed even when exceptions are thrown.\n\nSigned-off-by: Joel Holdsworth \n---\n git-p4.py | 13 +++++--------\n 1 file changed, 5 insertions(+), 8 deletions(-)\n\ndiff --git a/git-p4.py b/git-p4.py\nindex 2b4500226a..226cdef424 100755\n--- a/git-p4.py\n+++ b/git-p4.py\n@@ -1757,14 +1757,11 @@ def patchRCSKeywords(self, file, pattern):\n # Attempt to zap the RCS keywords in a p4 controlled file matching the given pattern\n (handle, outFileName) = tempfile.mkstemp(dir='.')\n try:\n- outFile = os.fdopen(handle, \"w+\")\n- inFile = open(file, \"r\")\n- regexp = re.compile(pattern, re.VERBOSE)\n- for line in inFile.readlines():\n- line = regexp.sub(r'$\\1$', line)\n- outFile.write(line)\n- inFile.close()\n- outFile.close()\n+ with os.fdopen(handle, \"w+\") as outFile, open(file, \"r\") as inFile:\n+ regexp = re.compile(pattern, re.VERBOSE)\n+ for line in inFile.readlines():\n+ line = regexp.sub(r'$\\1$', line)\n+ outFile.write(line)\n # Forcibly overwrite the original file\n os.unlink(file)\n shutil.move(outFileName, file)\n-- \n2.34.1\n\n"},{"id":"444221","messageId":"20211216134619.2048348-3-jholdsworth@nvidia.com","threadId":"57091","inReplyTo":"20211216134619.2048348-1-jholdsworth@nvidia.com","subject":"[PATCH v2 2/5] git-p4: pre-compile RCS keyword regexes","fromName":"Joel Holdsworth","fromEmail":"jholdsworth@nvidia.com","sentAt":"2021-12-16T13:46:16Z","receivedAt":"2021-12-16T13:47:26Z","isPatch":true,"sender":{"key":"jholdsworth@nvidia.com","avatar":"https://avatars.githubusercontent.com/u/1449493?v=4"},"body":"Previously git-p4.py would compile one of two regular expressions for\never RCS keyword-enabled file. This patch improves simplifies the code\nby pre-compiling the two regular expressions when the script first\nloads.\n\nSigned-off-by: Joel Holdsworth \n---\n git-p4.py | 48 ++++++++++++++++++------------------------------\n 1 file changed, 18 insertions(+), 30 deletions(-)\n\ndiff --git a/git-p4.py b/git-p4.py\nindex 226cdef424..0af83b9c72 100755\n--- a/git-p4.py\n+++ b/git-p4.py\n@@ -56,6 +56,9 @@\n \n p4_access_checked = False\n \n+re_ko_keywords = re.compile(r'\\$(Id|Header)(:[^$\\n]+)?\\$')\n+re_k_keywords = re.compile(r'\\$(Id|Header|Author|Date|DateTime|Change|File|Revision)(:[^$\\n]+)?\\$')\n+\n def p4_build_cmd(cmd):\n \"\"\"Build a suitable p4 command line.\n \n@@ -577,20 +580,12 @@ def p4_type(f):\n #\n def p4_keywords_regexp_for_type(base, type_mods):\n if base in (\"text\", \"unicode\", \"binary\"):\n- kwords = None\n if \"ko\" in type_mods:\n- kwords = 'Id|Header'\n+ return re_ko_keywords\n elif \"k\" in type_mods:\n- kwords = 'Id|Header|Author|Date|DateTime|Change|File|Revision'\n+ return re_k_keywords\n else:\n return None\n- pattern = r\"\"\"\n- \\$ # Starts with a dollar, followed by...\n- (%s) # one of the keywords, followed by...\n- (:[^$\\n]+)? # possibly an old expansion, followed by...\n- \\$ # another dollar\n- \"\"\" % kwords\n- return pattern\n else:\n return None\n \n@@ -1753,15 +1748,13 @@ def prepareLogMessage(self, template, message, jobs):\n \n return result\n \n- def patchRCSKeywords(self, file, pattern):\n- # Attempt to zap the RCS keywords in a p4 controlled file matching the given pattern\n+ def patchRCSKeywords(self, file, regexp):\n+ # Attempt to zap the RCS keywords in a p4 controlled file matching the given regex\n (handle, outFileName) = tempfile.mkstemp(dir='.')\n try:\n with os.fdopen(handle, \"w+\") as outFile, open(file, \"r\") as inFile:\n- regexp = re.compile(pattern, re.VERBOSE)\n for line in inFile.readlines():\n- line = regexp.sub(r'$\\1$', line)\n- outFile.write(line)\n+ outFile.write(regexp.sub(r'$\\1$', line))\n # Forcibly overwrite the original file\n os.unlink(file)\n shutil.move(outFileName, file)\n@@ -2088,25 +2081,22 @@ def applyCommit(self, id):\n # the patch to see if that's possible.\n if gitConfigBool(\"git-p4.attemptRCSCleanup\"):\n file = None\n- pattern = None\n kwfiles = {}\n for file in editedFiles | filesToDelete:\n # did this file's delta contain RCS keywords?\n- pattern = p4_keywords_regexp_for_file(file)\n-\n- if pattern:\n+ regexp = p4_keywords_regexp_for_file(file)\n+ if regexp:\n # this file is a possibility...look for RCS keywords.\n- regexp = re.compile(pattern, re.VERBOSE)\n for line in read_pipe_lines([\"git\", \"diff\", \"%s^..%s\" % (id, id), file]):\n if regexp.search(line):\n if verbose:\n- print(\"got keyword match on %s in %s in %s\" % (pattern, line, file))\n- kwfiles[file] = pattern\n+ print(\"got keyword match on %s in %s in %s\" % (regex.pattern, line, file))\n+ kwfiles[file] = regexp\n break\n \n- for file in kwfiles:\n+ for file, regexp in kwfiles.items():\n if verbose:\n- print(\"zapping %s with %s\" % (line,pattern))\n+ print(\"zapping %s with %s\" % (line, regexp.pattern))\n # File is being deleted, so not open in p4. Must\n # disable the read-only bit on windows.\n if self.isWindows and file not in editedFiles:\n@@ -3026,12 +3016,10 @@ def streamOneP4File(self, file, contents):\n \n # Note that we do not try to de-mangle keywords on utf16 files,\n # even though in theory somebody may want that.\n- pattern = p4_keywords_regexp_for_type(type_base, type_mods)\n- if pattern:\n- regexp = re.compile(pattern, re.VERBOSE)\n- text = ''.join(decode_text_stream(c) for c in contents)\n- text = regexp.sub(r'$\\1$', text)\n- contents = [ encode_text_stream(text) ]\n+ regexp = p4_keywords_regexp_for_type(type_base, type_mods)\n+ if regexp:\n+ contents = [encode_text_stream(regexp.sub(\n+ r'$\\1$', ''.join(decode_text_stream(c) for c in contents)))]\n \n if self.largeFileSystem:\n (git_mode, contents) = self.largeFileSystem.processContent(git_mode, relPath, contents)\n-- \n2.34.1\n\n"},{"id":"444222","messageId":"20211216134619.2048348-4-jholdsworth@nvidia.com","threadId":"57091","inReplyTo":"20211216134619.2048348-1-jholdsworth@nvidia.com","subject":"[PATCH v2 3/5] git-p4: add raw option to read_pipelines","fromName":"Joel Holdsworth","fromEmail":"jholdsworth@nvidia.com","sentAt":"2021-12-16T13:46:17Z","receivedAt":"2021-12-16T13:47:27Z","isPatch":true,"sender":{"key":"jholdsworth@nvidia.com","avatar":"https://avatars.githubusercontent.com/u/1449493?v=4"},"body":"Previously the read_lines function always decoded the result lines. In\norder to improve support for non-decoded binary processing of data in\ngit-p4.py, this patch adds a raw option to the function that allows\ndecoding to be disabled.\n\nSigned-off-by: Joel Holdsworth \n---\n git-p4.py | 8 +++++---\n 1 file changed, 5 insertions(+), 3 deletions(-)\n\ndiff --git a/git-p4.py b/git-p4.py\nindex 0af83b9c72..509feac2d8 100755\n--- a/git-p4.py\n+++ b/git-p4.py\n@@ -340,17 +340,19 @@ def p4_read_pipe(c, ignore_error=False, raw=False):\n real_cmd = p4_build_cmd(c)\n return read_pipe(real_cmd, ignore_error, raw=raw)\n \n-def read_pipe_lines(c):\n+def read_pipe_lines(c, raw=False):\n if verbose:\n sys.stderr.write('Reading pipe: %s\\n' % str(c))\n \n expand = not isinstance(c, list)\n p = subprocess.Popen(c, stdout=subprocess.PIPE, shell=expand)\n pipe = p.stdout\n- val = [decode_text_stream(line) for line in pipe.readlines()]\n+ lines = pipe.readlines()\n+ if not raw:\n+ lines = [decode_text_stream(line) for line in lines]\n if pipe.close() or p.wait():\n die('Command failed: %s' % str(c))\n- return val\n+ return lines\n \n def p4_read_pipe_lines(c):\n \"\"\"Specifically invoke p4 on the command supplied. \"\"\"\n-- \n2.34.1\n\n"},{"id":"444223","messageId":"20211216134619.2048348-5-jholdsworth@nvidia.com","threadId":"57091","inReplyTo":"20211216134619.2048348-1-jholdsworth@nvidia.com","subject":"[PATCH v2 4/5] git-p4: open temporary patch file for write only","fromName":"Joel Holdsworth","fromEmail":"jholdsworth@nvidia.com","sentAt":"2021-12-16T13:46:18Z","receivedAt":"2021-12-16T13:47:30Z","isPatch":true,"sender":{"key":"jholdsworth@nvidia.com","avatar":"https://avatars.githubusercontent.com/u/1449493?v=4"},"body":"The patchRCSKeywords method creates a temporary file in which to store\nthe patched output data. Previously this file was opened in \"w+\" mode\n(write and read), but the code never reads the contents of the file\nwhile open, so it only needs to be opened in \"w\" mode (write-only).\n\nSigned-off-by: Joel Holdsworth \n---\n git-p4.py | 2 +-\n 1 file changed, 1 insertion(+), 1 deletion(-)\n\ndiff --git a/git-p4.py b/git-p4.py\nindex 509feac2d8..7845210e69 100755\n--- a/git-p4.py\n+++ b/git-p4.py\n@@ -1754,7 +1754,7 @@ def patchRCSKeywords(self, file, regexp):\n # Attempt to zap the RCS keywords in a p4 controlled file matching the given regex\n (handle, outFileName) = tempfile.mkstemp(dir='.')\n try:\n- with os.fdopen(handle, \"w+\") as outFile, open(file, \"r\") as inFile:\n+ with os.fdopen(handle, \"w\") as outFile, open(file, \"r\") as inFile:\n for line in inFile.readlines():\n outFile.write(regexp.sub(r'$\\1$', line))\n # Forcibly overwrite the original file\n-- \n2.34.1\n\n"},{"id":"444224","messageId":"20211216134619.2048348-6-jholdsworth@nvidia.com","threadId":"57091","inReplyTo":"20211216134619.2048348-1-jholdsworth@nvidia.com","subject":"[PATCH v2 5/5] git-p4: resolve RCS keywords in bytes not utf-8","fromName":"Joel Holdsworth","fromEmail":"jholdsworth@nvidia.com","sentAt":"2021-12-16T13:46:19Z","receivedAt":"2021-12-16T13:47:33Z","isPatch":true,"sender":{"key":"jholdsworth@nvidia.com","avatar":"https://avatars.githubusercontent.com/u/1449493?v=4"},"body":"RCS keywords are strings that are replaced with information from\nPerforce. Examples include $Date$, $Author$, $File$, $Change$ etc.\n\nPerforce resolves these by expanding them with their expanded values\nwhen files are synced, but Git's data model requires these expanded\nvalues to be converted back into their unexpanded form.\n\nPreviously, git-p4.py would implement this behaviour through the use of\nregular expressions. However, the regular expression substitution was\napplied using decoded strings i.e. the content of incoming commit diffs\nwas first decoded from bytes into UTF-8, processed with regular\nexpressions, then converted back to bytes.\n\nNot only is this behaviour inefficient, but it is also a cause of a\ncommon issue caused by text files containing invalid UTF-8 data. For\nfiles created in Windows, CP1252 Smart Quote Characters (0x93 and 0x94)\nare seen fairly frequently. These codes are invalid in UTF-8, so if the\nscript encountered any file containing them, on Python 2 the symbols\nwill be corrupted, and on Python 3 the script will fail with an\nexception.\n\nThis patch replaces this decoding/encoding with bytes object regular\nexpressions, so that the substitution is performed directly upon the\nsource data with no conversions.\n\nA test for smart quote handling has been added to the\nt9810-git-p4-rcs.sh test suite.\n\nSigned-off-by: Joel Holdsworth \n---\n git-p4.py | 15 ++++++++-------\n t/t9810-git-p4-rcs.sh | 15 +++++++++++++++\n 2 files changed, 23 insertions(+), 7 deletions(-)\n\ndiff --git a/git-p4.py b/git-p4.py\nindex 7845210e69..986595bef0 100755\n--- a/git-p4.py\n+++ b/git-p4.py\n@@ -56,8 +56,8 @@\n \n p4_access_checked = False\n \n-re_ko_keywords = re.compile(r'\\$(Id|Header)(:[^$\\n]+)?\\$')\n-re_k_keywords = re.compile(r'\\$(Id|Header|Author|Date|DateTime|Change|File|Revision)(:[^$\\n]+)?\\$')\n+re_ko_keywords = re.compile(br'\\$(Id|Header)(:[^$\\n]+)?\\$')\n+re_k_keywords = re.compile(br'\\$(Id|Header|Author|Date|DateTime|Change|File|Revision)(:[^$\\n]+)?\\$')\n \n def p4_build_cmd(cmd):\n \"\"\"Build a suitable p4 command line.\n@@ -1754,9 +1754,9 @@ def patchRCSKeywords(self, file, regexp):\n # Attempt to zap the RCS keywords in a p4 controlled file matching the given regex\n (handle, outFileName) = tempfile.mkstemp(dir='.')\n try:\n- with os.fdopen(handle, \"w\") as outFile, open(file, \"r\") as inFile:\n+ with os.fdopen(handle, \"wb\") as outFile, open(file, \"rb\") as inFile:\n for line in inFile.readlines():\n- outFile.write(regexp.sub(r'$\\1$', line))\n+ outFile.write(regexp.sub(br'$\\1$', line))\n # Forcibly overwrite the original file\n os.unlink(file)\n shutil.move(outFileName, file)\n@@ -2089,7 +2089,9 @@ def applyCommit(self, id):\n regexp = p4_keywords_regexp_for_file(file)\n if regexp:\n # this file is a possibility...look for RCS keywords.\n- for line in read_pipe_lines([\"git\", \"diff\", \"%s^..%s\" % (id, id), file]):\n+ for line in read_pipe_lines(\n+ [\"git\", \"diff\", \"%s^..%s\" % (id, id), file],\n+ raw=True):\n if regexp.search(line):\n if verbose:\n print(\"got keyword match on %s in %s in %s\" % (regex.pattern, line, file))\n@@ -3020,8 +3022,7 @@ def streamOneP4File(self, file, contents):\n # even though in theory somebody may want that.\n regexp = p4_keywords_regexp_for_type(type_base, type_mods)\n if regexp:\n- contents = [encode_text_stream(regexp.sub(\n- r'$\\1$', ''.join(decode_text_stream(c) for c in contents)))]\n+ contents = [regexp.sub(br'$\\1$', c) for c in contents]\n \n if self.largeFileSystem:\n (git_mode, contents) = self.largeFileSystem.processContent(git_mode, relPath, contents)\ndiff --git a/t/t9810-git-p4-rcs.sh b/t/t9810-git-p4-rcs.sh\nindex e3836888ec..5fe83315ec 100755\n--- a/t/t9810-git-p4-rcs.sh\n+++ b/t/t9810-git-p4-rcs.sh\n@@ -4,6 +4,8 @@ test_description='git p4 rcs keywords'\n \n . ./lib-git-p4.sh\n \n+CP1252=\"\\223\\224\"\n+\n test_expect_success 'start p4d' '\n \tstart_p4d\n '\n@@ -32,6 +34,9 @@ test_expect_success 'init depot' '\n \t\tp4 submit -d \"filek\" &&\n \t\tp4 add -t text+ko fileko &&\n \t\tp4 submit -d \"fileko\" &&\n+\t\tprintf \"$CP1252\" >fileko_cp1252 &&\n+\t\tp4 add -t text+ko fileko_cp1252 &&\n+\t\tp4 submit -d \"fileko_cp1252\" &&\n \t\tp4 add -t text file_text &&\n \t\tp4 submit -d \"file_text\"\n \t)\n@@ -359,4 +364,14 @@ test_expect_failure 'Add keywords in git which do not match the default p4 value\n \t)\n '\n \n+test_expect_success 'check cp1252 smart quote are preserved through RCS keyword processing' '\n+\ttest_when_finished cleanup_git &&\n+\tgit p4 clone --dest=\"$git\" //depot &&\n+\t(\n+\t\tcd \"$git\" &&\n+\t\tprintf \"$CP1252\" >expect &&\n+\t\ttest_cmp_bin expect fileko_cp1252\n+\t)\n+'\n+\n test_done\n-- \n2.34.1\n\n"}]}