Simon Glass | 0d24de9 | 2012-01-14 15:12:45 +0000 | [diff] [blame] | 1 | # Copyright (c) 2011 The Chromium OS Authors. |
| 2 | # |
| 3 | # See file CREDITS for list of people who contributed to this |
| 4 | # project. |
| 5 | # |
| 6 | # This program is free software; you can redistribute it and/or |
| 7 | # modify it under the terms of the GNU General Public License as |
| 8 | # published by the Free Software Foundation; either version 2 of |
| 9 | # the License, or (at your option) any later version. |
| 10 | # |
| 11 | # This program is distributed in the hope that it will be useful, |
| 12 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 13 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 14 | # GNU General Public License for more details. |
| 15 | # |
| 16 | # You should have received a copy of the GNU General Public License |
| 17 | # along with this program; if not, write to the Free Software |
| 18 | # Foundation, Inc., 59 Temple Place, Suite 330, Boston, |
| 19 | # MA 02111-1307 USA |
| 20 | # |
| 21 | |
| 22 | import os |
| 23 | import re |
| 24 | import shutil |
| 25 | import tempfile |
| 26 | |
| 27 | import command |
| 28 | import commit |
| 29 | import gitutil |
| 30 | from series import Series |
| 31 | |
| 32 | # Tags that we detect and remove |
| 33 | re_remove = re.compile('^BUG=|^TEST=|^Change-Id:|^Review URL:' |
| 34 | '|Reviewed-on:|Reviewed-by:') |
| 35 | |
| 36 | # Lines which are allowed after a TEST= line |
| 37 | re_allowed_after_test = re.compile('^Signed-off-by:') |
| 38 | |
Ilya Yanok | 05e5b73 | 2012-08-06 23:46:05 +0000 | [diff] [blame] | 39 | # Signoffs |
| 40 | re_signoff = re.compile('^Signed-off-by:') |
| 41 | |
Simon Glass | 0d24de9 | 2012-01-14 15:12:45 +0000 | [diff] [blame] | 42 | # The start of the cover letter |
| 43 | re_cover = re.compile('^Cover-letter:') |
| 44 | |
| 45 | # Patch series tag |
| 46 | re_series = re.compile('^Series-(\w*): *(.*)') |
| 47 | |
| 48 | # Commit tags that we want to collect and keep |
Ilya Yanok | c737914 | 2012-08-06 23:46:08 +0000 | [diff] [blame^] | 49 | re_tag = re.compile('^(Tested-by|Acked-by|Cc): (.*)') |
Simon Glass | 0d24de9 | 2012-01-14 15:12:45 +0000 | [diff] [blame] | 50 | |
| 51 | # The start of a new commit in the git log |
| 52 | re_commit = re.compile('^commit (.*)') |
| 53 | |
| 54 | # We detect these since checkpatch doesn't always do it |
| 55 | re_space_before_tab = re.compile('^[+].* \t') |
| 56 | |
| 57 | # States we can be in - can we use range() and still have comments? |
| 58 | STATE_MSG_HEADER = 0 # Still in the message header |
| 59 | STATE_PATCH_SUBJECT = 1 # In patch subject (first line of log for a commit) |
| 60 | STATE_PATCH_HEADER = 2 # In patch header (after the subject) |
| 61 | STATE_DIFFS = 3 # In the diff part (past --- line) |
| 62 | |
| 63 | class PatchStream: |
| 64 | """Class for detecting/injecting tags in a patch or series of patches |
| 65 | |
| 66 | We support processing the output of 'git log' to read out the tags we |
| 67 | are interested in. We can also process a patch file in order to remove |
| 68 | unwanted tags or inject additional ones. These correspond to the two |
| 69 | phases of processing. |
| 70 | """ |
| 71 | def __init__(self, series, name=None, is_log=False): |
| 72 | self.skip_blank = False # True to skip a single blank line |
| 73 | self.found_test = False # Found a TEST= line |
| 74 | self.lines_after_test = 0 # MNumber of lines found after TEST= |
| 75 | self.warn = [] # List of warnings we have collected |
| 76 | self.linenum = 1 # Output line number we are up to |
| 77 | self.in_section = None # Name of start...END section we are in |
| 78 | self.notes = [] # Series notes |
| 79 | self.section = [] # The current section...END section |
| 80 | self.series = series # Info about the patch series |
| 81 | self.is_log = is_log # True if indent like git log |
| 82 | self.in_change = 0 # Non-zero if we are in a change list |
| 83 | self.blank_count = 0 # Number of blank lines stored up |
| 84 | self.state = STATE_MSG_HEADER # What state are we in? |
| 85 | self.tags = [] # Tags collected, like Tested-by... |
| 86 | self.signoff = [] # Contents of signoff line |
| 87 | self.commit = None # Current commit |
| 88 | |
| 89 | def AddToSeries(self, line, name, value): |
| 90 | """Add a new Series-xxx tag. |
| 91 | |
| 92 | When a Series-xxx tag is detected, we come here to record it, if we |
| 93 | are scanning a 'git log'. |
| 94 | |
| 95 | Args: |
| 96 | line: Source line containing tag (useful for debug/error messages) |
| 97 | name: Tag name (part after 'Series-') |
| 98 | value: Tag value (part after 'Series-xxx: ') |
| 99 | """ |
| 100 | if name == 'notes': |
| 101 | self.in_section = name |
| 102 | self.skip_blank = False |
| 103 | if self.is_log: |
| 104 | self.series.AddTag(self.commit, line, name, value) |
| 105 | |
| 106 | def CloseCommit(self): |
| 107 | """Save the current commit into our commit list, and reset our state""" |
| 108 | if self.commit and self.is_log: |
| 109 | self.series.AddCommit(self.commit) |
| 110 | self.commit = None |
| 111 | |
| 112 | def FormatTags(self, tags): |
| 113 | out_list = [] |
| 114 | for tag in sorted(tags): |
| 115 | if tag.startswith('Cc:'): |
| 116 | tag_list = tag[4:].split(',') |
| 117 | out_list += gitutil.BuildEmailList(tag_list, 'Cc:') |
| 118 | else: |
| 119 | out_list.append(tag) |
| 120 | return out_list |
| 121 | |
| 122 | def ProcessLine(self, line): |
| 123 | """Process a single line of a patch file or commit log |
| 124 | |
| 125 | This process a line and returns a list of lines to output. The list |
| 126 | may be empty or may contain multiple output lines. |
| 127 | |
| 128 | This is where all the complicated logic is located. The class's |
| 129 | state is used to move between different states and detect things |
| 130 | properly. |
| 131 | |
| 132 | We can be in one of two modes: |
| 133 | self.is_log == True: This is 'git log' mode, where most output is |
| 134 | indented by 4 characters and we are scanning for tags |
| 135 | |
| 136 | self.is_log == False: This is 'patch' mode, where we already have |
| 137 | all the tags, and are processing patches to remove junk we |
| 138 | don't want, and add things we think are required. |
| 139 | |
| 140 | Args: |
| 141 | line: text line to process |
| 142 | |
| 143 | Returns: |
| 144 | list of output lines, or [] if nothing should be output |
| 145 | """ |
| 146 | # Initially we have no output. Prepare the input line string |
| 147 | out = [] |
| 148 | line = line.rstrip('\n') |
| 149 | if self.is_log: |
| 150 | if line[:4] == ' ': |
| 151 | line = line[4:] |
| 152 | |
| 153 | # Handle state transition and skipping blank lines |
| 154 | series_match = re_series.match(line) |
| 155 | commit_match = re_commit.match(line) if self.is_log else None |
| 156 | tag_match = None |
| 157 | if self.state == STATE_PATCH_HEADER: |
| 158 | tag_match = re_tag.match(line) |
| 159 | is_blank = not line.strip() |
| 160 | if is_blank: |
| 161 | if (self.state == STATE_MSG_HEADER |
| 162 | or self.state == STATE_PATCH_SUBJECT): |
| 163 | self.state += 1 |
| 164 | |
| 165 | # We don't have a subject in the text stream of patch files |
| 166 | # It has its own line with a Subject: tag |
| 167 | if not self.is_log and self.state == STATE_PATCH_SUBJECT: |
| 168 | self.state += 1 |
| 169 | elif commit_match: |
| 170 | self.state = STATE_MSG_HEADER |
| 171 | |
| 172 | # If we are in a section, keep collecting lines until we see END |
| 173 | if self.in_section: |
| 174 | if line == 'END': |
| 175 | if self.in_section == 'cover': |
| 176 | self.series.cover = self.section |
| 177 | elif self.in_section == 'notes': |
| 178 | if self.is_log: |
| 179 | self.series.notes += self.section |
| 180 | else: |
| 181 | self.warn.append("Unknown section '%s'" % self.in_section) |
| 182 | self.in_section = None |
| 183 | self.skip_blank = True |
| 184 | self.section = [] |
| 185 | else: |
| 186 | self.section.append(line) |
| 187 | |
| 188 | # Detect the commit subject |
| 189 | elif not is_blank and self.state == STATE_PATCH_SUBJECT: |
| 190 | self.commit.subject = line |
| 191 | |
| 192 | # Detect the tags we want to remove, and skip blank lines |
| 193 | elif re_remove.match(line): |
| 194 | self.skip_blank = True |
| 195 | |
| 196 | # TEST= should be the last thing in the commit, so remove |
| 197 | # everything after it |
| 198 | if line.startswith('TEST='): |
| 199 | self.found_test = True |
| 200 | elif self.skip_blank and is_blank: |
| 201 | self.skip_blank = False |
| 202 | |
| 203 | # Detect the start of a cover letter section |
| 204 | elif re_cover.match(line): |
| 205 | self.in_section = 'cover' |
| 206 | self.skip_blank = False |
| 207 | |
| 208 | # If we are in a change list, key collected lines until a blank one |
| 209 | elif self.in_change: |
| 210 | if is_blank: |
| 211 | # Blank line ends this change list |
| 212 | self.in_change = 0 |
Ilya Yanok | 05e5b73 | 2012-08-06 23:46:05 +0000 | [diff] [blame] | 213 | elif line == '---' or re_signoff.match(line): |
| 214 | self.in_change = 0 |
| 215 | out = self.ProcessLine(line) |
Simon Glass | 0d24de9 | 2012-01-14 15:12:45 +0000 | [diff] [blame] | 216 | else: |
Ilya Yanok | a8840cb | 2012-08-06 23:46:06 +0000 | [diff] [blame] | 217 | if self.is_log: |
| 218 | self.series.AddChange(self.in_change, self.commit, line) |
Simon Glass | 0d24de9 | 2012-01-14 15:12:45 +0000 | [diff] [blame] | 219 | self.skip_blank = False |
| 220 | |
| 221 | # Detect Series-xxx tags |
| 222 | elif series_match: |
| 223 | name = series_match.group(1) |
| 224 | value = series_match.group(2) |
| 225 | if name == 'changes': |
| 226 | # value is the version number: e.g. 1, or 2 |
| 227 | try: |
| 228 | value = int(value) |
| 229 | except ValueError as str: |
| 230 | raise ValueError("%s: Cannot decode version info '%s'" % |
| 231 | (self.commit.hash, line)) |
| 232 | self.in_change = int(value) |
| 233 | else: |
| 234 | self.AddToSeries(line, name, value) |
| 235 | self.skip_blank = True |
| 236 | |
| 237 | # Detect the start of a new commit |
| 238 | elif commit_match: |
| 239 | self.CloseCommit() |
| 240 | self.commit = commit.Commit(commit_match.group(1)[:7]) |
| 241 | |
| 242 | # Detect tags in the commit message |
| 243 | elif tag_match: |
Simon Glass | 0d24de9 | 2012-01-14 15:12:45 +0000 | [diff] [blame] | 244 | # Remove Tested-by self, since few will take much notice |
Ilya Yanok | c737914 | 2012-08-06 23:46:08 +0000 | [diff] [blame^] | 245 | if (tag_match.group(1) == 'Tested-by' and |
Simon Glass | 0d24de9 | 2012-01-14 15:12:45 +0000 | [diff] [blame] | 246 | tag_match.group(2).find(os.getenv('USER') + '@') != -1): |
| 247 | self.warn.append("Ignoring %s" % line) |
| 248 | elif tag_match.group(1) == 'Cc': |
| 249 | self.commit.AddCc(tag_match.group(2).split(',')) |
| 250 | else: |
| 251 | self.tags.append(line); |
| 252 | |
| 253 | # Well that means this is an ordinary line |
| 254 | else: |
| 255 | pos = 1 |
| 256 | # Look for ugly ASCII characters |
| 257 | for ch in line: |
| 258 | # TODO: Would be nicer to report source filename and line |
| 259 | if ord(ch) > 0x80: |
| 260 | self.warn.append("Line %d/%d ('%s') has funny ascii char" % |
| 261 | (self.linenum, pos, line)) |
| 262 | pos += 1 |
| 263 | |
| 264 | # Look for space before tab |
| 265 | m = re_space_before_tab.match(line) |
| 266 | if m: |
| 267 | self.warn.append('Line %d/%d has space before tab' % |
| 268 | (self.linenum, m.start())) |
| 269 | |
| 270 | # OK, we have a valid non-blank line |
| 271 | out = [line] |
| 272 | self.linenum += 1 |
| 273 | self.skip_blank = False |
| 274 | if self.state == STATE_DIFFS: |
| 275 | pass |
| 276 | |
| 277 | # If this is the start of the diffs section, emit our tags and |
| 278 | # change log |
| 279 | elif line == '---': |
| 280 | self.state = STATE_DIFFS |
| 281 | |
| 282 | # Output the tags (signeoff first), then change list |
| 283 | out = [] |
Simon Glass | 0d24de9 | 2012-01-14 15:12:45 +0000 | [diff] [blame] | 284 | log = self.series.MakeChangeLog(self.commit) |
| 285 | out += self.FormatTags(self.tags) |
| 286 | out += [line] + log |
| 287 | elif self.found_test: |
| 288 | if not re_allowed_after_test.match(line): |
| 289 | self.lines_after_test += 1 |
| 290 | |
| 291 | return out |
| 292 | |
| 293 | def Finalize(self): |
| 294 | """Close out processing of this patch stream""" |
| 295 | self.CloseCommit() |
| 296 | if self.lines_after_test: |
| 297 | self.warn.append('Found %d lines after TEST=' % |
| 298 | self.lines_after_test) |
| 299 | |
| 300 | def ProcessStream(self, infd, outfd): |
| 301 | """Copy a stream from infd to outfd, filtering out unwanting things. |
| 302 | |
| 303 | This is used to process patch files one at a time. |
| 304 | |
| 305 | Args: |
| 306 | infd: Input stream file object |
| 307 | outfd: Output stream file object |
| 308 | """ |
| 309 | # Extract the filename from each diff, for nice warnings |
| 310 | fname = None |
| 311 | last_fname = None |
| 312 | re_fname = re.compile('diff --git a/(.*) b/.*') |
| 313 | while True: |
| 314 | line = infd.readline() |
| 315 | if not line: |
| 316 | break |
| 317 | out = self.ProcessLine(line) |
| 318 | |
| 319 | # Try to detect blank lines at EOF |
| 320 | for line in out: |
| 321 | match = re_fname.match(line) |
| 322 | if match: |
| 323 | last_fname = fname |
| 324 | fname = match.group(1) |
| 325 | if line == '+': |
| 326 | self.blank_count += 1 |
| 327 | else: |
| 328 | if self.blank_count and (line == '-- ' or match): |
| 329 | self.warn.append("Found possible blank line(s) at " |
| 330 | "end of file '%s'" % last_fname) |
| 331 | outfd.write('+\n' * self.blank_count) |
| 332 | outfd.write(line + '\n') |
| 333 | self.blank_count = 0 |
| 334 | self.Finalize() |
| 335 | |
| 336 | |
| 337 | def GetMetaData(start, count): |
| 338 | """Reads out patch series metadata from the commits |
| 339 | |
| 340 | This does a 'git log' on the relevant commits and pulls out the tags we |
| 341 | are interested in. |
| 342 | |
| 343 | Args: |
| 344 | start: Commit to start from: 0=HEAD, 1=next one, etc. |
| 345 | count: Number of commits to list |
| 346 | """ |
| 347 | pipe = [['git', 'log', '--reverse', 'HEAD~%d' % start, '-n%d' % count]] |
| 348 | stdout = command.RunPipe(pipe, capture=True) |
| 349 | series = Series() |
| 350 | ps = PatchStream(series, is_log=True) |
| 351 | for line in stdout.splitlines(): |
| 352 | ps.ProcessLine(line) |
| 353 | ps.Finalize() |
| 354 | return series |
| 355 | |
| 356 | def FixPatch(backup_dir, fname, series, commit): |
| 357 | """Fix up a patch file, by adding/removing as required. |
| 358 | |
| 359 | We remove our tags from the patch file, insert changes lists, etc. |
| 360 | The patch file is processed in place, and overwritten. |
| 361 | |
| 362 | A backup file is put into backup_dir (if not None). |
| 363 | |
| 364 | Args: |
| 365 | fname: Filename to patch file to process |
| 366 | series: Series information about this patch set |
| 367 | commit: Commit object for this patch file |
| 368 | Return: |
| 369 | A list of errors, or [] if all ok. |
| 370 | """ |
| 371 | handle, tmpname = tempfile.mkstemp() |
| 372 | outfd = os.fdopen(handle, 'w') |
| 373 | infd = open(fname, 'r') |
| 374 | ps = PatchStream(series) |
| 375 | ps.commit = commit |
| 376 | ps.ProcessStream(infd, outfd) |
| 377 | infd.close() |
| 378 | outfd.close() |
| 379 | |
| 380 | # Create a backup file if required |
| 381 | if backup_dir: |
| 382 | shutil.copy(fname, os.path.join(backup_dir, os.path.basename(fname))) |
| 383 | shutil.move(tmpname, fname) |
| 384 | return ps.warn |
| 385 | |
| 386 | def FixPatches(series, fnames): |
| 387 | """Fix up a list of patches identified by filenames |
| 388 | |
| 389 | The patch files are processed in place, and overwritten. |
| 390 | |
| 391 | Args: |
| 392 | series: The series object |
| 393 | fnames: List of patch files to process |
| 394 | """ |
| 395 | # Current workflow creates patches, so we shouldn't need a backup |
| 396 | backup_dir = None #tempfile.mkdtemp('clean-patch') |
| 397 | count = 0 |
| 398 | for fname in fnames: |
| 399 | commit = series.commits[count] |
| 400 | commit.patch = fname |
| 401 | result = FixPatch(backup_dir, fname, series, commit) |
| 402 | if result: |
| 403 | print '%d warnings for %s:' % (len(result), fname) |
| 404 | for warn in result: |
| 405 | print '\t', warn |
| 406 | print |
| 407 | count += 1 |
| 408 | print 'Cleaned %d patches' % count |
| 409 | return series |
| 410 | |
| 411 | def InsertCoverLetter(fname, series, count): |
| 412 | """Inserts a cover letter with the required info into patch 0 |
| 413 | |
| 414 | Args: |
| 415 | fname: Input / output filename of the cover letter file |
| 416 | series: Series object |
| 417 | count: Number of patches in the series |
| 418 | """ |
| 419 | fd = open(fname, 'r') |
| 420 | lines = fd.readlines() |
| 421 | fd.close() |
| 422 | |
| 423 | fd = open(fname, 'w') |
| 424 | text = series.cover |
| 425 | prefix = series.GetPatchPrefix() |
| 426 | for line in lines: |
| 427 | if line.startswith('Subject:'): |
| 428 | # TODO: if more than 10 patches this should save 00/xx, not 0/xx |
| 429 | line = 'Subject: [%s 0/%d] %s\n' % (prefix, count, text[0]) |
| 430 | |
| 431 | # Insert our cover letter |
| 432 | elif line.startswith('*** BLURB HERE ***'): |
| 433 | # First the blurb test |
| 434 | line = '\n'.join(text[1:]) + '\n' |
| 435 | if series.get('notes'): |
| 436 | line += '\n'.join(series.notes) + '\n' |
| 437 | |
| 438 | # Now the change list |
| 439 | out = series.MakeChangeLog(None) |
| 440 | line += '\n' + '\n'.join(out) |
| 441 | fd.write(line) |
| 442 | fd.close() |