src/svn.cpp

   1 /*
   2  *  Copyright (C) 2007  Thiago Macieira <thiago@kde.org>
   3  *
   4  *  This program is free software: you can redistribute it and/or modify
   5  *  it under the terms of the GNU General Public License as published by
   6  *  the Free Software Foundation, either version 2 of the License, or
   7  *  (at your option) any later version.
   8  *
   9  *  This program is distributed in the hope that it will be useful,
  10  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  11  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12  *  GNU General Public License for more details.
  13  *
  14  *  You should have received a copy of the GNU General Public License
  15  *  along with this program.  If not, see <http://www.gnu.org/licenses/>.
  16  */
  17
  18 /*
  19  * Based on svn-fast-export by Chris Lee <clee@kde.org>
  20  * License: MIT <http://www.opensource.org/licenses/mit-license.php>
  21  * URL: git://repo.or.cz/fast-import.git http://repo.or.cz/w/fast-export.git
  22  */
  23
  24 #define _XOPEN_SOURCE
  25 #define _LARGEFILE_SUPPORT
  26 #define _LARGEFILE64_SUPPORT
  27
  28 #include "svn.h"
  29
  30 #include <unistd.h>
  31 #include <string.h>
  32 #include <stdio.h>
  33 #include <time.h>
  34 #include <unistd.h>
  35
  36 #include <apr_lib.h>
  37 #include <apr_getopt.h>
  38 #include <apr_general.h>
  39
  40 #include <svn_fs.h>
  41 #include <svn_pools.h>
  42 #include <svn_repos.h>
  43 #include <svn_types.h>
  44
  45 #include <QFile>
  46 #include <QDebug>
  47
  48 #include "repository.h"
  49
  50 #undef SVN_ERR
  51 #define SVN_ERR(expr) SVN_INT_ERR(expr)
  52
  53 typedef QList<Rules::Match> MatchRuleList;
  54 typedef QHash<QString, Repository *> RepositoryHash;
  55 typedef QHash<QByteArray, QByteArray> IdentityHash;
  56
  57 class AprAutoPool
  58 {
  59     apr_pool_t *pool;
  60     AprAutoPool(const AprAutoPool &);
  61     AprAutoPool &operator=(const AprAutoPool &);
  62 public:
  63     inline AprAutoPool(apr_pool_t *parent = NULL)
  64         { pool = svn_pool_create(parent); }
  65     inline ~AprAutoPool()
  66         { svn_pool_destroy(pool); }
  67
  68     inline void clear() { svn_pool_clear(pool); }
  69     inline apr_pool_t *data() const { return pool; }
  70     inline operator apr_pool_t *() const { return pool; }
  71 };
  72
  73 class SvnPrivate
  74 {
  75 public:
  76     MatchRuleList matchRules;
  77     RepositoryHash repositories;
  78     IdentityHash identities;
  79
  80     SvnPrivate(const QString &pathToRepository);
  81     ~SvnPrivate();
  82     int youngestRevision();
  83     int exportRevision(int revnum);
  84
  85     int openRepository(const QString &pathToRepository);
  86
  87 private:
  88     AprAutoPool global_pool;
  89     svn_fs_t *fs;
  90     svn_revnum_t youngest_rev;
  91 };
  92
  93 void Svn::initialize()
  94 {
  95     // initialize APR or exit
  96     if (apr_initialize() != APR_SUCCESS) {
  97         fprintf(stderr, "You lose at apr_initialize().\n");
  98         exit(1);
  99     }
 100
 101     // static destructor
 102     static struct Destructor { ~Destructor() { apr_terminate(); } } destructor;
 103 }
 104
 105 Svn::Svn(const QString &pathToRepository)
 106     : d(new SvnPrivate(pathToRepository))
 107 {
 108 }
 109
 110 Svn::~Svn()
 111 {
 112     delete d;
 113 }
 114
 115 void Svn::setMatchRules(const MatchRuleList &matchRules)
 116 {
 117     d->matchRules = matchRules;
 118 }
 119
 120 void Svn::setRepositories(const RepositoryHash &repositories)
 121 {
 122     d->repositories = repositories;
 123 }
 124
 125 int Svn::youngestRevision()
 126 {
 127     return d->youngestRevision();
 128 }
 129
 130 bool Svn::exportRevision(int revnum)
 131 {
 132     return d->exportRevision(revnum) == EXIT_SUCCESS;
 133 }
 134
 135 SvnPrivate::SvnPrivate(const QString &pathToRepository)
 136     : global_pool(NULL)
 137 {
 138     openRepository(pathToRepository);
 139
 140     // get the youngest revision
 141     svn_fs_youngest_rev(&youngest_rev, fs, global_pool);
 142 }
 143
 144 SvnPrivate::~SvnPrivate()
 145 {
 146     svn_pool_destroy(global_pool);
 147 }
 148
 149 int SvnPrivate::youngestRevision()
 150 {
 151     return youngest_rev;
 152 }
 153
 154 int SvnPrivate::openRepository(const QString &pathToRepository)
 155 {
 156     svn_repos_t *repos;
 157     SVN_ERR(svn_repos_open(&repos, QFile::encodeName(pathToRepository), global_pool));
 158     fs = svn_repos_fs(repos);
 159
 160     return EXIT_SUCCESS;
 161 }
 162
 163 enum RuleType { AnyRule = 0, NoIgnoreRule = 0x01, NoRecurseRule = 0x02 };
 164
 165 static MatchRuleList::ConstIterator
 166 findMatchRule(const MatchRuleList &matchRules, int revnum, const QString &current,
 167               int ruleMask = AnyRule)
 168 {
 169     MatchRuleList::ConstIterator it = matchRules.constBegin(),
 170                                 end = matchRules.constEnd();
 171     for ( ; it != end; ++it) {
 172         if (it->minRevision > revnum)
 173             continue;
 174         if (it->maxRevision != -1 && it->maxRevision < revnum)
 175             continue;
 176         if (it->action == Rules::Match::Ignore && ruleMask & NoIgnoreRule)
 177             continue;
 178         if (it->action == Rules::Match::Recurse && ruleMask & NoRecurseRule)
 179             continue;
 180         if (it->rx.indexIn(current) == 0)
 181             return it;
 182     }
 183
 184     // no match
 185     return end;
 186 }
 187
 188 static void splitPathName(const Rules::Match &rule, const QString &pathName, QString *svnprefix_p,
 189                           QString *repository_p, QString *branch_p, QString *path_p)
 190 {
 191     QString svnprefix = pathName;
 192     svnprefix.truncate(rule.rx.matchedLength());
 193     if (svnprefix_p)
 194         *svnprefix_p = svnprefix;
 195
 196     if (repository_p) {
 197         *repository_p = svnprefix;
 198         repository_p->replace(rule.rx, rule.repository);
 199     }
 200
 201     if (branch_p) {
 202         *branch_p = svnprefix;
 203         branch_p->replace(rule.rx, rule.branch);
 204     }
 205
 206     if (path_p)
 207         *path_p = pathName.mid(svnprefix.length());
 208 }
 209
 210 static int pathMode(svn_fs_root_t *fs_root, const char *pathname, apr_pool_t *pool)
 211 {
 212     svn_string_t *propvalue;
 213     SVN_ERR(svn_fs_node_prop(&propvalue, fs_root, pathname, "svn:executable", pool));
 214     int mode = 0100644;
 215     if (propvalue)
 216         mode = 0100755;
 217
 218     // maybe it's a symlink?
 219     SVN_ERR(svn_fs_node_prop(&propvalue, fs_root, pathname, "svn:special", pool));
 220     if (propvalue && strcmp(propvalue->data, "symlink") == 0)
 221         mode = 0120000;
 222
 223     return mode;
 224 }
 225
 226 svn_error_t *QIODevice_write(void *baton, const char *data, apr_size_t *len)
 227 {
 228     QIODevice *device = reinterpret_cast<QIODevice *>(baton);
 229     device->write(data, *len);
 230
 231     if (device->bytesToWrite() > 16384)
 232         device->waitForBytesWritten(0);
 233     return SVN_NO_ERROR;
 234 }
 235
 236 static svn_stream_t *streamForDevice(QIODevice *device, apr_pool_t *pool)
 237 {
 238     svn_stream_t *stream = svn_stream_create(device, pool);
 239     svn_stream_set_write(stream, QIODevice_write);
 240
 241     return stream;
 242 }
 243
 244 static int dumpBlob(Repository::Transaction *txn, svn_fs_root_t *fs_root,
 245                     const char *pathname, const QString &finalPathName, apr_pool_t *pool)
 246 {
 247     AprAutoPool dumppool(pool);
 248     // what type is it?
 249     int mode = pathMode(fs_root, pathname, dumppool);
 250
 251     svn_filesize_t stream_length;
 252
 253     SVN_ERR(svn_fs_file_length(&stream_length, fs_root, pathname, dumppool));
 254     QIODevice *io = txn->addFile(finalPathName, mode, stream_length);
 255
 256 #ifndef DRY_RUN
 257     // open the file
 258     svn_stream_t *in_stream, *out_stream;
 259     SVN_ERR(svn_fs_file_contents(&in_stream, fs_root, pathname, dumppool));
 260
 261     // open a generic svn_stream_t for the QIODevice
 262     out_stream = streamForDevice(io, dumppool);
 263     SVN_ERR(svn_stream_copy(in_stream, out_stream, dumppool));
 264
 265     // print an ending newline
 266     io->putChar('\n');
 267 #endif
 268
 269     return EXIT_SUCCESS;
 270 }
 271
 272 static int recursiveDumpDir(Repository::Transaction *txn, svn_fs_root_t *fs_root,
 273                             const QByteArray &pathname, const QString &finalPathName,
 274                             apr_pool_t *pool)
 275 {
 276     // get the dir listing
 277     apr_hash_t *entries;
 278     SVN_ERR(svn_fs_dir_entries(&entries, fs_root, pathname, pool));
 279     AprAutoPool dirpool(pool);
 280
 281     for (apr_hash_index_t *i = apr_hash_first(pool, entries); i; i = apr_hash_next(i)) {
 282         dirpool.clear();
 283         const void *vkey;
 284         void *value;
 285         apr_hash_this(i, &vkey, NULL, &value);
 286
 287         svn_fs_dirent_t *dirent = reinterpret_cast<svn_fs_dirent_t *>(value);
 288         QByteArray entryName = pathname + '/' + dirent->name;
 289         QString entryFinalName;
 290         if (finalPathName.isEmpty())
 291             entryFinalName = dirent->name;
 292         else
 293             entryFinalName = finalPathName + '/' + dirent->name;
 294
 295         if (dirent->kind == svn_node_dir) {
 296             if (recursiveDumpDir(txn, fs_root, entryName, entryFinalName, dirpool) == EXIT_FAILURE)
 297                 return EXIT_FAILURE;
 298         } else if (dirent->kind == svn_node_file) {
 299             printf("+");
 300             fflush(stdout);
 301             if (dumpBlob(txn, fs_root, entryName, entryFinalName, dirpool) == EXIT_FAILURE)
 302                 return EXIT_FAILURE;
 303         }
 304     }
 305
 306     return EXIT_SUCCESS;
 307 }
 308
 309 static bool wasDir(svn_fs_t *fs, int revnum, const char *pathname, apr_pool_t *pool)
 310 {
 311     AprAutoPool subpool(pool);
 312     svn_fs_root_t *fs_root;
 313     if (svn_fs_revision_root(&fs_root, fs, revnum, subpool) != SVN_NO_ERROR)
 314         return false;
 315
 316     svn_boolean_t is_dir;
 317     if (svn_fs_is_dir(&is_dir, fs_root, pathname, subpool) != SVN_NO_ERROR)
 318         return false;
 319
 320     return is_dir;
 321 }
 322
 323 time_t get_epoch(char *svn_date)
 324 {
 325     struct tm tm;
 326     memset(&tm, 0, sizeof tm);
 327     QByteArray date(svn_date, strlen(svn_date) - 8);
 328     strptime(date, "%Y-%m-%dT%H:%M:%S", &tm);
 329     return mktime(&tm);
 330 }
 331
 332 class SvnRevision
 333 {
 334 public:
 335     AprAutoPool pool;
 336     QHash<QString, Repository::Transaction *> transactions;
 337     MatchRuleList matchRules;
 338     RepositoryHash repositories;
 339     IdentityHash identities;
 340
 341     svn_fs_t *fs;
 342     svn_fs_root_t *fs_root;
 343     int revnum;
 344
 345     SvnRevision(int revision, svn_fs_t *f, apr_pool_t *parent_pool)
 346         : pool(parent_pool), fs(f), fs_root(0), revnum(revision)
 347     {
 348     }
 349
 350     int open()
 351     {
 352         SVN_ERR(svn_fs_revision_root(&fs_root, fs, revnum, pool));
 353         return EXIT_SUCCESS;
 354     }
 355
 356     int prepareTransactions();
 357     int commit();
 358
 359     int exportEntry(const char *path, const svn_fs_path_change_t *change);
 360     int exportInternal(const char *path, const svn_fs_path_change_t *change,
 361                        const char *path_from, svn_revnum_t rev_from,
 362                        const QString &current, const Rules::Match &rule);
 363     int recurse(const char *path, const svn_fs_path_change_t *change,
 364                 const char *path_from, svn_revnum_t rev_from, apr_pool_t *pool);
 365 };
 366
 367 int SvnPrivate::exportRevision(int revnum)
 368 {
 369     SvnRevision rev(revnum, fs, global_pool);
 370     rev.matchRules = matchRules;
 371     rev.repositories = repositories;
 372     rev.identities = identities;
 373
 374     // open this revision:
 375     printf("Exporting revision %d ", revnum);
 376     fflush(stdout);
 377
 378     if (rev.open() == EXIT_FAILURE)
 379         return EXIT_FAILURE;
 380
 381     if (rev.prepareTransactions() == EXIT_FAILURE)
 382         return EXIT_FAILURE;
 383
 384     if (rev.transactions.isEmpty()) {
 385         printf(" nothing to do\n");
 386         return EXIT_SUCCESS;    // no changes?
 387     }
 388
 389     if (rev.commit() == EXIT_FAILURE)
 390         return EXIT_FAILURE;
 391
 392     printf(" done\n");
 393     return EXIT_SUCCESS;
 394 }
 395
 396 int SvnRevision::prepareTransactions()
 397 {
 398     // find out what was changed in this revision:
 399     apr_hash_t *changes;
 400     SVN_ERR(svn_fs_paths_changed(&changes, fs_root, pool));
 401     for (apr_hash_index_t *i = apr_hash_first(pool, changes); i; i = apr_hash_next(i)) {
 402         const void *vkey;
 403         void *value;
 404         apr_hash_this(i, &vkey, NULL, &value);
 405         const char *key = reinterpret_cast<const char *>(vkey);
 406         svn_fs_path_change_t *change = reinterpret_cast<svn_fs_path_change_t *>(value);
 407
 408         if (exportEntry(key, change) == EXIT_FAILURE)
 409             return EXIT_FAILURE;
 410     }
 411
 412     return EXIT_SUCCESS;
 413 }
 414
 415 int SvnRevision::commit()
 416 {
 417     // now create the commit
 418     apr_hash_t *revprops;
 419     SVN_ERR(svn_fs_revision_proplist(&revprops, fs, revnum, pool));
 420     svn_string_t *svnauthor = (svn_string_t*)apr_hash_get(revprops, "svn:author", APR_HASH_KEY_STRING);
 421     svn_string_t *svndate = (svn_string_t*)apr_hash_get(revprops, "svn:date", APR_HASH_KEY_STRING);
 422     svn_string_t *svnlog = (svn_string_t*)apr_hash_get(revprops, "svn:log", APR_HASH_KEY_STRING);
 423
 424     QByteArray log = (char *)svnlog->data;
 425     QByteArray authorident = svnauthor ? identities.value((char *)svnauthor->data) : QByteArray();
 426     time_t epoch = get_epoch((char*)svndate->data);
 427     if (authorident.isEmpty()) {
 428         if (!svnauthor || svn_string_isempty(svnauthor))
 429             authorident = "nobody <nobody@localhost>";
 430         else
 431             authorident = svnauthor->data + QByteArray(" <") +
 432                           svnauthor->data + QByteArray("@localhost>");
 433     }
 434
 435     foreach (Repository::Transaction *txn, transactions) {
 436         txn->setAuthor(authorident);
 437         txn->setDateTime(epoch);
 438         txn->setLog(log);
 439
 440         txn->commit();
 441         delete txn;
 442     }
 443
 444     return EXIT_SUCCESS;
 445 }
 446
 447 int SvnRevision::exportEntry(const char *key, const svn_fs_path_change_t *change)
 448 {
 449     AprAutoPool revpool(pool.data());
 450     QString current = QString::fromUtf8(key);
 451
 452     // was this copied from somewhere?
 453     svn_revnum_t rev_from;
 454     const char *path_from;
 455     SVN_ERR(svn_fs_copied_from(&rev_from, &path_from, fs_root, key, revpool));
 456
 457     // is this a directory?
 458     svn_boolean_t is_dir;
 459     SVN_ERR(svn_fs_is_dir(&is_dir, fs_root, key, revpool));
 460     if (is_dir) {
 461         if (path_from == NULL) {
 462             // no, it's a new directory being added
 463             // Git doesn't handle directories, so we don't either
 464             //qDebug() << "   mkdir ignored:" << key;
 465             return EXIT_SUCCESS;
 466         }
 467
 468         current += '/';
 469         qDebug() << "   " << key << "was copied from" << path_from;
 470     }
 471
 472     // find the first rule that matches this pathname
 473     MatchRuleList::ConstIterator match = findMatchRule(matchRules, revnum, current);
 474     if (match != matchRules.constEnd()) {
 475         const Rules::Match &rule = *match;
 476         switch (rule.action) {
 477         case Rules::Match::Ignore:
 478             // ignore rule
 479             qDebug() << "   " << qPrintable(current) << "rev" << revnum
 480                      << "-> ignored (rule line" << rule.lineNumber << ")";
 481             return EXIT_SUCCESS;
 482
 483         case Rules::Match::Recurse:
 484             // recurse rule
 485             if (is_dir)
 486                 return recurse(key, change, path_from, rev_from, revpool);
 487             if (change->change_kind != svn_fs_path_change_delete)
 488                 qWarning() << "   recurse rule " << rule.rx.pattern() << "line" << rule.lineNumber
 489                            << "applied to non-directory:" << qPrintable(current);
 490             return EXIT_SUCCESS;
 491
 492         case Rules::Match::Export:
 493             return exportInternal(key, change, path_from, rev_from, current, rule);
 494         }
 495     }
 496
 497     if (is_dir && path_from != NULL) {
 498         qDebug() << current << "is a copy-with-history, auto-recursing";
 499         return recurse(key, change, path_from, rev_from, revpool);
 500     } else if (wasDir(fs, revnum - 1, key, revpool)) {
 501         qDebug() << current << "was a directory; ignoring";
 502     } else if (change->change_kind == svn_fs_path_change_delete) {
 503         qDebug() << current << "is being deleted but I don't know anything about it; ignoring";
 504     } else {
 505         qCritical() << current << "did not match any rules; cannot continue";
 506         return EXIT_FAILURE;
 507     }
 508
 509     return EXIT_SUCCESS;
 510 }
 511
 512 int SvnRevision::exportInternal(const char *key, const svn_fs_path_change_t *change,
 513                                 const char *path_from, svn_revnum_t rev_from,
 514                                 const QString &current, const Rules::Match &rule)
 515 {
 516     QString svnprefix, repository, branch, path;
 517     splitPathName(rule, current, &svnprefix, &repository, &branch, &path);
 518
 519     printf(".");
 520     fflush(stdout);
 521 //                qDebug() << "   " << qPrintable(current) << "rev" << revnum << "->"
 522 //                         << qPrintable(repository) << qPrintable(branch) << qPrintable(path);
 523
 524     if (path.isEmpty() && path_from != NULL) {
 525         QString previous = QString::fromUtf8(path_from) + '/';
 526         MatchRuleList::ConstIterator prevmatch =
 527             findMatchRule(matchRules, rev_from, previous, NoRecurseRule | NoIgnoreRule);
 528         if (prevmatch != matchRules.constEnd()) {
 529             QString prevsvnprefix, prevrepository, prevbranch, prevpath;
 530             splitPathName(*prevmatch, previous, &prevsvnprefix, &prevrepository,
 531                           &prevbranch, &prevpath);
 532
 533             if (!prevpath.isEmpty()) {
 534                 qDebug() << qPrintable(current) << "is a partial branch of repository"
 535                          << qPrintable(prevrepository) << "branch"
 536                          << qPrintable(prevbranch) << "subdir"
 537                          << qPrintable(prevpath);
 538             } else if (prevrepository != repository) {
 539                 qWarning() << qPrintable(current) << "rev" << revnum
 540                            << "is a cross-repository copy (from repository"
 541                            << qPrintable(prevrepository) << "branch"
 542                            << qPrintable(prevbranch) << "path"
 543                            << qPrintable(prevpath) << "rev" << rev_from << ")";
 544             } else if (prevbranch == branch) {
 545                 // same branch and same repository
 546                 qDebug() << qPrintable(current) << "rev" << revnum
 547                          << "is an SVN rename from"
 548                          << qPrintable(previous) << "rev" << rev_from;
 549                 return EXIT_SUCCESS;
 550             } else {
 551                 // same repository but not same branch
 552                 // this means this is a plain branch
 553                 qDebug() << qPrintable(repository) << ": branch"
 554                          << qPrintable(branch) << "is branching from"
 555                          << qPrintable(prevbranch);
 556
 557                 Repository *repo = repositories.value(repository, 0);
 558                 if (!repo) {
 559                     qCritical() << "Rule" << rule.rx.pattern() << "line" << rule.lineNumber
 560                                 << "references unknown repository" << repository;
 561                     return EXIT_FAILURE;
 562                 }
 563
 564                 repo->createBranch(branch, revnum, prevbranch, rev_from);
 565             }
 566         }
 567     }
 568
 569     Repository::Transaction *txn = transactions.value(repository, 0);
 570     if (!txn) {
 571         Repository *repo = repositories.value(repository, 0);
 572         if (!repo) {
 573             qCritical() << "Rule" << rule.rx.pattern() << "line" << rule.lineNumber
 574                         << "references unknown repository" << repository;
 575             return EXIT_FAILURE;
 576         }
 577
 578         txn = repo->newTransaction(branch, svnprefix, revnum);
 579         if (!txn)
 580             return EXIT_FAILURE;
 581
 582         transactions.insert(repository, txn);
 583     }
 584
 585     if (change->change_kind == svn_fs_path_change_delete) {
 586         txn->deleteFile(path);
 587     } else if (!current.endsWith('/')) {
 588         dumpBlob(txn, fs_root, key, path, pool);
 589     } else {
 590         txn->deleteFile(path);
 591         recursiveDumpDir(txn, fs_root, key, path, pool);
 592     }
 593
 594     return EXIT_SUCCESS;
 595 }
 596
 597 int SvnRevision::recurse(const char *path, const svn_fs_path_change_t *change,
 598                          const char *path_from, svn_revnum_t rev_from,
 599                          apr_pool_t *pool)
 600 {
 601     // get the dir listing
 602     apr_hash_t *entries;
 603     SVN_ERR(svn_fs_dir_entries(&entries, fs_root, path, pool));
 604
 605     AprAutoPool dirpool(pool);
 606     for (apr_hash_index_t *i = apr_hash_first(pool, entries); i; i = apr_hash_next(i)) {
 607         dirpool.clear();
 608         const void *vkey;
 609         void *value;
 610         apr_hash_this(i, &vkey, NULL, &value);
 611
 612         svn_fs_dirent_t *dirent = reinterpret_cast<svn_fs_dirent_t *>(value);
 613         QByteArray entry = path + QByteArray("/") + dirent->name;
 614         QByteArray entryFrom;
 615         if (path_from)
 616             entryFrom = path_from + QByteArray("/") + dirent->name;
 617
 618         QString current = QString::fromUtf8(entry);
 619         if (dirent->kind == svn_node_dir)
 620             current += '/';
 621
 622         // find the first rule that matches this pathname
 623         MatchRuleList::ConstIterator match = findMatchRule(matchRules, revnum, current);
 624         if (match != matchRules.constEnd()) {
 625             if (exportInternal(entry, change, entryFrom.isNull() ? 0 : entryFrom.constData(),
 626                                rev_from, current, *match) == EXIT_FAILURE)
 627                 return EXIT_FAILURE;
 628         } else {
 629             qCritical() << current << "did not match any rules; cannot continue";
 630             return EXIT_FAILURE;
 631         }
 632     }
 633 }