~ubuntu-branches/ubuntu/saucy/lxml/saucy-updates

<tt class="docutils literal">lxml.html</tt> would interpret CSS like <tt class="docutils literal"><span class="pre">url("link")</span></tt> incorrectly

247

(treating the quotation marks as part of the link).</li>

248

<li>Failing import on systems that have an <tt class="docutils literal">io</tt> module.</li>

249

</ul>

250

</div>

251

</div>

252

253

254

255

<h2>Bugs fixed</h2>

256

257

<li>Potential memory leak on exception handling. This was due to a

258

problem in Cython, not lxml itself.</li>

259

<li>Failing import on systems that have an <tt class="docutils literal">io</tt> module.</li>

260

</ul>

261

</div>

262

</div>

263

264

265

266

<h2>Features added</h2>

267

268

<li>Allow <tt class="docutils literal">lxml.html.diff.htmldiff</tt> to accept Element objects, not

269

just HTML strings.</li>

270

</ul>

271

</div>

272

273

<h2>Bugs fixed</h2>

274

275

<li>Crash when using an XPath evaluator in multiple threads.</li>

276

<li>Fixed missing whitespace before <tt class="docutils literal"><span class="pre">Link:...</span></tt> in <tt class="docutils literal">lxml.html.diff</tt>.</li>

277

</ul>

278

</div>

279

280

<h2>Other changes</h2>

281

282

<li>Export <tt class="docutils literal">lxml.html.parse</tt>.</li>

283

</ul>

284

</div>

285

</div>

286

287

288

289

<h2>Bugs fixed</h2>

290

291

<li>Crash when using an XPath evaluator in multiple threads.</li>

292

</ul>

293

</div>

294

</div>

295

296

297

298

<h2>Bugs fixed</h2>

299

300

<li>Crash when using an XPath evaluator in multiple threads.</li>

301

</ul>

302

</div>

303

</div>

304

305

<h1>2.2alpha1 (2008-11-23)</h1>

306

307

<h2>Features added</h2>

308

309

<li>Support for XSLT result tree fragments in XPath/XSLT extension

310

functions.</li>

311

<li>QName objects have new properties <tt class="docutils literal">namespace</tt> and <tt class="docutils literal">localname</tt>.</li>

312

<li>New options for exclusive C14N and C14N without comments.</li>

313

<li>Instantiating a custom Element classes creates a new Element.</li>

314

</ul>

315

</div>

316

317

<h2>Bugs fixed</h2>

318

319

<li>XSLT didn't inherit the parse options of the input document.</li>

320

<li>0-bytes could slip through the API when used inside of Unicode

321

strings.</li>

322

<li>With <tt class="docutils literal">lxml.html.clean.autolink</tt>, links with balanced parenthesis,

323

that end in a parenthesis, will be linked in their entirety (typical

324

with Wikipedia links).</li>

325

</ul>

326

</div>

327

328

<h2>Other changes</h2>

329

</div>

330

</div>

331

332

333

334

<h2>Features added</h2>

335

</div>

336

337

<h2>Bugs fixed</h2>

338

339

<li>Ref-count leaks when lxml enters a try-except statement while an

340

outside exception lives in sys.exc_*(). This was due to a problem in

341

Cython, not lxml itself.</li>

342

<li>Parser Unicode decoding errors could get swallowed by other

343

exceptions.</li>

344

<li>Name/import errors in some Python modules.</li>

345

<li>Internal DTD subsets that did not specify a system or public ID were

346

not serialised and did not appear in the docinfo property of

347

ElementTrees.</li>

348

<li>Fix a pre-Py3k warning when parsing from a gzip file in Py2.6.</li>

349

<li>Test suite fixes for libxml2 2.7.</li>

350

<li>Resolver.resolve_string() did not work for non-ASCII byte strings.</li>

351

<li>Resolver.resolve_file() was broken.</li>

352

<li>Overriding the parser encoding didn't work for many encodings.</li>

353

</ul>

354

</div>

355

356

<h2>Other changes</h2>

357

</div>

358

</div>

359

360

361

362

<h2>Bugs fixed</h2>

363

364

<li>Ref-count leaks when lxml enters a try-except statement while an

365

outside exception lives in sys.exc_*(). This was due to a problem in

366

Cython, not lxml itself.</li>

367

</ul>

368

</div>

369

</div>

370

371

372

373

<h2>Features added</h2>

374

375

<li>lxml.etree now tries to find the absolute path name of files when

376

parsing from a file-like object. This helps custom resolvers when

377

resolving relative URLs, as lixbml2 can prepend them with the path

378

of the source document.</li>

379

</ul>

380

</div>

381

382

<h2>Bugs fixed</h2>

383

384

<li>Memory problem when passing documents between threads.</li>

385

<li>Target parser did not honour the <tt class="docutils literal">recover</tt> option and raised an

386

exception instead of calling <tt class="docutils literal">.close()</tt> on the target.</li>

387

</ul>

388

</div>

389

390

<h2>Other changes</h2>

391

</div>

392

</div>

393

394

395

396

<h2>Bugs fixed</h2>

397

398

<li>Memory problem when passing documents between threads.</li>

399

<li>Target parser did not honour the <tt class="docutils literal">recover</tt> option and raised an

400

exception instead of calling <tt class="docutils literal">.close()</tt> on the target.</li>

401

</ul>

402

</div>

403

</div>

404

405

406

407

<h2>Features added</h2>

408

</div>

409

410

<h2>Bugs fixed</h2>

411

412

<li>Crash when parsing XSLT stylesheets in a thread and using them in

413

another.</li>

414

<li>Encoding problem when including text with ElementInclude under

415

Python 3.</li>

416

</ul>

417

</div>

418

419

<h2>Other changes</h2>

420

</div>

421

</div>

422

423

424

425

<h2>Features added</h2>

426

427

<li><tt class="docutils literal">lxml.html.rewrite_links()</tt> strips links to work around documents

428

with whitespace in URL attributes.</li>

429

</ul>

430

</div>

431

432

<h2>Bugs fixed</h2>

433

434

<li>Crash when parsing XSLT stylesheets in a thread and using them in

435

another.</li>

436

<li>CSS selector parser dropped remaining expression after a function

437

with parameters.</li>

438

</ul>

439

</div>

440

441

<h2>Other changes</h2>

442

</div>

443

</div>

444

445

446

447

<h2>Features added</h2>

448

449

<li>Smart strings can be switched off in XPath (<tt class="docutils literal">smart_strings</tt>

450

keyword option).</li>

451

<li><tt class="docutils literal">lxml.html.rewrite_links()</tt> strips links to work around documents

452

with whitespace in URL attributes.</li>

453

</ul>

454

</div>

455

456

<h2>Bugs fixed</h2>

457

458

<li>Custom resolvers were not used for XMLSchema includes/imports and

459

XInclude processing.</li>

460

<li>CSS selector parser dropped remaining expression after a function

461

with parameters.</li>

462

</ul>

463

</div>

464

465

<h2>Other changes</h2>

466

467

<li><tt class="docutils literal">objectify.enableRecursiveStr()</tt> was removed, use

468

<tt class="docutils literal">objectify.enable_recursive_str()</tt> instead</li>

469

<li>Speed-up when running XSLTs on documents from other threads</li>

470

</ul>

471

</div>

472

</div>

473

474

475

476

<h2>Features added</h2>

477

478

<li>Pickling <tt class="docutils literal">ElementTree</tt> objects in lxml.objectify.</li>

479

</ul>

480

</div>

481

482

<h2>Bugs fixed</h2>

483

484

<li>Descending dot-separated classes in CSS selectors were not resolved

485

correctly.</li>

486

<li><tt class="docutils literal">ElementTree.parse()</tt> didn't handle target parser result.</li>

487

<li>Potential threading problem in XInclude.</li>

488

<li>Crash in Element class lookup classes when the __init__() method of

489

the super class is not called from Python subclasses.</li>

490

</ul>

491

</div>

492

493

<h2>Other changes</h2>

494

495

<li>Non-ASCII characters in attribute values are no longer escaped on

496

serialisation.</li>

497

</ul>

498

</div>

499

</div>

500

501

502

503

<h2>Features added</h2>

504

505

<li>Major overhaul of <tt class="docutils literal">tools/xpathgrep.py</tt> script.</li>

506

<li>Pickling <tt class="docutils literal">ElementTree</tt> objects in lxml.objectify.</li>

507

<li>Support for parsing from file-like objects that return unicode

508

strings.</li>

509

<li>New function <tt class="docutils literal">etree.cleanup_namespaces(el)</tt> that removes unused

510

namespace declarations from a (sub)tree (experimental).</li>

511

<li>XSLT results support the buffer protocol in Python 3.</li>

512

<li>Polymorphic functions in <tt class="docutils literal">lxml.html</tt> that accept either a tree or

513

a parsable string will return either a UTF-8 encoded byte string, a

514

unicode string or a tree, based on the type of the input.

515

Previously, the result was always a byte string or a tree.</li>

516

<li>Support for Python 2.6 and 3.0 beta.</li>

517

<li>File name handling now uses a heuristic to convert between byte

518

strings (usually filenames) and unicode strings (usually URLs).</li>

519

<li>Parsing from a plain file object frees the GIL under Python 2.x.</li>

520

<li>Running <tt class="docutils literal">iterparse()</tt> on a plain file (or filename) frees the GIL

521

on reading under Python 2.x.</li>

522

<li>Conversion functions <tt class="docutils literal">html_to_xhtml()</tt> and <tt class="docutils literal">xhtml_to_html()</tt> in

523

lxml.html (experimental).</li>

524

<li>Most features in lxml.html work for XHTML namespaced tag names

525

(experimental).</li>

526

</ul>

527

</div>

528

529

<h2>Bugs fixed</h2>

530

531

<li><tt class="docutils literal">ElementTree.parse()</tt> didn't handle target parser result.</li>

532

<li>Crash in Element class lookup classes when the __init__() method of

533

the super class is not called from Python subclasses.</li>

534

<li>A number of problems related to unicode/byte string conversion of

535

filenames and error messages were fixed.</li>

536

<li>Building on MacOS-X now passes the "flat_namespace" option to the C

537

compiler, which reportedly prevents build quirks and crashes on this

538

platform.</li>

539

<li>Windows build was broken.</li>

540

<li>Rare crash when serialising to a file object with certain encodings.</li>

541

</ul>

542

</div>

543

544

<h2>Other changes</h2>

545

546

<li>Non-ASCII characters in attribute values are no longer escaped on

547

serialisation.</li>

548

<li>Passing non-ASCII byte strings or invalid unicode strings as .tag,

549

namespaces, etc. will result in a ValueError instead of an

550

AssertionError (just like the tag well-formedness check).</li>

551

<li>Up to several times faster attribute access (i.e. tree traversal) in

552

lxml.objectify.</li>

553

</ul>

554

</div>

555

</div>

556

557

558

559

<h2>Features added</h2>

560

</div>

561

562

<h2>Bugs fixed</h2>

563

564

<li>Incorrect evaluation of <tt class="docutils literal"><span class="pre">el.find("tag[child]")</span></tt>.</li>

565

<li>Windows build was broken.</li>

566

<li>Moving a subtree from a document created in one thread into a

567

document of another thread could crash when the rest of the source

568

document is deleted while the subtree is still in use.</li>

569

<li>Rare crash when serialising to a file object with certain encodings.</li>

570

</ul>

571

</div>

572

573

<h2>Other changes</h2>

574

575

<li>lxml should now build without problems on MacOS-X.</li>

576

</ul>

577

</div>

578

</div>

579

580

581

582

<h2>Features added</h2>

583

584

<li>All parse functions in lxml.html take a <tt class="docutils literal">parser</tt> keyword argument.</li>

585

<li>lxml.html has a new parser class <tt class="docutils literal">XHTMLParser</tt> and a module

586

attribute <tt class="docutils literal">xhtml_parser</tt> that provide XML parsers that are

587

pre-configured for the lxml.html package.</li>

588

</ul>

589

</div>

590

591

<h2>Bugs fixed</h2>

592

593

<li>Moving a subtree from a document created in one thread into a

594

document of another thread could crash when the rest of the source

595

document is deleted while the subtree is still in use.</li>

596

<li>Passing an nsmap when creating an Element will no longer strip

597

redundantly defined namespace URIs. This prevented the definition

598

of more than one prefix for a namespace on the same Element.</li>

599

</ul>

600

</div>

601

602

<h2>Other changes</h2>

603

604

<li>If the default namespace is redundantly defined with a prefix on the

605

same Element, the prefix will now be preferred for subelements and

606

attributes. This allows users to work around a problem in libxml2

607

where attributes from the default namespace could serialise without

608

a prefix even when they appear on an Element with a different

609

namespace (i.e. they would end up in the wrong namespace).</li>

610

</ul>

611

</div>

612

</div>

613

614

615

616

<h2>Features added</h2>

617

</div>

618

619

<h2>Bugs fixed</h2>

620

621

<li>Resolving to a filename in custom resolvers didn't work.</li>

622

<li>lxml did not honour libxslt's second error state "STOPPED", which

623

let some XSLT errors pass silently.</li>

624

<li>Memory leak in Schematron with libxml2 >= 2.6.31.</li>

625

</ul>

626

</div>

627

628

<h2>Other changes</h2>

629

</div>

630

</div>

631

632

633

634

<h2>Features added</h2>

635

636

<li>Error logging in Schematron (requires libxml2 2.6.32 or later).</li>

637

<li>Parser option <tt class="docutils literal">strip_cdata</tt> for normalising or keeping CDATA

638

sections. Defaults to <tt class="docutils literal">True</tt> as before, thus replacing CDATA

639

sections by their text content.</li>

640

<li><tt class="docutils literal">CDATA()</tt> factory to wrap string content as CDATA section.</li>

641

</ul>

642

</div>

643

644

<h2>Bugs fixed</h2>

645

646

<li>Resolving to a filename in custom resolvers didn't work.</li>

647

<li>lxml did not honour libxslt's second error state "STOPPED", which

648

let some XSLT errors pass silently.</li>

649

<li>Memory leak in Schematron with libxml2 >= 2.6.31.</li>

650

<li>lxml.etree accepted non well-formed namespace prefix names.</li>

651

</ul>

652

</div>

653

654

<h2>Other changes</h2>

655

656

<li>Major cleanup in internal <tt class="docutils literal">moveNodeToDocument()</tt> function, which

657

takes care of namespace cleanup when moving elements between

658

different namespace contexts.</li>

659

<li>New Elements created through the <tt class="docutils literal">makeelement()</tt> method of an HTML

660

parser or through lxml.html now end up in a new HTML document

661

(doctype HTML 4.01 Transitional) instead of a generic XML document.

662

This mostly impacts the serialisation and the availability of a DTD

663

context.</li>

664

</ul>

665

</div>

666

</div>

667

668

669

670

<h2>Features added</h2>

671

</div>

672

673

<h2>Bugs fixed</h2>

674

675

<li>Hanging thread in conjunction with GTK threading.</li>

676

<li>Crash bug in iterparse when moving elements into other documents.</li>

677

<li>HTML elements' <tt class="docutils literal">.cssselect()</tt> method was broken.</li>

678

<li><tt class="docutils literal"><span class="pre">ElementTree.find*()</span></tt> didn't accept QName objects.</li>

679

</ul>

680

</div>

681

682

<h2>Other changes</h2>

683

</div>

684

</div>

685

686

<h1>2.1alpha1 (2008-03-27)</h1>

687

688

<h2>Features added</h2>

689

690

<li>New event types 'comment' and 'pi' in <tt class="docutils literal">iterparse()</tt>.</li>

691

<li><tt class="docutils literal">XSLTAccessControl</tt> instances have a property <tt class="docutils literal">options</tt> that

692

returns a dict of access configuration options.</li>

693

<li>Constant instances <tt class="docutils literal">DENY_ALL</tt> and <tt class="docutils literal">DENY_WRITE</tt> on

694

<tt class="docutils literal">XSLTAccessControl</tt> class.</li>

695

<li>Extension elements for XSLT (experimental!)</li>

696

<li><tt class="docutils literal">Element.base</tt> property returns the xml:base or HTML base URL of

697

an Element.</li>

698

<li><tt class="docutils literal">docinfo.URL</tt> property is writable.</li>

699

</ul>

700

</div>

701

702

<h2>Bugs fixed</h2>

703

704

<li>Default encoding for plain text serialisation was different from

705

that of XML serialisation (UTF-8 instead of ASCII).</li>

706

</ul>

707

</div>

708

709

<h2>Other changes</h2>

710

711

<li>Minor API speed-ups.</li>

712

<li>The benchmark suite now uses tail text in the trees, which makes the

713

absolute numbers incomparable to previous results.</li>

714

<li>Generating the HTML documentation now requires <a class="reference external" href="http://pygments.org/">Pygments</a>, which is

715

used to enable syntax highlighting for the doctest examples.</li>

716

</ul>

717

<p>Most long-time deprecated functions and methods were removed:</p>

718

<ul>

719

<li><p class="first"><tt class="docutils literal">etree.clearErrorLog()</tt>, use <tt class="docutils literal">etree.clear_error_log()</tt></p>

720

</li>

721

<li><p class="first"><tt class="docutils literal">etree.useGlobalPythonLog()</tt>, use

722

<tt class="docutils literal">etree.use_global_python_log()</tt></p>

723

</li>

724

<li><p class="first"><tt class="docutils literal">etree.ElementClassLookup.setFallback()</tt>, use

725

<tt class="docutils literal">etree.ElementClassLookup.set_fallback()</tt></p>

726

</li>

727

<li><p class="first"><tt class="docutils literal">etree.getDefaultParser()</tt>, use <tt class="docutils literal">etree.get_default_parser()</tt></p>

728

</li>

729

<li><p class="first"><tt class="docutils literal">etree.setDefaultParser()</tt>, use <tt class="docutils literal">etree.set_default_parser()</tt></p>

730

</li>

731

<li><p class="first"><tt class="docutils literal">etree.setElementClassLookup()</tt>, use

732

<tt class="docutils literal">etree.set_element_class_lookup()</tt></p>

733

<p>Note that <tt class="docutils literal">parser.setElementClassLookup()</tt> has not been removed

734

yet, although <tt class="docutils literal">parser.set_element_class_lookup()</tt> should be used

735

instead.</p>

736

</li>

737

<li><p class="first"><tt class="docutils literal">xpath_evaluator.registerNamespace()</tt>, use

738

<tt class="docutils literal">xpath_evaluator.register_namespace()</tt></p>

739

</li>

740

<li><p class="first"><tt class="docutils literal">xpath_evaluator.registerNamespaces()</tt>, use

741

<tt class="docutils literal">xpath_evaluator.register_namespaces()</tt></p>

742

</li>

743

<li><p class="first"><tt class="docutils literal">objectify.setPytypeAttributeTag</tt>, use

744

<tt class="docutils literal">objectify.set_pytype_attribute_tag</tt></p>

745

</li>

746

<li><p class="first"><tt class="docutils literal">objectify.setDefaultParser()</tt>, use

747

<tt class="docutils literal">objectify.set_default_parser()</tt></p>

748

</li>

749

</ul>

750

</div>

751

</div>

752

753

754

755

<h2>Features added</h2>

756

757

<li>soupparser.parse() allows passing keyword arguments on to

758

BeautifulSoup.</li>

759

<li><tt class="docutils literal">fromstring()</tt> method in <tt class="docutils literal">lxml.html.soupparser</tt>.</li>

760

</ul>

761

</div>

762

763

<h2>Bugs fixed</h2>

764

765

<li><tt class="docutils literal">lxml.html.diff</tt> didn't treat empty tags properly (e.g.,

766

767

<li>Handle entity replacements correctly in target parser.</li>

768

<li>Crash when using <tt class="docutils literal">iterparse()</tt> with XML Schema validation.</li>

769

<li>The BeautifulSoup parser (soupparser.py) did not replace entities,

770

which made them turn up in text content.</li>

771

<li>Attribute assignment of custom PyTypes in objectify could fail to

772

correctly serialise the value to a string.</li>

773

</ul>

774

</div>

775

776

<h2>Other changes</h2>

777

778

<li><tt class="docutils literal">lxml.html.ElementSoup</tt> was replaced by a new module

779

<tt class="docutils literal">lxml.html.soupparser</tt> with a more consistent API. The old module

780

remains for compatibility with ElementTree's own ElementSoup module.</li>

781

<li>Setting the XSLT_CONFIG and XML2_CONFIG environment variables at

782

build time will let setup.py pick up the <tt class="docutils literal"><span class="pre">xml2-config</span></tt> and

783

<tt class="docutils literal"><span class="pre">xslt-config</span></tt> scripts from the supplied path name.</li>

784

<li>Passing <tt class="docutils literal"><span class="pre">--with-xml2-config=/path/to/xml2-config</span></tt> to setup.py will

785

override the <tt class="docutils literal"><span class="pre">xml2-config</span></tt> script that is used to determine the C

786

compiler options. The same applies for the <tt class="docutils literal"><span class="pre">--with-xslt-config</span></tt>

787

option.</li>

788

</ul>

789

</div>

790

</div>

791

792

793

794

<h2>Features added</h2>

795

796

<li>Support passing <tt class="docutils literal">base_url</tt> to file parser functions to override

797

the filename of the file(-like) object.</li>

798

</ul>

799

</div>

800

801

<h2>Bugs fixed</h2>

802

803

<li>The prefix for objectify's pytype namespace was missing from the set

804

of default prefixes.</li>

805

<li>Memory leak in Schematron (fixed only for libxml2 2.6.31+).</li>

806

<li>Error type names in RelaxNG were reported incorrectly.</li>

807

<li>Slice deletion bug fixed in objectify.</li>

808

</ul>

809

</div>

810

811

<h2>Other changes</h2>

812

813

<li>Enabled doctests for some Python modules (especially <tt class="docutils literal">lxml.html</tt>).</li>

814

<li>Add a <tt class="docutils literal">method</tt> argument to <tt class="docutils literal">lxml.html.tostring()</tt>

815

(<tt class="docutils literal"><span class="pre">method="xml"</span></tt> for XHTML output).</li>

816

<li>Make it clearer that methods like <tt class="docutils literal">lxml.html.fromstring()</tt> take a

817

<tt class="docutils literal">base_url</tt> argument.</li>

818

</ul>

819

</div>

820

</div>

821

822

823

824

<h2>Features added</h2>

825

826

<li>Child iteration in <tt class="docutils literal">lxml.pyclasslookup</tt>.</li>

827

<li>Loads of new docstrings reflect the signature of functions and

828

methods to make them visible in API docs and <tt class="docutils literal">help()</tt></li>

829

</ul>

830

</div>

831

832

<h2>Bugs fixed</h2>

833

834

<li>The module <tt class="docutils literal">lxml.html.builder</tt> was duplicated as

835

<tt class="docutils literal">lxml.htmlbuilder</tt></li>

836

<li>Form elements would return None for <tt class="docutils literal">form.fields.keys()</tt> if there

837

was an unnamed input field. Now unnamed input fields are completely

838

ignored.</li>

839

<li>Setting an element slice in objectify could insert slice-overlapping

840

elements at the wrong position.</li>

841

</ul>

842

</div>

843

844

<h2>Other changes</h2>

845

846

<li>The generated API documentation was cleaned up and disburdened from

847

non-public classes etc.</li>

848

<li>The previously public module <tt class="docutils literal">lxml.html.setmixin</tt> was renamed to

849

<tt class="docutils literal">lxml.html._setmixin</tt> as it is not an official part of lxml. If

850

you want to use it, feel free to copy it over to your own source

851

base.</li>

852

<li>Passing <tt class="docutils literal"><span class="pre">--with-xslt-config=/path/to/xslt-config</span></tt> to setup.py will

853

override the <tt class="docutils literal"><span class="pre">xslt-config</span></tt> script that is used to determine the C

854

compiler options.</li>

855

</ul>

856

</div>

857

</div>

858

859

860

861

<h2>Features added</h2>

862

863

<li>Passing the <tt class="docutils literal">unicode</tt> type as <tt class="docutils literal">encoding</tt> to <tt class="docutils literal">tostring()</tt> will

864

serialise to unicode. The <tt class="docutils literal">tounicode()</tt> function is now

865

deprecated.</li>

866

<li><tt class="docutils literal">XMLSchema()</tt> and <tt class="docutils literal">RelaxNG()</tt> can parse from StringIO.</li>

867

<li><tt class="docutils literal">makeparser()</tt> function in <tt class="docutils literal">lxml.objectify</tt> to create a new

868

parser with the usual objectify setup.</li>

869

<li>Plain ASCII XPath string results are no longer forced into unicode

870

objects as in 2.0beta1, but are returned as plain strings as before.</li>

871

<li>All XPath string results are 'smart' objects that have a

872

<tt class="docutils literal">getparent()</tt> method to retrieve their parent Element.</li>

873

<li><tt class="docutils literal">with_tail</tt> option in serialiser functions.</li>

874

<li>More accurate exception messages in validator creation.</li>

875

<li>Parse-time XML schema validation (<tt class="docutils literal">schema</tt> parser keyword).</li>

876

<li>XPath string results of the <tt class="docutils literal">text()</tt> function and attribute

877

selection make their Element container accessible through a

878

<tt class="docutils literal">getparent()</tt> method. As a side-effect, they are now always

879

unicode objects (even ASCII strings).</li>

880

<li><tt class="docutils literal">XSLT</tt> objects are usable in any thread - at the cost of a deep

881

copy if they were not created in that thread.</li>

882

<li>Invalid entity names and character references will be rejected by

883

the <tt class="docutils literal">Entity()</tt> factory.</li>

884

<li><tt class="docutils literal">entity.text</tt> returns the textual representation of the entity,

885

e.g. <tt class="docutils literal">&amp;</tt>.</li>

886

<li>New properties <tt class="docutils literal">position</tt> and <tt class="docutils literal">code</tt> on ParseError exception (as

887

in ET 1.3)</li>

888

<li>Rich comparison of <tt class="docutils literal">element.attrib</tt> proxies.</li>

889

<li>ElementTree compatible TreeBuilder class.</li>

890

<li>Use default prefixes for some common XML namespaces.</li>

891

<li><tt class="docutils literal">lxml.html.clean.Cleaner</tt> now allows for a <tt class="docutils literal">host_whitelist</tt>, and

892

two overridable methods: <tt class="docutils literal">allow_embedded_url(el, url)</tt> and the

893

more general <tt class="docutils literal">allow_element(el)</tt>.</li>

894

<li>Extended slicing of Elements as in <tt class="docutils literal"><span class="pre">element[1:-1:2]</span></tt>, both in

895

etree and in objectify</li>

896

<li>Resolvers can now provide a <tt class="docutils literal">base_url</tt> keyword argument when

897

resolving a document as string data.</li>

898

<li>When using <tt class="docutils literal">lxml.doctestcompare</tt> you can give the doctest option

899

<tt class="docutils literal">NOPARSE_MARKUP</tt> (like <tt class="docutils literal"># doctest: +NOPARSE_MARKUP</tt>) to suppress

900

the special checking for one test.</li>

901

<li>Separate <tt class="docutils literal">feed_error_log</tt> property for the feed parser interface.

902

The normal parser interface and <tt class="docutils literal">iterparse</tt> continue to use

903

<tt class="docutils literal">error_log</tt>.</li>

904

<li>The normal parsers and the feed parser interface are now separated

905

and can be used concurrently on the same parser instance.</li>

906

<li><tt class="docutils literal">fromstringlist()</tt> and <tt class="docutils literal">tostringlist()</tt> functions as in

907

ElementTree 1.3</li>

908

<li><tt class="docutils literal">iterparse()</tt> accepts an <tt class="docutils literal">html</tt> boolean keyword argument for

909

parsing with the HTML parser (note that this interface may be

910

subject to change)</li>

911

<li>Parsers accept an <tt class="docutils literal">encoding</tt> keyword argument that overrides the encoding

912

of the parsed documents.</li>

913

<li>New C-API function <tt class="docutils literal">hasChild()</tt> to test for children</li>

914

<li><tt class="docutils literal">annotate()</tt> function in objectify can annotate with Python types and XSI

915

types in one step. Accompanied by <tt class="docutils literal">xsiannotate()</tt> and <tt class="docutils literal">pyannotate()</tt>.</li>

916

<li><tt class="docutils literal">ET.write()</tt>, <tt class="docutils literal">tostring()</tt> and <tt class="docutils literal">tounicode()</tt> now accept a keyword

917

argument <tt class="docutils literal">method</tt> that can be one of 'xml' (or None), 'html' or 'text' to

918

serialise as XML, HTML or plain text content.</li>

919

<li><tt class="docutils literal">iterfind()</tt> method on Elements returns an iterator equivalent to

920

<tt class="docutils literal">findall()</tt></li>

921

<li><tt class="docutils literal">itertext()</tt> method on Elements</li>

922

<li>Setting a QName object as value of the .text property or as an attribute

923

will resolve its prefix in the respective context</li>

924

<li>ElementTree-like parser target interface as described in

925

<a class="reference external" href="http://effbot.org/elementtree/elementtree-xmlparser.htm">http://effbot.org/elementtree/elementtree-xmlparser.htm</a></li>

926

<li>ElementTree-like feed parser interface on XMLParser and HTMLParser

927

(<tt class="docutils literal">feed()</tt> and <tt class="docutils literal">close()</tt> methods)</li>

928

<li>Reimplemented <tt class="docutils literal">objectify.E</tt> for better performance and improved

929

integration with objectify. Provides extended type support based on

930

registered PyTypes.</li>

931

<li>XSLT objects now support deep copying</li>

932

<li>New <tt class="docutils literal">makeSubElement()</tt> C-API function that allows creating a new

933

subelement straight with text, tail and attributes.</li>

934

<li>XPath extension functions can now access the current context node

935

(<tt class="docutils literal">context.context_node</tt>) and use a context dictionary

936

(<tt class="docutils literal">context.eval_context</tt>) from the context provided in their first

937

parameter</li>

938

<li>HTML tag soup parser based on BeautifulSoup in <tt class="docutils literal">lxml.html.ElementSoup</tt></li>

939

<li>New module <tt class="docutils literal">lxml.doctestcompare</tt> by Ian Bicking for writing simplified

940

doctests based on XML/HTML output. Use by importing <tt class="docutils literal">lxml.usedoctest</tt> or

941

<tt class="docutils literal">lxml.html.usedoctest</tt> from within a doctest.</li>

942

<li>New module <tt class="docutils literal">lxml.cssselect</tt> by Ian Bicking for selecting Elements with CSS

943

selectors.</li>

944

<li>New package <tt class="docutils literal">lxml.html</tt> written by Ian Bicking for advanced HTML

945

treatment.</li>

946

<li>Namespace class setup is now local to the <tt class="docutils literal">ElementNamespaceClassLookup</tt>

947

instance and no longer global.</li>

948

<li>Schematron validation (incomplete in libxml2)</li>

949

<li>Additional <tt class="docutils literal">stringify</tt> argument to <tt class="docutils literal">objectify.PyType()</tt> takes a

950

conversion function to strings to support setting text values from arbitrary

951

types.</li>

952

<li>Entity support through an <tt class="docutils literal">Entity</tt> factory and element classes. XML

953

parsers now have a <tt class="docutils literal">resolve_entities</tt> keyword argument that can be set to

954

False to keep entities in the document.</li>

955

<li><tt class="docutils literal">column</tt> field on error log entries to accompany the <tt class="docutils literal">line</tt> field</li>

956

<li>Error specific messages in XPath parsing and evaluation

957

NOTE: for evaluation errors, you will now get an XPathEvalError instead of

958

an XPathSyntaxError. To catch both, you can except on <tt class="docutils literal">XPathError</tt></li>

959

<li>The regular expression functions in XPath now support passing a node-set

960

instead of a string</li>

961

<li>Extended type annotation in objectify: new <tt class="docutils literal">xsiannotate()</tt> function</li>

962

<li>EXSLT RegExp support in standard XPath (not only XSLT)</li>

963

</ul>

964

</div>

965

966

<h2>Bugs fixed</h2>

967

968

<li>Missing import in <tt class="docutils literal">lxml.html.clean</tt>.</li>

969

<li>Some Python 2.4-isms prevented lxml from building/running under

970

Python 2.3.</li>

971

<li>XPath on ElementTrees could crash when selecting the virtual root

972

node of the ElementTree.</li>

973

<li>Compilation <tt class="docutils literal"><span class="pre">--without-threading</span></tt> was buggy in alpha5/6.</li>

974

<li>Memory leak in the <tt class="docutils literal">parse()</tt> function.</li>

975

<li>Minor bugs in XSLT error message formatting.</li>

976

<li>Result document memory leak in target parser.</li>

977

<li>Target parser failed to report comments.</li>

978

<li>In the <tt class="docutils literal">lxml.html</tt> <tt class="docutils literal">iter_links</tt> method, links in <tt class="docutils literal"><object></tt>

979

tags weren't recognized. (Note: plugin-specific link parameters

980

still aren't recognized.) Also, the <tt class="docutils literal"><embed></tt> tag, though not

981

standard, is now included in <tt class="docutils literal">lxml.html.defs.special_inline_tags</tt>.</li>

982

<li>Using custom resolvers on XSLT stylesheets parsed from a string

983

could request ill-formed URLs.</li>

984

<li>With <tt class="docutils literal">lxml.doctestcompare</tt> if you do <tt class="docutils literal"><tag <span class="pre">xmlns="..."></span></tt> in your

985

output, it will then be namespace-neutral (before the ellipsis was

986

treated as a real namespace).</li>

987

<li>AttributeError in feed parser on parse errors</li>

988

<li>XML feed parser setup problem</li>

989

<li>Type annotation for unicode strings in <tt class="docutils literal">DataElement()</tt></li>

990

<li>lxml failed to serialise namespace declarations of elements other than the

991

root node of a tree</li>

992

<li>Race condition in XSLT where the resolver context leaked between concurrent

993

XSLT calls</li>

994

<li>lxml.etree did not check tag/attribute names</li>

995

<li>The XML parser did not report undefined entities as error</li>

996

<li>The text in exceptions raised by XML parsers, validators and XPath

997

evaluators now reports the first error that occurred instead of the last</li>

998

<li>Passing '' as XPath namespace prefix did not raise an error</li>

999

<li>Thread safety in XPath evaluators</li>

1000

</ul>

1001

</div>

1002

1003

<h2>Other changes</h2>

1004

1005

<li>Exceptions carry only the part of the error log that is related to

1006

the operation that caused the error.</li>

1007

<li><tt class="docutils literal">XMLSchema()</tt> and <tt class="docutils literal">RelaxNG()</tt> now enforce passing the source

1008

file/filename through the <tt class="docutils literal">file</tt> keyword argument.</li>

1009

<li>The test suite now skips most doctests under Python 2.3.</li>

1010

<li><tt class="docutils literal">make clean</tt> no longer removes the .c files (use <tt class="docutils literal">make

1011

realclean</tt> instead)</li>

1012

<li>Minor performance tweaks for Element instantiation and subelement

1013

creation</li>

1014

<li>Various places in the XPath, XSLT and iteration APIs now require

1015

keyword-only arguments.</li>

1016

<li>The argument order in <tt class="docutils literal">element.itersiblings()</tt> was changed to

1017

match the order used in all other iteration methods. The second

1018

argument ('preceding') is now a keyword-only argument.</li>

1019

<li>The <tt class="docutils literal">getiterator()</tt> method on Elements and ElementTrees was

1020

reverted to return an iterator as it did in lxml 1.x. The ET API

1021

specification allows it to return either a sequence or an iterator,

1022

and it traditionally returned a sequence in ET and an iterator in

1023

lxml. However, it is now deprecated in favour of the <tt class="docutils literal">iter()</tt>

1024

method, which should be used in new code wherever possible.</li>

1025

<li>The 'pretty printed' serialisation of ElementTree objects now

1026

inserts newlines at the root level between processing instructions,

1027

comments and the root tag.</li>

1028

<li>A 'pretty printed' serialisation is now terminated with a newline.</li>

1029

<li>Second argument to <tt class="docutils literal">lxml.etree.Extension()</tt> helper is no longer

1030

required, third argument is now a keyword-only argument <tt class="docutils literal">ns</tt>.</li>

1031

<li><tt class="docutils literal">lxml.html.tostring</tt> takes an <tt class="docutils literal">encoding</tt> argument.</li>

1032

<li>The module source files were renamed to "lxml.*.pyx", such as

1033

"lxml.etree.pyx". This was changed for consistency with the way

1034

Pyrex commonly handles package imports. The main effect is that

1035

classes now know about their fully qualified class name, including

1036

the package name of their module.</li>

1037

<li>Keyword-only arguments in some API functions, especially in the

1038

parsers and serialisers.</li>

1039

<li>Tag name validation in lxml.etree (and lxml.html) now distinguishes

1040

between HTML tags and XML tags based on the parser that was used to

1041

parse or create them. HTML tags no longer reject any non-ASCII

1042

characters in tag names but only spaces and the special characters

1043

1044

<li>lxml.etree now emits a warning if you use XPath with libxml2 2.6.27

1045

(which can crash on certain XPath errors)</li>

1046

<li>Type annotation in objectify now preserves the already annotated type by

1047

default to prevent loosing type information that is already there.</li>

1048

<li><tt class="docutils literal">element.getiterator()</tt> returns a list, use <tt class="docutils literal">element.iter()</tt> to retrieve

1049

an iterator (ElementTree 1.3 compatible behaviour)</li>

1050

<li>objectify.PyType for None is now called "NoneType"</li>

1051

<li><tt class="docutils literal">el.getiterator()</tt> renamed to <tt class="docutils literal">el.iter()</tt>, following ElementTree 1.3 -

1052

original name is still available as alias</li>

1053

<li>In the public C-API, <tt class="docutils literal">findOrBuildNodeNs()</tt> was replaced by the more

1054

generic <tt class="docutils literal">findOrBuildNodeNsPrefix</tt></li>

1055

<li>Major refactoring in XPath/XSLT extension function code</li>

1056

<li>Network access in parsers disabled by default</li>

1057

</ul>

1058

</div>

1059

</div>

1060

1061

1062

1063

<h2>Bugs fixed</h2>

1064

1065

<li>Backported decref crash fix from 2.0</li>

1066

<li>Well hidden free-while-in-use crash bug in ObjectPath</li>

1067

</ul>

1068

</div>

1069

1070

<h2>Other changes</h2>

1071

1072

<li>The test suites now run <tt class="docutils literal">gc.collect()</tt> in the <tt class="docutils literal">tearDown()</tt>

1073

methods. While this makes them take a lot longer to run, it also

1074

makes it easier to link a specific test to garbage collection

1075

problems that would otherwise appear in later tests.</li>

1076

</ul>

1077

</div>

1078

</div>

1079

1080

1081

1082

<h2>Features added</h2>

1083

</div>

1084

1085

<h2>Bugs fixed</h2>

1086

1087

<li>lxml.etree could crash when adding more than 10000 namespaces to a

1088

document</li>

1089

<li>lxml failed to serialise namespace declarations of elements other

1090

than the root node of a tree</li>

1091

</ul>

1092

</div>

1093

</div>

1094

1095

1096

1097

<h2>Features added</h2>

1098

1099

<li>The <tt class="docutils literal">ElementMaker</tt> in <tt class="docutils literal">lxml.builder</tt> now accepts the keyword arguments

1100

<tt class="docutils literal">namespace</tt> and <tt class="docutils literal">nsmap</tt> to set a namespace and nsmap for the Elements it

1101

creates.</li>

1102

<li>The <tt class="docutils literal">docinfo</tt> on ElementTree objects has new properties <tt class="docutils literal">internalDTD</tt>

1103

and <tt class="docutils literal">externalDTD</tt> that return a DTD object for the internal or external

1104

subset of the document respectively.</li>

1105

<li>Serialising an ElementTree now includes any internal DTD subsets that are

1106

part of the document, as well as comments and PIs that are siblings of the

1107

root node.</li>

1108

</ul>

1109

</div>

1110

1111

<h2>Bugs fixed</h2>

1112

1113

<li>Parsing with the <tt class="docutils literal">no_network</tt> option could fail</li>

1114

</ul>

1115

</div>

1116

1117

<h2>Other changes</h2>

1118

1119

<li>lxml now raises a TagNameWarning about tag names containing ':' instead of

1120

an Error as 1.3.3 did. The reason is that a number of projects currently

1121

misuse the previous lack of tag name validation to generate namespace

1122

prefixes without declaring namespaces. Apart from the danger of generating

1123

broken XML this way, it also breaks most of the namespace-aware tools in

1124

XML, including XPath, XSLT and validation. lxml 1.3.x will continue to

1125

support this bug with a Warning, while lxml 2.0 will be strict about

1126

well-formed tag names (not only regarding ':').</li>

1127

<li>Serialising an Element no longer includes its comment and PI siblings (only

1128

ElementTree serialisation includes them).</li>

1129

</ul>

1130

</div>

1131

</div>

1132

1133

1134

1135

<h2>Features added</h2>

1136

1137

<li>ElementTree compatible parser <tt class="docutils literal">ETCompatXMLParser</tt> strips processing

1138

instructions and comments while parsing XML</li>

1139

<li>Parsers now support stripping PIs (keyword argument 'remove_pis')</li>

1140

<li><tt class="docutils literal">etree.fromstring()</tt> now supports parsing both HTML and XML, depending on

1141

the parser you pass.</li>

1142

<li>Support <tt class="docutils literal">base_url</tt> keyword argument in <tt class="docutils literal">HTML()</tt> and <tt class="docutils literal">XML()</tt></li>

1143

</ul>

1144

</div>

1145

1146

<h2>Bugs fixed</h2>

1147

1148

<li>Parsing from Python Unicode strings failed on some platforms</li>

1149

<li><tt class="docutils literal">Element()</tt> did not raise an exception on tag names containing ':'</li>

1150

<li><tt class="docutils literal">Element.getiterator(tag)</tt> did not accept <tt class="docutils literal">Comment</tt> and

1151

<tt class="docutils literal">ProcessingInstruction</tt> as tags. It also accepts <tt class="docutils literal">Element</tt> now.</li>

1152

</ul>

1153

</div>

1154

</div>

1155

1156

1157

1158

<h2>Features added</h2>

1159

</div>

1160

1161

<h2>Bugs fixed</h2>

1162

1163

<li>"deallocating None" crash bug</li>

1164

</ul>

1165

</div>

1166

</div>

1167

1168

1169

1170

<h2>Features added</h2>

1171

1172

<li>objectify.DataElement now supports setting values from existing data

1173

elements (not just plain Python types) and reuses defined namespaces etc.</li>

1174

<li>E-factory support for lxml.objectify (<tt class="docutils literal">objectify.E</tt>)</li>

1175

</ul>

1176

</div>

1177

1178

<h2>Bugs fixed</h2>

1179

1180

<li>Better way to prevent crashes in Element proxy cleanup code</li>

1181

<li>objectify.DataElement didn't set up None value correctly</li>

1182

<li>objectify.DataElement didn't check the value against the provided type hints</li>

1183

<li>Reference-counting bug in <tt class="docutils literal">Element.attrib.pop()</tt></li>

1184

</ul>

1185

</div>

1186

</div>

1187

1188

1189

1190

<h2>Features added</h2>

1191

1192

<li>Module <tt class="docutils literal">lxml.pyclasslookup</tt> module implements an Element class lookup

1193

scheme that can access the entire tree in read-only mode to help determining

1194

a suitable Element class</li>

1195

<li>Parsers take a <tt class="docutils literal">remove_comments</tt> keyword argument that skips over comments</li>

1196

<li><tt class="docutils literal">parse()</tt> function in <tt class="docutils literal">objectify</tt>, corresponding to <tt class="docutils literal">XML()</tt> etc.</li>

1197

<li><tt class="docutils literal">Element.addnext(el)</tt> and <tt class="docutils literal">Element.addprevious(el)</tt> methods to support

1198

adding processing instructions and comments around the root node</li>

1199

<li><tt class="docutils literal">Element.attrib</tt> was missing <tt class="docutils literal">clear()</tt> and <tt class="docutils literal">pop()</tt> methods</li>

1200

<li>Extended type annotation in objectify: cleaner annotation namespace setup

1201

plus new <tt class="docutils literal">deannotate()</tt> function</li>

1202

<li>Support for custom Element class instantiation in lxml.sax: passing a

1203

<tt class="docutils literal">makeelement</tt> function to the ElementTreeContentHandler will reuse the

1204

lookup context of that function</li>

1205

<li>'.' represents empty ObjectPath (identity)</li>

1206

<li><tt class="docutils literal">Element.values()</tt> to accompany the existing <tt class="docutils literal">.keys()</tt> and <tt class="docutils literal">.items()</tt></li>

1207

<li><tt class="docutils literal">collectAttributes()</tt> C-function to build a list of attribute

1208

keys/values/items for a libxml2 node</li>

1209

<li><tt class="docutils literal">DTD</tt> validator class (like <tt class="docutils literal">RelaxNG</tt> and <tt class="docutils literal">XMLSchema</tt>)</li>

1210

<li>HTML generator helpers by Fredrik Lundh in <tt class="docutils literal">lxml.htmlbuilder</tt></li>

1211

<li><tt class="docutils literal">ElementMaker</tt> XML generator by Fredrik Lundh in <tt class="docutils literal">lxml.builder.E</tt></li>

1212

<li>Support for pickeling <tt class="docutils literal">objectify.ObjectifiedElement</tt> objects to XML</li>

1213

<li><tt class="docutils literal">update()</tt> method on Element.attrib</li>

1214

<li>Optimised replacement for libxml2's _xmlReconsiliateNs(). This allows lxml

1215

a better handling of namespaces when moving elements between documents.</li>

1216

</ul>

1217

</div>

1218

1219

<h2>Bugs fixed</h2>

1220

1221

<li>Removing Elements from a tree could make them loose their namespace

1222

declarations</li>

1223

<li><tt class="docutils literal">ElementInclude</tt> didn't honour base URL of original document</li>

1224

<li>Replacing the children slice of an Element would cut off the tails of the

1225

original children</li>

1226

<li><tt class="docutils literal">Element.getiterator(tag)</tt> did not accept <tt class="docutils literal">Comment</tt> and

1227

<tt class="docutils literal">ProcessingInstruction</tt> as tags</li>

1228

<li>API functions now check incoming strings for XML conformity. Zero bytes or

1229

low ASCII characters are no longer accepted (AssertionError).</li>

1230

<li>XSLT parsing failed to pass resolver context on to imported documents</li>

1231

<li>passing '' as namespace prefix in nsmap could be passed through to libxml2</li>

1232

<li>Objectify couldn't handle prefixed XSD type names in <tt class="docutils literal">xsi:type</tt></li>

1233

<li>More ET compatible behaviour when writing out XML declarations or not</li>

1234

<li>More robust error handling in <tt class="docutils literal">iterparse()</tt></li>

1235

<li>Documents lost their top-level PIs and comments on serialisation</li>

1236

<li>lxml.sax failed on comments and PIs. Comments are now properly ignored and

1237

PIs are copied.</li>

1238

<li>Possible memory leaks in namespace handling when moving elements between

1239

documents</li>

1240

</ul>

1241

</div>

1242

1243

<h2>Other changes</h2>

1244

1245

<li>major restructuring in the documentation</li>

1246

</ul>

1247

</div>

1248

</div>

1249

1250

1251

1252

<h2>Bugs fixed</h2>

1253

1254

<li>Build fixes for MS compiler</li>

1255

<li>Item assignments to special names like <tt class="docutils literal"><span class="pre">element["text"]</span></tt> failed</li>

1256

<li>Renamed ObjectifiedDataElement.__setText() to _setText() to make it easier

1257

to access</li>

1258

<li>The pattern for attribute names in ObjectPath was too restrictive</li>

1259

</ul>

1260

</div>

1261

</div>

1262

1263

1264

1265

<h2>Features added</h2>

1266

1267

<li>Rich comparison of QName objects</li>

1268

<li>Support for regular expressions in benchmark selection</li>

1269

<li>get/set emulation (not .attrib!) for attributes on processing instructions</li>

1270

<li>ElementInclude Python module for ElementTree compatible XInclude processing

1271

that honours custom resolvers registered with the source document</li>

1272

<li>ElementTree.parser property holds the parser used to parse the document</li>

1273

<li>setup.py has been refactored for greater readability and flexibility</li>

1274

<li>--rpath flag to setup.py to induce automatic linking-in of dynamic library

1275

runtime search paths has been renamed to --auto-rpath. This makes it

1276

possible to pass an --rpath directly to distutils; previously this was being

1277

shadowed.</li>

1278

</ul>

1279

</div>

1280

1281

<h2>Bugs fixed</h2>

1282

1283

<li>Element instantiation now uses locks to prevent race conditions with threads</li>

1284

<li>ElementTree.write() did not raise an exception when the file was not writable</li>

1285

<li>Error handling could crash under Python <= 2.4.1 - fixed by disabling thread

1286

support in these environments</li>

1287

<li>Element.find*() did not accept QName objects as path</li>

1288

</ul>

1289

</div>

1290

1291

<h2>Other changes</h2>

1292

1293

<li>code cleanup: redundant _NodeBase super class merged into _Element class

1294

Note: although the impact should be zero in most cases, this change breaks

1295

the compatibiliy of the public C-API</li>

1296

</ul>

1297

</div>

1298

</div>

1299

1300

1301

1302

<h2>Features added</h2>

1303

1304

<li>Data elements in objectify support repr(), which is now used by dump()</li>

1305

<li>Source distribution now ships with a patched Pyrex</li>

1306

<li>New C-API function makeElement() to create new elements with text,

1307

tail, attributes and namespaces</li>

1308

<li>Reuse original parser flags for XInclude</li>

1309

<li>Simplified support for handling XSLT processing instructions</li>

1310

</ul>

1311

</div>

1312

1313

<h2>Bugs fixed</h2>

1314

1315

<li>Parser resources were not freed before the next parser run</li>

1316

<li>Open files and XML strings returned by Python resolvers were not

1317

closed/freed</li>

1318

<li>Crash in the IDDict returned by XMLDTDID</li>

1319

<li>Copying Comments and ProcessingInstructions failed</li>

1320

<li>Memory leak for external URLs in _XSLTProcessingInstruction.parseXSL()</li>

1321

<li>Memory leak when garbage collecting tailed root elements</li>

1322

<li>HTML script/style content was not propagated to .text</li>

1323

<li>Show text xincluded between text nodes correctly in .text and .tail</li>

1324

<li>'integer * objectify.StringElement' operation was not supported</li>

1325

</ul>

1326

</div>

1327

</div>

1328

1329

1330

1331

<h2>Features added</h2>

1332

1333

<li>XSLT profiling support (<tt class="docutils literal">profile_run</tt> keyword)</li>

1334

<li>countchildren() method on objectify.ObjectifiedElement</li>

1335

<li>Support custom elements for tree nodes in lxml.objectify</li>

1336

</ul>

1337

</div>

1338

1339

<h2>Bugs fixed</h2>

1340

1341

<li>lxml.objectify failed to support long data values (e.g., "123L")</li>

1342

<li>Error messages from XSLT did not reach <tt class="docutils literal">XSLT.error_log</tt></li>

1343

<li>Factories objectify.Element() and objectify.DataElement() were missing

1344

<tt class="docutils literal">attrib</tt> and <tt class="docutils literal">nsmap</tt> keyword arguments</li>

1345

<li>Changing the default parser in lxml.objectify did not update the factories

1346

Element() and DataElement()</li>

1347

<li>Let lxml.objectify.Element() always generate tree elements (not data

1348

elements)</li>

1349

<li>Build under Windows failed ('0' bug in patched Pyrex version)</li>

1350

</ul>

1351

</div>

1352

</div>

1353

1354

1355

1356

<h2>Features added</h2>

1357

1358

<li>Comments and processing instructions return '' and

1359

'<?pi-target content?>' for repr()</li>

1360

<li>Parsers are now the preferred (and default) place where element class lookup

1361

schemes should be registered. Namespace lookup is no longer supported by

1362

default.</li>

1363

<li>Support for Python 2.5 beta</li>

1364

<li>Unlock the GIL for deep copying documents and for XPath()</li>

1365

<li>New <tt class="docutils literal">compact</tt> keyword argument for parsing read-only documents</li>

1366

<li>Support for parser options in iterparse()</li>

1367

<li>The <tt class="docutils literal">namespace</tt> axis is supported in XPath and returns (prefix, URI)

1368

tuples</li>

1369

<li>The XPath expression "/" now returns an empty list instead of raising an

1370

exception</li>

1371

<li>XML-Object API on top of lxml (lxml.objectify)</li>

1372

<li>Customizable Element class lookup:<ul>

1373

<li>different pre-implemented lookup mechanisms</li>

1374

<li>support for externally provided lookup functions</li>

1375

</ul>

1376

</li>

1377

<li>Support for processing instructions (ET-like, not compatible)</li>

1378

<li>Public C-level API for independent extension modules</li>

1379

<li>Module level <tt class="docutils literal">iterwalk()</tt> function as 'iterparse' for trees</li>

1380

<li>Module level <tt class="docutils literal">iterparse()</tt> function similar to ElementTree (see

1381

documentation for differences)</li>

1382

<li>Element.nsmap property returns a mapping of all namespace prefixes known at

1383

the Element to their namespace URI</li>

1384

<li>Reentrant threading support in RelaxNG, XMLSchema and XSLT</li>

1385

<li>Threading support in parsers and serializers:<ul>

1386

<li>All in-memory operations (tostring, parse(StringIO), etc.) free the GIL</li>

1387

<li>File operations (on file names) free the GIL</li>

1388

<li>Reading from file-like objects frees the GIL and reacquires it for reading</li>

1389

<li>Serialisation to file-like objects is single-threaded (high lock overhead)</li>

1390

</ul>

1391

</li>

1392

<li>Element iteration over XPath axes:<ul>

1393

<li>Element.iterdescendants() iterates over the descendants of an element</li>

1394

<li>Element.iterancestors() iterates over the ancestors of an element (from

1395

parent to parent)</li>

1396

<li>Element.itersiblings() iterates over either the following or preceding

1397

siblings of an element</li>

1398

<li>Element.iterchildren() iterates over the children of an element in either

1399

direction</li>

1400

<li>All iterators support the <tt class="docutils literal">tag</tt> keyword argument to restrict the

1401

generated elements</li>

1402

</ul>

1403

</li>

1404

<li>Element.getnext() and Element.getprevious() return the direct siblings of an

1405

element</li>

1406

</ul>

1407

</div>

1408

1409

<h2>Bugs fixed</h2>

1410

1411

<li>filenames with local 8-bit encoding were not supported</li>

1412

<li>1.1beta did not compile under Python 2.3</li>

1413

<li>ignore unknown 'pyval' attribute values in objectify</li>

1414

<li>objectify.ObjectifiedElement.addattr() failed to accept Elements and Lists</li>

1415

<li>objectify.ObjectPath.setattr() failed to accept Elements and Lists</li>

1416

<li>XPathSyntaxError now inherits from XPathError</li>

1417

<li>Threading race conditions in RelaxNG and XMLSchema</li>

1418

<li>Crash when mixing elements from XSLT results into other trees, concurrent

1419

XSLT is only allowed when the stylesheet was parsed in the main thread</li>

1420

<li>The EXSLT <tt class="docutils literal">regexp:match</tt> function now works as defined (except for some

1421

differences in the regular expression syntax)</li>

1422

<li>Setting element.text to '' returned None on request, not the empty string</li>

1423

<li><tt class="docutils literal">iterparse()</tt> could crash on long XML files</li>

1424

<li>Creating documents no longer copies the parser for later URL resolving. For

1425

performance reasons, only a reference is kept. Resolver updates on the

1426

parser will now be reflected by documents that were parsed before the

1427

change. Although this should rarely become visible, it is a behavioral

1428

change from 1.0.</li>

1429

</ul>

1430

</div>

1431

</div>

1432

1433

1434

1435

<h2>Features added</h2>

1436

1437

<li>List-like <tt class="docutils literal">Element.extend()</tt> method</li>

1438

</ul>

1439

</div>

1440

1441

<h2>Bugs fixed</h2>

1442

1443

<li>Crash in tail handling in <tt class="docutils literal">Element.replace()</tt></li>

1444

</ul>

1445

</div>

1446

</div>

1447

1448

1449

1450

<h2>Features added</h2>

1451

1452

<li>Element.replace(old, new) method to replace a subelement by another one</li>

1453

</ul>

1454

</div>

1455

1456

<h2>Bugs fixed</h2>

1457

1458

<li>Crash when mixing elements from XSLT results into other trees</li>

1459

<li>Copying/deepcopying did not work for ElementTree objects</li>

1460

<li>Setting an attribute to a non-string value did not raise an exception</li>

1461

<li>Element.remove() deleted the tail text from the removed Element</li>

1462

</ul>

1463

</div>

1464

</div>

1465

1466

1467

1468

<h2>Features added</h2>

1469

1470

<li>Support for setting a custom default Element class as opposed to namespace

1471

specific classes (which still override the default class)</li>

1472

</ul>

1473

</div>

1474

1475

<h2>Bugs fixed</h2>

1476

1477

<li>Rare exceptions in Python list functions were not handled</li>

1478

<li>Parsing accepted unicode strings with XML encoding declaration in certain

1479

cases</li>

1480

<li>Parsing 8-bit encoded strings from StringIO objects raised an exception</li>

1481

<li>Module function <tt class="docutils literal">initThread()</tt> was removed - useless (and never worked)</li>

1482

<li>XSLT and parser exception messages include the error line number</li>

1483

</ul>

1484

</div>

1485

</div>

1486

1487

1488

1489

<h2>Features added</h2>

1490

1491

<li>Repeated calls to Element.attrib now efficiently return the same instance</li>

1492

</ul>

1493

</div>

1494

1495

<h2>Bugs fixed</h2>

1496

1497

<li>Document deallocation could crash in certain garbage collection scenarios</li>

1498

<li>Extension function calls in XSLT variable declarations could break the

1499

stylesheet and crash on repeated calls</li>

1500

<li>Deep copying Elements could loose namespaces declared in parents</li>

1501

<li>Deep copying Elements did not copy tail</li>

1502

<li>Parsing file(-like) objects failed to load external entities</li>

1503

<li>Parsing 8-bit strings from file(-like) objects raised an exception</li>

1504

<li>xsl:include failed when the stylesheet was parsed from a file-like object</li>

1505

<li>lxml.sax.ElementTreeProducer did not call startDocument() / endDocument()</li>

1506

<li>MSVC compiler complained about long strings (supports only 2048 bytes)</li>

1507

</ul>

1508

</div>

1509

</div>

1510

1511

1512

1513

<h2>Features added</h2>

1514

1515

<li>Element.getiterator() and the findall() methods support finding arbitrary

1516

elements from a namespace (pattern <tt class="docutils literal">{namespace}*</tt>)</li>

1517

<li>Another speedup in tree iteration code</li>

1518

<li>General speedup of Python Element object creation and deallocation</li>

1519

<li>Writing C14N no longer serializes in memory (reduced memory footprint)</li>

1520

<li>PyErrorLog for error logging through the Python <tt class="docutils literal">logging</tt> module</li>

1521

<li><tt class="docutils literal">Element.getroottree()</tt> returns an ElementTree for the root node of the

1522

document that contains the element.</li>

1523

<li>ElementTree.getpath(element) returns a simple, absolute XPath expression to

1524

find the element in the tree structure</li>

1525

<li>Error logs have a <tt class="docutils literal">last_error</tt> attribute for convenience</li>

1526

<li>Comment texts can be changed through the API</li>

1527

<li>Formatted output via <tt class="docutils literal">pretty_print</tt> keyword in serialization functions</li>

1528

<li>XSLT can block access to file system and network via <tt class="docutils literal">XSLTAccessControl</tt></li>

1529

<li>ElementTree.write() no longer serializes in memory (reduced memory

1530

footprint)</li>

1531

<li>Speedup of Element.findall(tag) and Element.getiterator(tag)</li>

1532

<li>Support for writing the XML representation of Elements and ElementTrees to

1533

Python unicode strings via <tt class="docutils literal">etree.tounicode()</tt></li>

1534

<li>Support for writing XSLT results to Python unicode strings via <tt class="docutils literal">unicode()</tt></li>

1535

<li>Parsing a unicode string no longer copies the string (reduced memory

1536

footprint)</li>

1537

<li>Parsing file-like objects reads chunks rather than the whole file (reduced

1538

memory footprint)</li>

1539

<li>Parsing StringIO objects from the start avoids copying the string (reduced

1540

memory footprint)</li>

1541

<li>Read-only 'docinfo' attribute in ElementTree class holds DOCTYPE

1542

information, original encoding and XML version as seen by the parser</li>

1543

<li>etree module can be compiled without libxslt by commenting out the line

1544

<tt class="docutils literal">include "xslt.pxi"</tt> near the end of the etree.pyx source file</li>

1545

<li>Better error messages in parser exceptions</li>

1546

<li>Error reporting also works in XSLT</li>

1547

<li>Support for custom document loaders (URI resolvers) in parsers and XSLT,

1548

resolvers are registered at parser level</li>

1549

<li>Implementation of exslt:regexp for XSLT based on the Python 're' module,

1550

enabled by default, can be switched off with 'regexp=False' keyword argument</li>

1551

<li>Support for exslt extensions (libexslt) and libxslt extra functions

1552

(node-set, document, write, output)</li>

1553

<li>Substantial speedup in XPath.evaluate()</li>

1554

<li>HTMLParser for parsing (broken) HTML</li>

1555

<li>XMLDTDID function parses XML into tuple (root node, ID dict) based on xml:id

1556

implementation of libxml2 (as opposed to ET compatible XMLID)</li>

1557

</ul>

1558

</div>

1559

1560

<h2>Bugs fixed</h2>

1561

1562

<li>Memory leak in Element.__setitem__</li>

1563

<li>Memory leak in Element.attrib.items() and Element.attrib.values()</li>

1564

<li>Memory leak in XPath extension functions</li>

1565

<li>Memory leak in unicode related setup code</li>

1566

<li>Element now raises ValueError on empty tag names</li>

1567

<li>Namespace fixing after moving elements between documents could fail if the

1568

source document was freed too early</li>

1569

<li>Setting namespace-less tag names on namespaced elements ('{ns}t' -> 't')

1570

didn't reset the namespace</li>

1571

<li>Unknown constants from newer libxml2 versions could raise exceptions in the

1572

error handlers</li>

1573

<li>lxml.etree compiles much faster</li>

1574

<li>On libxml2 <= 2.6.22, parsing strings with encoding declaration could fail

1575

in certain cases</li>

1576

<li>Document reference in ElementTree objects was not updated when the root

1577

element was moved to a different document</li>

1578

<li>Running absolute XPath expressions on an Element now evaluates against the

1579

root tree</li>

1580

<li>Evaluating absolute XPath expressions (<tt class="docutils literal">/*</tt>) on an ElementTree could fail</li>

1581

<li>Crashes when calling XSLT, RelaxNG, etc. with uninitialized ElementTree

1582

objects</li>

1583

<li>Removed public function <tt class="docutils literal">initThreadLogging()</tt>, replaced by more general

1584

<tt class="docutils literal">initThread()</tt> which fixes a number of setup problems in threads</li>

1585

<li>Memory leak when using iconv encoders in tostring/write</li>

1586

<li>Deep copying Elements and ElementTrees maintains the document information</li>

1587

<li>Serialization functions raise LookupError for unknown encodings</li>

1588

<li>Memory deallocation crash resulting from deep copying elements</li>

1589

<li>Some ElementTree methods could crash if the root node was not initialized

1590

(neither file nor element passed to the constructor)</li>

1591

<li>Element/SubElement failed to set attribute namespaces from passed <tt class="docutils literal">attrib</tt>

1592

dictionary</li>

1593

<li><tt class="docutils literal">tostring()</tt> adds an XML declaration for non-ASCII encodings</li>

1594

<li><tt class="docutils literal">tostring()</tt> failed to serialize encodings that contain 0-bytes</li>

1595

<li>ElementTree.xpath() and XPathDocumentEvaluator were not using the

1596

ElementTree root node as reference point</li>

1597

<li>Calling <tt class="docutils literal"><span class="pre">document('')</span></tt> in XSLT failed to return the stylesheet</li>

1598

</ul>

1599

</div>

1600

</div>

1601

1602

1603

1604

<h2>Features added</h2>

1605

1606

<li>Speedup for Element.makeelement(): the new element reuses the original

1607

libxml2 document instead of creating a new empty one</li>

1608

<li>Speedup for reversed() iteration over element children (Py2.4+ only)</li>

1609

<li>ElementTree compatible QName class</li>

1610

<li>RelaxNG and XMLSchema accept any Element, not only ElementTrees</li>

1611

</ul>

1612

</div>

1613

1614

<h2>Bugs fixed</h2>

1615

1616

<li>str(xslt_result) was broken for XSLT output other than UTF-8</li>

1617

<li>Memory leak if write_c14n fails to write the file after conversion</li>

1618

<li>Crash in XMLSchema and RelaxNG when passing non-schema documents</li>

1619

<li>Memory leak in RelaxNG() when RelaxNGParseError is raised</li>

1620

</ul>

1621

</div>

1622

</div>

1623

1624

1625

1626

<h2>Features added</h2>

1627

1628

<li>lxml.sax.ElementTreeContentHandler checks closing elements and raises

1629

SaxError on mismatch</li>

1630

<li>lxml.sax.ElementTreeContentHandler supports namespace-less SAX events

1631

(startElement, endElement) and defaults to empty attributes (keyword

1632

argument)</li>

1633

<li>Speedup for repeatedly accessing element tag names</li>

1634

<li>Minor API performance improvements</li>

1635

</ul>

1636

</div>

1637

1638

<h2>Bugs fixed</h2>

1639

1640

<li>Memory deallocation bug when using XSLT output method "html"</li>

1641

<li>sax.py was handling UTF-8 encoded tag names where it shouldn't</li>

1642

<li>lxml.tests package will no longer be installed (is still in source tar)</li>

1643

</ul>

1644

</div>

1645

</div>

1646

1647

1648

1649

<h2>Features added</h2>

1650

1651

<li>Error logging API for libxml2 error messages</li>

1652

<li>Various performance improvements</li>

1653

<li>Benchmark script for lxml, ElementTree and cElementTree</li>

1654

<li>Support for registering extension functions through new FunctionNamespace

1655

class (see doc/extensions.txt)</li>

1656

<li>ETXPath class for XPath expressions in ElementTree notation ('//{ns}tag')</li>

1657

<li>Support for variables in XPath expressions (also in XPath class)</li>

1658

<li>XPath class for compiled XPath expressions</li>

1659

<li>XMLID module level function (ElementTree compatible)</li>

1660

<li>XMLParser API for customized libxml2 parser configuration</li>

1661

<li>Support for custom Element classes through new Namespace API (see

1662

doc/namespace_extensions.txt)</li>

1663

<li>Common exception base class LxmlError for module exceptions</li>

1664

<li>real iterator support in iter(Element), Element.getiterator()</li>

1665

<li>XSLT objects are callable, result trees support str()</li>

1666

<li>Added MANIFEST.in for easier creation of RPM files.</li>

1667

<li>'getparent' method on elements allows navigation to an element's

1668

parent element.</li>

1669

<li>Python core compatible SAX tree builder and SAX event generator. See

1670

doc/sax.txt for more information.</li>

1671

</ul>

1672

</div>

1673

1674

<h2>Bugs fixed</h2>

1675

1676

<li>Segfaults and memory leaks in various API functions of Element</li>

1677

<li>Segfault in XSLT.tostring()</li>

1678

<li>ElementTree objects no longer interfere, Elements can be root of different

1679

ElementTrees at the same time</li>

1680

<li>document('') works in XSLT documents read from files (in-memory documents

1681

cannot support this due to libxslt deficiencies)</li>

1682

</ul>

1683

</div>

1684

</div>

1685

1686

1687

1688

<h2>Features added</h2>

1689

1690

<li>Support for copy.deepcopy() on elements. copy.copy() works also, but

1691

does the same thing, and does <em>not</em> create a shallow copy, as that

1692

makes no sense in the context of libxml2 trees. This means a

1693

potential incompatibility with ElementTree, but there's more chance

1694

that it works than if copy.copy() isn't supported at all.</li>

1695

<li>Increased compatibility with (c)ElementTree; .parse() on ElementTree is

1696

supported and parsing of gzipped XML files works.</li>

1697

<li>implemented index() on elements, allowing one to find the index of a

1698

SubElement.</li>

1699

</ul>

1700

</div>

1701

1702

<h2>Bugs fixed</h2>

1703

1704

<li>Use xslt-config instead of xml2-config to find out libxml2

1705

directories to take into account a case where libxslt is installed

1706

in a different directory than libxslt.</li>

1707

<li>Eliminate crash condition in iteration when text nodes are changed.</li>

1708

<li>Passing 'None' to tostring() does not result in a segfault anymore,

1709

but an AssertionError.</li>

1710

<li>Some test fixes for Windows.</li>

1711

<li>Raise XMLSyntaxError and XPathSyntaxError instead of plain python

1712

syntax errors. This should be less confusing.</li>

1713

<li>Fixed error with uncaught exception in Pyrex code.</li>

1714

<li>Calling lxml.etree.fromstring('') throws XMLSyntaxError instead of a

1715

segfault.</li>

1716

<li>has_key() works on attrib. 'in' tests also work correctly on attrib.</li>

1717

<li>INSTALL.txt was saying 2.2.16 instead of 2.6.16 as a supported

1718

libxml2 version, as it should.</li>

1719

<li>Passing a UTF-8 encoded string to the XML() function would fail;

1720

fixed.</li>

1721

</ul>

1722

</div>

1723

</div>

1724

1725

1726

1727

<h2>Features added</h2>

1728

1729

<li>parameters (XPath expressions) can be passed to XSLT using keyword

1730

parameters.</li>

1731

<li>Simple XInclude support. Calling the xinclude() method on a tree

1732

will process any XInclude statements in the document.</li>

1733

<li>XMLSchema support. Use the XMLSchema class or the convenience

1734

xmlschema() method on a tree to do XML Schema (XSD) validation.</li>

1735

<li>Added convenience xslt() method on tree. This is less efficient

1736

than the XSLT object, but makes it easier to write quick code.</li>

1737

<li>Added convenience relaxng() method on tree. This is less efficient

1738

than the RelaxNG object, but makes it easier to write quick code.</li>

1739

<li>Make it possible to use XPathEvaluator with elements as well. The

1740

XPathEvaluator in this case will retain the element so multiple

1741

XPath queries can be made against one element efficiently. This

1742

replaces the second argument to the .evaluate() method that existed

1743

previously.</li>

1744

<li>Allow registerNamespace() to be called on an XPathEvaluator, after

1745

creation, to add additional namespaces. Also allow registerNamespaces(),

1746

which does the same for a namespace dictionary.</li>

1747

<li>Add 'prefix' attribute to element to be able to read prefix information.

1748

This is entirely read-only.</li>

1749

<li>It is possible to supply an extra nsmap keyword parameter to

1750

the Element() and SubElement() constructors, which supplies a

1751

prefix to namespace URI mapping. This will create namespace

1752

prefix declarations on these elements and these prefixes will show up

1753

in XML serialization.</li>

1754

</ul>

1755

</div>

1756

1757

<h2>Bugs fixed</h2>

1758

1759

<li>Killed yet another memory management related bug: trees created

1760

using newDoc would not get a libxml2-level dictionary, which caused

1761

problems when deallocating these documents later if they contained a

1762

node that came from a document with a dictionary.</li>

1763

<li>Moving namespaced elements between documents was problematic as

1764

references to the original document would remain. This has been fixed

1765

by applying xmlReconciliateNs() after each move operation.</li>

1766

<li>Can pass None to 'dump()' without segfaults.</li>

1767

<li>tostring() works properly for non-root elements as well.</li>

1768

<li>Cleaned out the tostring() method so it should handle encoding

1769

correctly.</li>

1770

<li>Cleaned out the ElementTree.write() method so it should handle encoding

1771

correctly. Writing directly to a file should also be faster, as there is no

1772

need to go through a Python string in that case. Made sure the test cases

1773

test both serializing to StringIO as well as serializing to a real file.</li>

1774

</ul>

1775

</div>

1776

</div>

1777

1778

1779

1780

<h2>Features added</h2>

1781

1782

<li>Changed setup.py so that library_dirs is also guessed. This should

1783

help with compilation on the Mac OS X platform, where otherwise the

1784

wrong library (shipping with the OS) could be picked up.</li>

1785

<li>Tweaked setup.py so that it picks up the version from version.txt.</li>

1786

</ul>

1787

</div>

1788

1789

<h2>Bugs fixed</h2>

1790

1791

<li>Do the right thing when handling namespaced attributes.</li>

1792

<li>fix bug where tostring() moved nodes into new documents. tostring()

1793

had very nasty side-effects before this fix, sorry!</li>

1794

</ul>

1795

</div>

1796

</div>

1797

1798

1799

1800

<li>Python 2.2 compatibility fixes.</li>

1801

<li>unicode fixes in Element() and Comment() as well as XML(); unicode

1802

input wasn't properly being UTF-8 encoded.</li>

1803

</ul>

1804

</div>

1805

1806

1807

<p>Initial public release.</p>

1808

</div>

1809

</div>

1810

1811

1812

Generated on: 2010-03-02.

1813

1814

</div>

1815

</body>

1816

</html>

Older »