Cool. Thanks! On Wed, Aug 19, 2009 at 2:58 PM, Ralf Schmitt<[email protected]> wrote: > > Ryan Lane <[email protected]> writes: > >> Where is the changelog? What things were improved? I spent the last 5 >> minutes looking for one and can't find it. > > we try to keep an changelog inside the readme in the top level > directory. they also get uploaded to the pypi description page: > > http://pypi.python.org/pypi/mwlib/ > http://pypi.python.org/pypi/mwlib.rl/ > > mwlib.rl/README.txt contains the following: > > ,---- > | 2009-08-17 release 0.12.1 > | ------------------------- > | * fixes > | * add Table of Contents > | * improve support for CJK languages > | * fix printing problems > | * use formatter class to style text > | * add CustomNodeTransformer > `---- > > > mwlib/README.txt contains: > > ,---- > | 2009-8-18 release 0.12.2 > | ------------------------ > | - fix status callbacks to pod partner > | > | 2009-8-17 release 0.12.1 > | ------------------------ > | - added mw-client and mw-check-service > | - mw-serve-ctl can now send report mails > | - fixes for race conditions in mwlib.filequeue (mw-watch) > | - lots of other improvements... > `---- > > where lots of other improvements is (according to > 'git shortlog 0.11.2..0.12.2'): > > ,---- > | [email protected] (65): > | Added tag 0.11.2 for changeset 95728196cb62 > | remove UTF-8 BOM > | typo > | upversion > | fallbacks for MW API w/out magicwords and namespacealiases > | unicode2str > | merge > | fix old zipwiki > | fix getContributorsFromInformationTempalte() > | merge > | only pass redirects=1 if titles are passed, *not* revids > | fix: remove tmpdir on clear() > | fix: some wikis do not return a query warning > | merge > | fallback for MW APIs that do not support namespacealiases > | no revs => no authors > | optional interwikimap > | another optional thing. not sure which degree of brokeness we should > support here... > | removed getLinkURL(). almost deprecated mwlib.wikidbbase. > | proper casing for class names > | added > | fix for #607 > | workaround until #609 is fixed > | fetch licenses before giving ctl to reactor > | fix (but the bug did not really hurt) > | prevent race conditions > | FS on WMF server has no O_EXLOCK, so we now simply require the > portable lockfile module > | fallback to English prefix if an image description page is not found > | fix > | merge > | support --imagesize again > | moved code to new modules mwlib.apps.{serve|watch}. added > --max-running-time option. > | refactored buildzip: always use unique tmpdirs when creating nuwikis > | merge > | renamed hack() to make_zip(), no longer use **kwargs and **locals() > | allow /tmp to be on a different fs > | close tmpfile > | made manifest > | merge > | test adjusted: getURL() for images is deprecated > | merge > | use zipfile + adapt instead of (tmp) nuwiki dir > | fix > | use httplib2 > | tidied up > | do not write files in report(). do not cgi.escape() mail body. allow > setting of mail headers. > | do not send mails on wrong parameters etc. > | fix > | tidy > | swap order, indent=4 > | no json on older Pythons > | add mail-sending capabilities to mw-serve-ctl > | fix: one beer too much? :-P > | merge > | fixes to keep mw-watch from crashing > | added mw-client > | do not JSON decode if command is download > | added mw-check-service > | added some changelog entries > | added collection.get_wiki() > | better impl > | --config/base_url not required any more > | fix > | use WebOb instead of custom request, response handling > | fix > | > | [email protected] (309): > | fix #595 > | removed. use mwlib/siteinfo/fetch_siteinfo.py > | add test for 595 > | merge > | fix 599 > | test for #607 > | merge > | Backed out changeset 11eb025ecb8a > | merge > | fix #607 > | add entry for pnt in fix_wikipedia_siteinfo > | implement nuwiki's interface in cdbwiki and use nuwiki.adapt > | do not read old cdb file format > | specify language in wikiconf for cdb wiki > | follow redirects > | tags inside comments should not be replaced with uniq patterns > | better output > | replace comments while replacing with UNIQ strings > | test comments > | remove unused Overlay code > | remove remove_comments > | mediawiki sends wrong query-continue replies. do not ask for the same > url again. > | update MANIFEST.in > | fix re2c detection > | stop looking for vlist on first [[ (fix #578) > | test for #578 > | merge > | merge > | python 2.4 fix > | use odfpy 0.9 > | better tree output > | add get_siteinfo method to dummydb > | fix #530 > | normalize redirect names. > | merge > | always set headers > | assume any unknown wiki config starting with ":" is on wikipedia.org > | add some (failing) tests for pre tags > | remove nowiki tags and replace html entities inside pre tags > | python 2.4 fix > | typo > | save all page names cdbwiki tried to load while resolving redirects > | fix documentation > | remove fzipcreator > | remove mwscan, scanner > | update MANIFEST.in > | shorten code > | clean code > | add tagparser > | use tagparser > | use tagparser for h tags > | make token walker return a list instead of walking the tree while it > is being changed > | remove unused blist code > | allow cython generated files in other dirs > | some cython stuff to speed things up a bit > | test s/var tag > | support s/var tags > | update > | get rid of the skip_types argument. set some artificial tagname for > sections > | merge > | speedup token walker a bit by not looking into empty lists > | clean code > | merge > | add articles method to nuwiki in order to make mw-parse -a work with > nuwikis > | switch defaultns to 0 when a name starts with a colon > | add test for empty interwiki links > | allow empty interwiki links > | replace get_recursive_tag_parser with tagparser. clean code. > | remove tools/build_namespace_langs.py > | add login method > | update MANIFEST.in > | support password protected wikis in twisted_api.py > | make twisted_api work without passwords > | remove old commented-out code > | add try_api_urls > | remove debug output > | add find_api_for_url > | shorten code > | fix refcounting > | start splitting twisted_api > | rename twisted_api to net/fetch > | update MANIFEST.in > | add pool class > | add linuxmem for debugging > | add multiplier class > | use multiplier. > | I need CAPSLOCK to type FSOutput, which I don't have. > | add missing import > | use empty status message > | remove unused imports > | remove netdb > | update MANIFEST.in > | use mwapi.pool > | remove some test code. clean code. > | remove zipcreator and jobsched. > | get rid of dependency to mwapidb > | update MANIFEST.in > | remove no-threads/num-threads argument > | add pytest_twisted > | merge > | unconditionally use twisted plugin > | typo > | handle multiple --config arguments in option parser. > | evaluate style="display:..." and set blocknode attribute > | make it work with latest pylib > | merge > | remove some unused variables > | make sure pytest_twisted can be found > | update MANIFEST.in > | make pyflakes happy > | make tests mostly work. > | merge > | add -xnet to keywords if XNET=1 this will allow us to skip network > tests per test function not per file. > | merge > | mark test as xnet > | make test_validatetags a generator test > | test_xhtmlwriter: make test_validatetags a generator test > | merge > | restart system calls on SIGCHLD signals (needed for OS X) this only > works in python >= 2.6. > | merge > | var, s tag support > | remove example_conf. does not work > | update MANIFEST.in > | try to make tests work > | script to build xml catalog file. > | add docbook xml cache > | update urls > | do not download if file is alredy available > | add more cached urls > | enable docbookwriter tests again > | update MANIFEST.in > | test source tag with and without enclose attribute. > | make source tags with enclose=None inline nodes. > | do not set dummy start/len attribute > | mark tests as xfail > | add test for urllink inside link > | parse url links before parsing links. this moves parse_urls before > six other parser steps and might cause havoc. > | merge > | do not lowercase enclose attribute > | fix test for #602. > | be quiet when running sdist > | be quiet > | test [[http://example.com bla]] > | Backed out changeset 4028435971b6 it didn't fix anything... > | merge > | move make_nuwiki to own file > | do not start/stop reactor in fetcher. > | make --collectionpage work again > | save images to temporary file first > | allow to pass in img_size > | update MANIFEST.in > | prevent recursion error. count each query-continued query. see > http://twistedmatrix.com/trac/ticket/411 > | pass cmlimit. the default of 10 is just too low > | test spaces before indented tables. > | remove spaces before indented tables when analyzing that stuff. > | test that sub/sup close each other > | make stray closing sup/sub close each other > | typo > | handle slice objects. > | revert to 1e555742aa81: do not use httplib2 > | merge > | merge > | no capslock here > | multiwiki > | merge > | shorten > | fix > | allow spaces in closing html tags > | test that spaces in closing tags work > | merge > | mark tests as xfail > | adapt to latest py lib > | use timelib if available > | test relative date (e.g. -12 days) > | use timelib > | test that parser function names can be constructed from templates > | fix expected value > | check magic_nodes registry after expanding template name > | remove duplicate/inline implementations of switch/if > | use classes for objects in a metabook and provide custom json > encoder/decoder, which can handle metabook objects. > | update MANIFEST.in > | fix interwiki class > | also show tree after treecleaner has run > | append slash to base_url in getURL > | pass sort_keys=True when dumping json > | update siteinfo > | make fix_wikipedia_siteinfo do nothing. pnt/arz language links are > now returned in the siteinfo from api.php > | make get return the default instead of None > | deferreds may be called immediately. > | do not create zip files when the fetch deferred returned an error. > reset signal handlers in order to make ctrl-c work again in mw-render. > | dump json with sort_keys=True > | show warning when using dict-like interface > | use dot syntax > | merge > | shorten calc_checksum > | remove unused import > | use myjson > | fix interwiki loading from old zip files > | merge > | use dot syntax > | use dot syntax > | fix: use metabook source class > | define walk/articles method on collections and use them > | python 2.4 fix > | implement make_interwiki > | update MANIFEST.in > | use dot syntax > | add wikiconf to metabook. > | remove namespace.py, recorddb, and mwapidb. > | update MANIFEST.in > | fix mw-render. > | remove lang.py/lang.txt > | remove code using mwapidb > | we've had enough of these hacks. refactor instead. > | rip mw-testserve, mw-html programs and mwlib/web.py > | rip rendermath, texmap > | update MANIFEST.in > | rip htmlwriter > | update MANIFEST.in > | remove resources > | rip wikidbbase > | get rid of getTemplate/getArticle > | bring zipwiki back to life > | update MANIFEST.in > | remove mwlib.resources > | fix > | make those mails a bit readable > | merge > | use dot syntax. store environment in each article from metabook. > | remove compat methods > | use dot syntax > | add wiki property to metabook articles > | do not copy properties into instance.__dict__ > | add multi-nuwiki > | set _wiki attribute on licenses > | return license objects > | use source object, not dict > | merge > | add images property > | shorten > | fix > | do not pass parameters explicitly > | pass print_template_pattern et al as arguments to makewiki. > | replace --login with --username, --password, --domain. > | fix > | handle print_template_prefix in parse_args. > | shorten > | add wikiconf > | fetch from multiple wikis in parallel > | fix > | add shared_progress > | feed status object > | use metabook.wikiconf instead of dummy_web_wiki and store in the the > wikiconf attribute since we are not really instantiating a real wiki object. > | fix > | create needed multiwiki data > | shorten. > | fix > | add multicoll.py > | fix getURL > | try rendering an old zip file with all working writers > | pass kwargs > | fix > | use pytest_generate_tests plugin. > | search for zip files > | add nuwiki zip file > | add multiwiki example zip > | up version to 0.12.1.dev > | fix > | merge > | merge > | fix license handling > | update MANIFEST.in > | remove debug output > | fix --collectionpage option > | do not accidentally use the python-json module. > | test that [[ inside #iferror works (test for > http://code.pediapress.com/wiki/ticket/648) > | fix issue with [[ inside #iferror > | test left-to-right and right-to-left mark > | strip left-to-right and right-to-left mark > | implement dl, dt, dd tags > | remove wiki property. > | better deprecation warnings > | shorten code > | make tests work with py 1.0 work around bug in inspect.isclass called > from py.test on test_refine.empty, which subsequently called > issubclass(empty, UnitTest)... > | do not use subdirectories when creating a nuwiki from one wiki, even > if it has an ident. > | check image links blocknode attribute > | set blocknode attribute of image links > | merge > | test that color values with hash marks work in parseParams > | fix color values with hash marks work in parseParams > | merge > | do not fetch the complete excluded-in-print category as it can get > rather large. fetch categories with each article/template... > | fix mw-client download command > | shorten code > | fix error handling in do_download > | fix handling of login_credentials. mw-render now uses > --user,--password,... instead of --login > | fix test: we do not fetch the whole excluded-in-print category > anymore. > | handle empty strings as revision. > | up version to 0.12.1 > | update docs > | Added tag 0.12.1 for changeset 59b2de9ad990 > | update version to 0.12.2.dev > | fix error handling for non POST requests > | fix mw-show > | workaround for #656 > | fix tests > | do not remove caption only tables > | enable status callbacks. > | up version to 0.12.2. update docs > | > | [email protected] (54): > | make cleaner method list class attr > | add method to export image license info > | better detection of big tables that need to be split > | merge > | treecleaner: remove empty rows at the end of a table. remove wrapper > tables > | fix > | merge > | use basedir as tmpdir > | fix for alignment > | merge > | fix image names > | merge > | merge > | allow empty chapters > | merge > | fix bug in styleutils.getTextAlign. add test > | linearize tables with large lists > | merge > | add skipMethods as class attr > | add formatter > | add grouping chars to formatter > | merge > | optimize heuristics for splitting tables into columns > | pass break_long arg to cleanText > | fix status > | handle spacer DIVs > | merge > | add box which has a border > | remove empty sections from parse tree > | add splitTableLists method to treecleaner. fix for formatter > | merge > | remove debug output > | merge > | merge > | merge > | improve status ranges > | merge > | optimize splitTableLists method > | add markShortParagraph treecleaner method > | merge > | correct test > | avoid recursions > | fix > | fix and test for splitTableLists > | fix option parsing > | add DefinitionList, Term and Description to advtree > | dont remove sections containing only textless galleries > | fix font switching in source tags > | add editor to metabook > | merge > | fix for t.numcols > | clean up > | add reports to treecleaner > | fix > | > | zencoder (1): > | use os.urandom instead of reading /dev/urandom > `---- > > Maintaining the changelog is a real PITA. Unfortunately automatically > creating that changelog also leaves some room for improvements... > > Regards, > - Ralf > > > >
--~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "mwlib" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/mwlib?hl=en -~----------~----~----~----~------~----~------~--~---
