I am sure we'll find some corners that need to be ironed but it is considerably better than 1.99.7. It adds lot of new features and improves many existing ones:
- 57,000 new lines of code and closed 279 issues since 1.99.7.
- Has a new scheduler, a built-in wiki, new language and pluralization system, better markmin with oembed support and better scaffolding app, increased security.
- Lots of experimental features including GIS support, mongodb support, built-in auth.wiki(), and more.
Should be 100% backward compatible. If you run into any issue let us know ASAP.
I personally want to thank the major contributors to this release (in alphabetic order)
Alan, Andrew, Anthony, Bruno, Christian, Dave, Dominic, Iceberg, Jonathan, Marc, Mariano, Marin, Martin, Mark, Michael, Michele, Niphlod, Patrick, Vladyslav,
They spend many nights testing, coding, debugging at a very fast pace.
Many many people have contributed.
## 2.00.2
### DAL Improvements
- Support for DAL(lazy_tables=True) and db.define_table(on_define=lambda table:), thanks Jonathan
- db(...).select(cacheable=True) make select 30% faster
- db(...).select(cache=(cache.ram,3600)) now caches parsed data 100x faster
- db(...).count(cache=(cache.ram,3600)) now supported
- MongoDB support in DAL (experimental), thanks Mark Breedveld
- geodal and spatialite, thanks Denes and Fran (experimental)
- db.mytable._before_insert, _after_insert, _before_update, _after_update, _before_delete. _after_delete (list of callbacks)
- db(...).update_naive(...) same as update but ignores table._before_update and table._after_update
- DAL BIGINT support and DAL(...,bigint_id=True)
- IS_IN_DB(..., distinct=True)
- new syntax: db.mytable.insert(myuploadfield=open(....)), thank you Iceberg
- db(...).select(db.mytable.myfield.count(distinct=True))
- db.mytable.myfield.filter_in, filter_out
- db.mytable._enable_record_versioning(db) adds versioning to this table
- teradata adapter, thanks Andrew Willimott
- experimental Sybase Adapter
- added db.table.field.avg()
- Support for Google App Engine projections, thanks Christian
- Field(... 'upload', default=path) now accepts a path to a local file as default value, if user does not upload a file. Relative path looks inside current application folder, thanks Marin
- executesql(...,fields=,columns=) allows parsing of results in Rows, thanks Anthony
### Auth improvements
- auth.enable_record_versioning(db) adds full versioning to all tables
- @auth.requires_login(otherwise=URL(...))
- auth supports salt and compatible with third party data, thanks Dave Stoll
- CRYPT now defaults to pbkdf2(1000,20,sha1)
- Built-in wiki with menu, tags, search, media, permissions. def index: return auth.wiki()
- auth.settings.everybody_group_id
- allow storage of uploads on any PyFileSystem (including amazon)
### Form improvements
- FORM.confirm('Are you sure?',{'Back':URL(...)})
- SQLFORM.smartdictform(dict)
- form.add_button(value,link)
- SQLFORM.grid(groupby='...')
- fixed security issue with SQLFORM.grid and SQLFORM.smartgrid
- more export options in SQLFORM.grid and SQLFORM.smartgrid (html, xml, csv, ...)
### Admin improvements
- new admin pages: manage_students, bulk_regsiter, and progress reports
- increased security in admin against CSRF
- experimental Git integration
- experimental OpenShift deployment
- multi-language pluralization engine
- ace text web editor in admin
- Ukrainian translations, thanks Vladyslav Kozlovskyy
- Romanian translation for welcome, thanks ionel
- support for mercurial 2.6, thanks Vlad
### Scheduler Improvements (thanks to niphlod, ykessler, dhx, toomim)
- web2py.py -K myapp -X starts the myapp scheduler alongside the webserver
- tasks are marked EXPIRED (if stop_time passed)
- functions with no result don't end up in scheduler_run
- more options: web2py.py -E -b -L
- scheduler can now handle 10k tasks with 20 concurrent workers and with no issues
- new params:
tasks can be found in the environment (no need to define the tasks parameter)
max_empty_runs kills the workers automatically if no new tasks are found in queue (nice for "spikes" of processing power)
discard_results to completely discard the results (if you don't need the output of the task)
utc_time enables datetime calculations with UTC time
- scheduler_task changes:
task_name is no longer required (filled automatically with function_name if found empty)
uuid makes easy to coordinate scheduler_task maintenance (filled automatically if not provided)
stop_time has no default (previously was today+1)
retry_failed to requeue automatically failed tasks
sync_output refreshes automatically the output (nice to report percentages)
- workers can be:
DISABLED (put to sleep and do nothing if not sending the heartbeat every 30 seconds)
TERMINATE (complete the current task and then die)
KILL (kill ASAP)
### Other Improvements
- gluon/contrib/webclient.py makes it easy to create functional tests for app
- DIV(..).elements(...replace=...), thanks Anthony
- new layout based on Twitter Bootstrap
- New generic views: generic.ics (Mac Mail Calendar) and generic.map (Google Maps)
- request.args(0,default=0, cast=int, otherwise=URL(...)), thanks Anthony
- redirect(...,type='auto') will be handled properly in ajax responses
- routes in can redirect outside with routes_in=[('/path','303->http://..')]
- better memcache support
- improved spreadsheet, thanks Alan
- new internationalization engine, thanks Vladyslav
- pluralization engine, thanks Vladyslav
- new markmin with support for nested lists, <i>, <em>, autolinks, thanks Vladyslav
- new syntax: {{=BR()*5}}
- gluon.cache.lazy_cache decorator allows caching functions in modules
- .coffee and .less support in response.files, thanks Sam Sheftel
- ldap certificate support
- pg8000 postgresql driver support (experimental)
- @cache('%(name)s%(args)s%(vars)s',5) and cache.autokey
- added tox.ini, thanks Marc
- web2py.py --run_system_tests, thanks Marc Abramowitz
- html.py (and web2py helpers) can be used without web2py dependencies
- new fpdf, thanks Mariano