Skip to content

Commit 9cc834d

Browse files
authored
Update changelog (#727)
1 parent f91af79 commit 9cc834d

1 file changed

Lines changed: 13 additions & 3 deletions

File tree

README.rst

Lines changed: 13 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -154,25 +154,31 @@ Changelog
154154
Python version eariler then 3.6 were likely broken with zeroconf
155155
already, however the version is now explictly checked.
156156

157-
* BREAKING CHANGE: RecordUpdateListener now uses update_records instead of update_record (#419) @bdraco
157+
* BREAKING CHANGE: RecordUpdateListener now uses async_update_records instead of update_record (#419, #726) @bdraco
158158

159159
This allows the listener to receive all the records that have
160160
been updated in a single transaction such as a packet or
161161
cache expiry.
162162

163-
update_record has been deprecated in favor of update_records
163+
update_record has been deprecated in favor of async_update_records
164164
A compatibility shim exists to ensure classes that use
165165
RecordUpdateListener as a base class continue to have
166166
update_record called, however they should be updated
167167
as soon as possible.
168168

169-
A new method update_records_complete is now called on each
169+
A new method async_update_records_complete is now called on each
170170
listener when all listeners have completed processing updates
171171
and the cache has been updated. This allows ServiceBrowsers
172172
to delay calling handlers until they are sure the cache
173173
has been updated as its a common pattern to call for
174174
ServiceInfo when a ServiceBrowser handler fires.
175175

176+
The async_ prefix was choosen to make it clear that these
177+
functions run in the eventloop and should never do blocking
178+
I/O. Before 0.32+ these functions ran in a select() loop and
179+
should not have been doing any blocking I/O, but it was not
180+
clear to implementors that I/O would block the loop.
181+
176182
* BREAKING CHANGE: Ensure listeners do not miss initial packets if Engine starts too quickly (#387) @bdraco
177183

178184
When manually creating a zeroconf.Engine object, it is no longer started automatically.
@@ -231,6 +237,10 @@ Changelog
231237

232238
* MAJOR BUG: Fix queries for AAAA records (#616) @bdraco
233239

240+
* Fix ServiceInfo with multiple A records (#725) @bdraco
241+
242+
* Synchronize time for fate sharing (#718) @bdraco
243+
234244
* Cleanup typing in zero._core and document ignores (#714) @bdraco
235245

236246
* Cleanup typing in zeroconf._logger (#715) @bdraco

0 commit comments

Comments
 (0)