summaryrefslogtreecommitdiff
path: root/CHANGES.rst
Commit message (Collapse)AuthorAgeFilesLines
* release version 2.3.42.3.4David Lord2023-05-081-1/+1
|
* preserve invalid itms-services url schemeDavid Lord2023-05-061-0/+2
|
* Do not apply max_form_parts to non-multipart dataAdrian Moennich2023-05-051-0/+2
|
* Remove uses of warnings.catch_warnings (#2690) (#2692)Jeff Dairiki2023-05-031-0/+1
|
* WWWAuthenticate.from_header handles base64 padding in tokenDavid Lord2023-05-011-2/+2
|
* Authorization.from_header handles base64 padding in tokenNicolas Simonds2023-05-011-0/+3
| | | | co-authored-by: David Lord <davidism@gmail.com>
* start version 2.3.4David Lord2023-05-011-0/+6
|
* release version 2.3.3David Lord2023-05-011-1/+1
| | | | | cherry pick 7a33be5edacfe823067a90fae5cbb19551895d9c, which is the 2.3.3 tag commit, merged into main instead of 2.3.x
* Fix the parsing of large multipart bodiespgjones2023-05-011-0/+2
| | | | | | | | | | | | | | There were two issues to fix. Firstly if a boundary couldn't be found the parser should have parsed up to the end of the buffer or last newline (whichever is earlier). However the last newline would be the first character since 082e0e5b9c01fa3178ac0153413f082616f10914 as the DATA_START state would have a buffer that starts with newline. This was fixed by changing the last_newline method to take the data to search as an argument. Secondly the parsing was slow as the shortcut search for the boundary was removed resulting in full regex matches on each iteration. Restoring the shortcut restores the previous performance.
* cookie path=/ defaultDavid Lord2023-05-011-2/+3
|
* start version 2.3.3David Lord2023-05-011-0/+6
|
* release version 2.3.22.3.2David Lord2023-04-281-1/+1
|
* cookie path=/ defaultDavid Lord2023-04-281-0/+2
|
* max_content_length only with wsgi.input_terminatedDavid Lord2023-04-281-0/+2
|
* test client parses set-cookie more accuratelyDavid Lord2023-04-271-0/+2
|
* start version 2.3.2David Lord2023-04-271-0/+6
|
* release version 2.3.12.3.1David Lord2023-04-271-1/+1
|
* Authorization and WWWAuthenticate __eq__David Lord2023-04-261-0/+2
|
* include missing pyi filesDavid Lord2023-04-261-0/+1
|
* allow more characters in unquoted cookie valuesDavid Lord2023-04-261-0/+1
|
* quote plus in urlsDavid Lord2023-04-261-0/+2
|
* start version 2.3.1David Lord2023-04-261-0/+6
|
* release version 2.3.02.3.0David Lord2023-04-251-1/+1
|
* support scryptDavid Lord2023-04-251-0/+2
| | | | deprecate plain hash methods
* fix issue linksDavid Lord2023-04-241-5/+5
|
* deprecate custom form parser content typesDavid Lord2023-04-241-0/+3
| | | | deprecate charset and errors missed in previous deprecation PR
* escape ASCII control characters in request logDavid Lord2023-04-241-0/+2
| | | | Co-authored-by: David Leadbeater, G-Research <dgl@users.noreply.github.com>
* drop support for python 3.7David Lord2023-04-201-0/+1
|
* add __future__ annotationsDavid Lord2023-04-191-0/+1
|
* fix ProfilerMiddleware stream param type hintErik Hartwig2023-04-151-0/+2
|
* no path by defaultDavid Lord2023-04-111-1/+1
|
* deprecate request and response charsetDavid Lord2023-04-111-3/+6
|
* restrict multipart charsetsDavid Lord2023-04-101-0/+2
|
* remove request.url_charsetDavid Lord2023-04-101-2/+3
|
* deprecate charset in routingDavid Lord2023-04-101-0/+2
|
* deprecate bytes where str is expectedDavid Lord2023-04-101-0/+6
|
* Correctly implement multipart parsing for empty fieldspgjones2023-04-101-0/+1
| | | | | | | | | | | | | According to RFC2046 body-part := MIME-part-headers [CRLF *OCTET] Meaning there should only be a blank line if there is are body octets, hence if the field is empty it should not be present. The previous implementation assumed the blacnk line would always be present. This fixes both the decoder and encoder to ensure that empty fields are correctly decoded and encoded.
* update type annotation for Accept.best_matchJens Heinrich2023-04-031-0/+2
|
* refactor test client cookie implementationDavid Lord2023-04-031-0/+14
| | | | | | no longer use http.cookiejar add get_cookie add properties for decoded key and value
* no path by defaultDavid Lord2023-04-031-0/+1
|
* refactor parse_cookie and dump_cookieDavid Lord2023-04-031-0/+8
| | | | | | faster, based on parse_options_header refactor deprecate bytes dump_cookie allows domain without dot, strips leading dot
* more windows special filesDavid Lord2023-03-181-0/+1
|
* development server discards header keys with underscoresDavid Lord2023-03-171-0/+2
|
* check max_content_length consistentlyDavid Lord2023-03-141-0/+8
|
* refactor Authorization and WWWAuthenticateDavid Lord2023-03-101-0/+17
| | | | | | | | | add 'token' attribute to support token schemes such as Bearer don't subclass dict move parsing to from_header class methods assign to response.www_authenticate to set header assign list to set multiple header values prefer creating and assigning new WWWAuthenticate instance
* refactor parse and dump header functionsDavid Lord2023-03-091-0/+9
| | | | | | | | | deprecate some unused parameters and behaviors deprecate passing bytes handle quoting empty string dumping options skips None values unquote inline if calling the function would do some redundant work parse dict handles values with charsets
* improve parse_accept_headerDavid Lord2023-03-061-0/+2
| | | | | | follow RFC 9110 use parse_list_header and parse_options_header discard items with invalid quality
* refactor parse_options_headerDavid Lord2023-03-061-0/+3
| | | | | | | | | follow RFC 9110, discarding invalid parts 2-3 times faster disallow arbitrary encodings handle encoded quotes for HTTP headers and multipart form data Co-authored-by: L. Kärkkäinen <98187+Tronic@users.noreply.github.com>
* increase PBKDF2 default to 600,000 iterationsIllia Volochii2023-03-041-0/+1
|
* deprecate make_line_iter and make_chunk_iterDavid Lord2023-03-041-0/+1
|
View Lorry Controller Status