d4cf98bb | 06-Apr-2022 |
Daniel Stenberg |
connecache: remove duplicate connc->closure_handle check The superfluous extra check could cause analyzer false positives and doesn't serve any purpose. Closes #8676 |
ba83c522 | 05-Apr-2022 |
MAntoniak <47522782+MAntoniak@users.noreply.github.com> |
mbedtls: remove server_fd from backend Closes #8682 |
92c01235 | 05-Apr-2022 |
Tatsuhiro Tsujikawa |
ngtcp2: use token when detecting :status header field Closes #8679 |
913b80d5 | 05-Apr-2022 |
Tatsuhiro Tsujikawa |
ngtcp2: make curl 1ms faster Pass 0 for an already expired timer. Closes #8678 |
1b1443a8 | 05-Apr-2022 |
Tatsuhiro Tsujikawa |
ngtcp2: fix QUIC_IDLE_TIMEOUT QUIC_IDLE_TIMEOUT should be of type ngtcp2_duration which is nanoseconds resolution. Closes #8678 |
4a8f6869 | 05-Apr-2022 |
Daniel Stenberg |
English: use American spelling consistently Authorization, Initialization, Organization etc. Closes #8673 |
b47bda55 | 04-Apr-2022 |
Sascha Zengler |
BUGS: Fix incorrect punctuation Closes #8672 Reviewed-by: Daniel Gustafsson <daniel@yesql.se> |
f48a0ce8 | 04-Apr-2022 |
Daniel Stenberg |
tool_listhelp.c: uppercase URL |
41ea595e | 04-Apr-2022 |
Daniel Stenberg |
RELEASE-NOTES: synced |
854ec765 | 01-Apr-2022 |
Daniel Stenberg |
http: streamclose "already downloaded" Instead of connclose()ing, since when HTTP/2 is used it doesn't need to close the connection as stopping the current transfer is enough. R
http: streamclose "already downloaded" Instead of connclose()ing, since when HTTP/2 is used it doesn't need to close the connection as stopping the current transfer is enough. Reported-by: Evangelos Foutras Closes #8665
show more ...
|
9ca6cd92 | 29-Mar-2022 |
Jay Satiro |
ftp: fix error message for partial file upload - Show the count of bytes written on partial file upload. Prior to this change the error message mistakenly showed the count of by
ftp: fix error message for partial file upload - Show the count of bytes written on partial file upload. Prior to this change the error message mistakenly showed the count of bytes read, not written. Bug: https://github.com/curl/curl/discussions/8637 Reported-by: Taras Kushnir Closes https://github.com/curl/curl/pull/8649
show more ...
|
218cc700 | 01-Apr-2022 |
Daniel Stenberg |
http: correct the header error message to say colon Not semicolon Reported-by: Gisle Vanem Ref: #8666 Closes #8667 |
b716b5aa | 01-Apr-2022 |
Daniel Stenberg |
lib: #ifdef on USE_HTTP2 better ... as nghttp2 might not be the library that provides HTTP/2 support. Closes #8661 |
266627b0 | 01-Apr-2022 |
MAntoniak <47522782+MAntoniak@users.noreply.github.com> |
mbedtls: remove 'protocols' array from backend when ALPN is not used Closes #8663 |
fda4b816 | 01-Apr-2022 |
Daniel Stenberg |
http2: RST the stream if we stop it on our own will For the "simulated 304" case the done-call isn't considered "premature" but since the server didn't close the stream it needs to be re
http2: RST the stream if we stop it on our own will For the "simulated 304" case the done-call isn't considered "premature" but since the server didn't close the stream it needs to be reset to stop delivering data. Closes #8664
show more ...
|
3fa634a3 | 01-Apr-2022 |
Daniel Stenberg |
http: close the stream (not connection) on time condition abort Closes #8664 |
b5a96805 | 01-Apr-2022 |
Daniel Stenberg |
http2: handle DONE called for the paused stream As it could otherwise stall all streams on the connection Reported-by: Evangelos Foutras Fixes #8626 Closes #8664 |
7c1acaf6 | 31-Mar-2022 |
Daniel Stenberg |
tls: make mbedtls and NSS check for h2, not nghttp2 This makes them able to also negotiate HTTP/2 even when built to use hyper for h2. Closes #8656 |
0239a5fb | 31-Mar-2022 |
Daniel Stenberg |
tests/libtest/lib670.c: fixup the copyright year range follow-up to b54e18640ea4b7 |
b54e1864 | 31-Mar-2022 |
Leandro Coutinho |
lib670: avoid double check result Closes #8660 |
34ebf3f9 | 31-Mar-2022 |
Daniel Stenberg |
vtls: use a generic "ALPN, server accepted" message Closes #8657 |
55043b40 | 31-Mar-2022 |
Daniel Stenberg |
vtls: use a backend standard message for "ALPN: offers %s" I call it VTLS_INFOF_ALPN_OFFER_1STR, the '1str' meaning that the infof() call also needs a string argument: the ALPN ID.
vtls: use a backend standard message for "ALPN: offers %s" I call it VTLS_INFOF_ALPN_OFFER_1STR, the '1str' meaning that the infof() call also needs a string argument: the ALPN ID. Closes #8657
show more ...
|
62d58182 | 31-Mar-2022 |
MonkeybreadSoftware |
strcase.h: add comment about the return code Tool often we run into expecting this to work like strcmp, but it returns 1 instead of 0 for match. Closes #8658 |
3bc5b32d | 30-Mar-2022 |
Daniel Stenberg |
vtls: provide a unified APLN-disagree string for all backends Also rephrase to make it sound less dangerous: "ALPN: server did not agree on a protocol. Uses default." Repo
vtls: provide a unified APLN-disagree string for all backends Also rephrase to make it sound less dangerous: "ALPN: server did not agree on a protocol. Uses default." Reported-by: Nick Coghlan Fixes #8643 Closes #8651
show more ...
|
08ca1f2b | 30-Mar-2022 |
Daniel Stenberg |
projects/README: converted to markdown Closes #8652 |