|
6 | 6 |
|
7 | 7 | Changelog
|
8 | 8 |
|
| 9 | +Version 7.79.1 (22 Sep 2021) |
| 10 | + |
| 11 | +Daniel Stenberg (22 Sep 2021) |
| 12 | +- RELEASE-NOTES: synced |
| 13 | + |
| 14 | + curl 7.79.1 release |
| 15 | + |
| 16 | +- THANKS: added names from the 7.79.1 release |
| 17 | + |
| 18 | +- test897: verify delivery of IMAP post-body header content |
| 19 | + |
| 20 | + The "content" is delivered as "body" by curl, but the envelope continues |
| 21 | + after the body and the rest of it should be delivered as header. |
| 22 | + |
| 23 | + The IMAP server can now get 'POSTFETCH' set to include more data to |
| 24 | + include after the body and test 897 is done to verify that such "extra" |
| 25 | + header data is in fact delivered by curl as header. |
| 26 | + |
| 27 | + Ref: #7284 but fails to reproduce the issue |
| 28 | + |
| 29 | + Closes #7748 |
| 30 | + |
| 31 | +- KNOWN_BUGS: connection migration doesn't work |
| 32 | + |
| 33 | + Closes #7695 |
| 34 | + |
| 35 | +- RELEASE-NOTES: synced |
| 36 | + |
| 37 | +- http: fix the broken >3 digit response code detection |
| 38 | + |
| 39 | + When the "reason phrase" in the HTTP status line starts with a digit, |
| 40 | + that was treated as the forth response code digit and curl would claim |
| 41 | + the response to be non-compliant. |
| 42 | + |
| 43 | + Added test 1466 to verify this case. |
| 44 | + |
| 45 | + Regression brought by 5dc594e44f73b17 |
| 46 | + Reported-by: Glenn de boer |
| 47 | + Fixes #7738 |
| 48 | + Closes #7739 |
| 49 | + |
| 50 | +Jay Satiro (17 Sep 2021) |
| 51 | +- strerror: use sys_errlist instead of strerror on Windows |
| 52 | + |
| 53 | + - Change Curl_strerror to use sys_errlist[errnum] instead of strerror to |
| 54 | + retrieve the error message on Windows. |
| 55 | + |
| 56 | + Windows' strerror writes to a static buffer and is not thread-safe. |
| 57 | + |
| 58 | + Follow-up to 2f0bb86 which removed most instances of strerror in favor |
| 59 | + of calling Curl_strerror (which calls strerror_r for other platforms). |
| 60 | + |
| 61 | + Ref: https://github.com/curl/curl/pull/7685 |
| 62 | + Ref: https://github.com/curl/curl/commit/2f0bb86 |
| 63 | + |
| 64 | + Closes https://github.com/curl/curl/pull/7735 |
| 65 | + |
| 66 | +Daniel Stenberg (16 Sep 2021) |
| 67 | +- dist: provide lib/.checksrc in the tarball |
| 68 | + |
| 69 | + So that debug builds work (checksrc really) |
| 70 | + |
| 71 | + Reported-by: Marcel Raad |
| 72 | + Reported-by: tawmoto on github |
| 73 | + Fixes #7733 |
| 74 | + Closes #7734 |
| 75 | + |
| 76 | +- TODO: Improve documentation about fork safety |
| 77 | + |
| 78 | + Closes #6968 |
| 79 | + |
| 80 | +- hsts: CURLSTS_FAIL from hsts read callback should fail transfer |
| 81 | + |
| 82 | + ... and have CURLE_ABORTED_BY_CALLBACK returned. |
| 83 | + |
| 84 | + Extended test 1915 to verify. |
| 85 | + |
| 86 | + Reported-by: Jonathan Cardoso |
| 87 | + Fixes #7726 |
| 88 | + Closes #7729 |
| 89 | + |
| 90 | +- test1184: disable |
| 91 | + |
| 92 | + The test should be fine and it works for me repeated when run manually, |
| 93 | + but clearly it causes CI failures and it needs more research. |
| 94 | + |
| 95 | + Reported-by: RiderALT on github |
| 96 | + Fixes #7725 |
| 97 | + Closes #7732 |
| 98 | + |
| 99 | +- Curl_http2_setup: don't change connection data on repeat invokes |
| 100 | + |
| 101 | + Regression from 3cb8a748670ab88c (releasde in 7.79.0). That change moved |
| 102 | + transfer oriented inits to before the check but also erroneously moved a |
| 103 | + few connection oriented ones, which causes problems. |
| 104 | + |
| 105 | + Reported-by: Evangelos Foutras |
| 106 | + Fixes #7730 |
| 107 | + Closes #7731 |
| 108 | + |
| 109 | +- RELEASE-NOTES: synced |
| 110 | + |
| 111 | + and bump to 7.79.1 |
| 112 | + |
| 113 | +Kamil Dudka (16 Sep 2021) |
| 114 | +- tests/sshserver.pl: make it work with openssh-8.7p1 |
| 115 | + |
| 116 | + ... by not using options with no argument where an argument is required: |
| 117 | + |
| 118 | + === Start of file tests/log/ssh_server.log |
| 119 | + curl_sshd_config line 6: no argument after keyword "DenyGroups" |
| 120 | + curl_sshd_config line 7: no argument after keyword "AllowGroups" |
| 121 | + curl_sshd_config line 10: Deprecated option AuthorizedKeysFile2 |
| 122 | + curl_sshd_config line 29: Deprecated option KeyRegenerationInterval |
| 123 | + curl_sshd_config line 39: Deprecated option RhostsRSAAuthentication |
| 124 | + curl_sshd_config line 40: Deprecated option RSAAuthentication |
| 125 | + curl_sshd_config line 41: Deprecated option ServerKeyBits |
| 126 | + curl_sshd_config line 45: Deprecated option UseLogin |
| 127 | + curl_sshd_config line 56: no argument after keyword "AcceptEnv" |
| 128 | + curl_sshd_config: terminating, 3 bad configuration options |
| 129 | + === End of file tests/log/ssh_server.log |
| 130 | + |
| 131 | + === Start of file log/sftp_server.log |
| 132 | + curl_sftp_config line 33: Unsupported option "rhostsrsaauthentication" |
| 133 | + curl_sftp_config line 34: Unsupported option "rsaauthentication" |
| 134 | + curl_sftp_config line 52: no argument after keyword "sendenv" |
| 135 | + curl_sftp_config: terminating, 1 bad configuration options |
| 136 | + Connection closed. |
| 137 | + Connection closed |
| 138 | + === End of file log/sftp_server.log |
| 139 | + |
| 140 | + Closes #7724 |
| 141 | + |
| 142 | +Daniel Stenberg (15 Sep 2021) |
| 143 | +- hsts: handle unlimited expiry |
| 144 | + |
| 145 | + When setting a blank expire string, meaning unlimited, curl would pass |
| 146 | + TIME_T_MAX to getime_r() when creating the output, while on 64 bit |
| 147 | + systems such a large value cannot be convetered to a tm struct making |
| 148 | + curl to exit the loop with an error instead. It can't be converted |
| 149 | + because the year it would represent doesn't fit in the 'int tm_year' |
| 150 | + field! |
| 151 | + |
| 152 | + Starting now, unlimited expiry is instead handled differently by using a |
| 153 | + human readable expiry date spelled out as "unlimited" instead of trying |
| 154 | + to use a distant actual date. |
| 155 | + |
| 156 | + Test 1660 and 1915 have been updated to help verify this change. |
| 157 | + |
| 158 | + Reported-by: Jonathan Cardoso |
| 159 | + Fixes #7720 |
| 160 | + Closes #7721 |
| 161 | + |
| 162 | +- curl_multi_fdset: make FD_SET() not operate on sockets out of range |
| 163 | + |
| 164 | + The VALID_SOCK() macro was made to only check for FD_SETSIZE if curl was |
| 165 | + built to use select(), even though the curl_multi_fdset() function |
| 166 | + always and unconditionally uses FD_SET and needs the check. |
| 167 | + |
| 168 | + Reported-by: 0xee on github |
| 169 | + Fixes #7718 |
| 170 | + Closes #7719 |
| 171 | + |
| 172 | +- FAQ: add GOPHERS + curl works on data, not files |
| 173 | + |
9 | 174 | Version 7.79.0 (14 Sep 2021)
|
10 | 175 |
|
11 | 176 | Daniel Stenberg (14 Sep 2021)
|
@@ -8109,174 +8274,3 @@ Daniel Stenberg (20 Jan 2021)
|
8109 | 8274 | Closes #6473
|
8110 | 8275 |
|
8111 | 8276 | - gopher: remove accidental conn->data leftover
|
8112 |
| - |
8113 |
| -- libssh: avoid plain free() of libssh-memory |
8114 |
| - |
8115 |
| - Since curl's own memory debugging system redefines free() calls to track |
8116 |
| - and fiddle with memory, it cannot be used on memory allocated by 3rd |
8117 |
| - party libraries. |
8118 |
| - |
8119 |
| - Third party libraries SHOULD NOT require free() to release allocated |
8120 |
| - resources for this reason - and libs can use separate healp allocators |
8121 |
| - on some systems (like Windows) so free() doesn't necessarily work |
8122 |
| - anyway. |
8123 |
| - |
8124 |
| - Filed as an issue with libssh: https://bugs.libssh.org/T268 |
8125 |
| - |
8126 |
| - Closes #6481 |
8127 |
| - |
8128 |
| -- send: assert that Curl_write_plain() has a ->conn when called |
8129 |
| - |
8130 |
| - To help catch bad invokes. |
8131 |
| - |
8132 |
| - Closes #6476 |
8133 |
| - |
8134 |
| -- test410: verify HTTPS GET with a 49K request header |
8135 |
| - |
8136 |
| - skip test 410 for mesalink in the CI as it otherwise hangs "forever" |
8137 |
| - |
8138 |
| -- lib: pass in 'struct Curl_easy *' to most functions |
8139 |
| - |
8140 |
| - ... in most cases instead of 'struct connectdata *' but in some cases in |
8141 |
| - addition to. |
8142 |
| - |
8143 |
| - - We mostly operate on transfers and not connections. |
8144 |
| - |
8145 |
| - - We need the transfer handle to log, store data and more. Everything in |
8146 |
| - libcurl is driven by a transfer (the CURL * in the public API). |
8147 |
| - |
8148 |
| - - This work clarifies and separates the transfers from the connections |
8149 |
| - better. |
8150 |
| - |
8151 |
| - - We should avoid "conn->data". Since individual connections can be used |
8152 |
| - by many transfers when multiplexing, making sure that conn->data |
8153 |
| - points to the current and correct transfer at all times is difficult |
8154 |
| - and has been notoriously error-prone over the years. The goal is to |
8155 |
| - ultimately remove the conn->data pointer for this reason. |
8156 |
| - |
8157 |
| - Closes #6425 |
8158 |
| - |
8159 |
| -Emil Engler (17 Jan 2021) |
8160 |
| -- docs: fix typos in NEW-PROTOCOL.md |
8161 |
| - |
8162 |
| - This fixes a misspelled "it" and a grammatically wrong "-ing" suffix. |
8163 |
| - |
8164 |
| - Closes #6471 |
8165 |
| - |
8166 |
| -Daniel Stenberg (16 Jan 2021) |
8167 |
| -- RELEASE-NOTES: synced |
8168 |
| - |
8169 |
| -Jay Satiro (16 Jan 2021) |
8170 |
| -- [Razvan Cojocaru brought this change] |
8171 |
| - |
8172 |
| - cmake: expose CURL_DISABLE_OPENSSL_AUTO_LOAD_CONFIG |
8173 |
| - |
8174 |
| - This does for cmake builds what --disable-openssl-auto-load-config |
8175 |
| - does for autoconf builds. |
8176 |
| - |
8177 |
| - Closes https://github.com/curl/curl/pull/6435 |
8178 |
| - |
8179 |
| -Daniel Stenberg (15 Jan 2021) |
8180 |
| -- test1918: verify curl_easy_option_by_name() and curl_easy_option_by_id() |
8181 |
| - |
8182 |
| - ... and as a practical side-effect, make sure that the |
8183 |
| - Curl_easyopts_check() function is asserted in debug builds, which we |
8184 |
| - want to detect mismatches between the options list in easyoptions.c and |
8185 |
| - the options in curl.h |
8186 |
| - |
8187 |
| - Found-by: Gisle Vanem |
8188 |
| - Bug: https://github.com/curl/curl/commit/08e8455dddc5e48e58a12ade3815c01ae3da3b64#commitcomment-45991815 |
8189 |
| - |
8190 |
| - Closes #6461 |
8191 |
| - |
8192 |
| -- [Gisle Vanem brought this change] |
8193 |
| - |
8194 |
| - easyoptions: add the missing AWS_SIGV4 |
8195 |
| - |
8196 |
| - Follow-up from AWS_SIGV4 |
8197 |
| - |
8198 |
| -- schannel_verify: fix safefree call typo |
8199 |
| - |
8200 |
| - Follow-up from e87ad71d1ba00519 |
8201 |
| - |
8202 |
| - Closes #6459 |
8203 |
| - |
8204 |
| -- mime: make sure setting MIMEPOST to NULL resets properly |
8205 |
| - |
8206 |
| - ... so that a function can first use MIMEPOST and then set it to NULL to |
8207 |
| - reset it back to a blank POST. |
8208 |
| - |
8209 |
| - Added test 584 to verify the fix. |
8210 |
| - |
8211 |
| - Reported-by: Christoph M. Becker |
8212 |
| - |
8213 |
| - Fixes #6455 |
8214 |
| - Closes #6456 |
8215 |
| - |
8216 |
| -- multi: set the PRETRANSFER time-stamp when we switch to PERFORM |
8217 |
| - |
8218 |
| - ... instead of at end of the DO state. This makes the timer more |
8219 |
| - accurate for the protocols that use the DOING state (such as FTP), and |
8220 |
| - simplifies how the function (now called init_perform) is called. |
8221 |
| - |
8222 |
| - The timer will then include the entire procedure up to PERFORM - |
8223 |
| - including all instructions for getting the transfer started. |
8224 |
| - |
8225 |
| - Closes #6454 |
8226 |
| - |
8227 |
| -- CURLINFO_PRETRANSFER_TIME.3: clarify |
8228 |
| - |
8229 |
| - ... the timer *does* include the instructions for getting the remote |
8230 |
| - file. |
8231 |
| - |
8232 |
| - Ref: #6452 |
8233 |
| - Closes #6453 |
8234 |
| - |
8235 |
| -- [Gisle Vanem brought this change] |
8236 |
| - |
8237 |
| - schannel: plug a memory-leak |
8238 |
| - |
8239 |
| - ... when built without -DUNICODE. |
8240 |
| - |
8241 |
| - Closes #6457 |
8242 |
| - |
8243 |
| -Jay Satiro (14 Jan 2021) |
8244 |
| -- gitattributes: Set batch files to CRLF line endings on checkout |
8245 |
| - |
8246 |
| - If a batch file is run without CRLF line endings (ie LF-only) then |
8247 |
| - arbitrary behavior may occur. I consider that a bug in Windows, however |
8248 |
| - the effects can be serious enough (eg unintended code executed) that |
8249 |
| - we're fixing it in the repo by requiring CRLF line endings for batch |
8250 |
| - files on checkout. |
8251 |
| - |
8252 |
| - Prior to this change the checked-out line endings of batch files were |
8253 |
| - dependent on a user's git preferences. On Windows it is common for git |
8254 |
| - users to have automatic CRLF conversion enabled (core.autocrlf true), |
8255 |
| - but those users that don't would run into this behavior. |
8256 |
| - |
8257 |
| - For example a user has reported running the Visual Studio project |
8258 |
| - generator batch file (projects/generate.bat) and it looped forever. |
8259 |
| - Output showed that the Windows OS interpreter was occasionally jumping |
8260 |
| - to arbitrary points in the batch file and executing commands. This |
8261 |
| - resulted in unintended files being removed (a removal sequence called) |
8262 |
| - and looping forever. |
8263 |
| - |
8264 |
| - Ref: https://serverfault.com/q/429594 |
8265 |
| - Ref: https://stackoverflow.com/q/232651 |
8266 |
| - Ref: https://www.dostips.com/forum/viewtopic.php?t=8988 |
8267 |
| - Ref: https://git-scm.com/docs/gitattributes#_checking_out_and_checking_in |
8268 |
| - Ref: https://git-scm.com/book/en/v2/Customizing-Git-Git-Configuration#_core_autocrlf |
8269 |
| - |
8270 |
| - Bug: https://github.com/curl/curl/discussions/6427 |
8271 |
| - Reported-by: Ganesh Kamath |
8272 |
| - |
8273 |
| - Closes https://github.com/curl/curl/pull/6442 |
8274 |
| - |
8275 |
| -Daniel Stenberg (14 Jan 2021) |
8276 |
| -- tool_operate: spellfix a comment |
8277 |
| - |
8278 |
| -- ROADMAP: refreshed |
8279 |
| - |
8280 |
| - o removed HSTS - already implemented |
8281 |
| - o added HTTPS RR records |
8282 |
| - o mention HTTP/3 completion |
0 commit comments