]>
git.ipfire.org Git - thirdparty/curl.git/log
Daniel Stenberg [Fri, 17 Oct 2025 07:58:20 +0000 (09:58 +0200)]
imap: treat capabilities case insensitively
Reported-by: Joshua Rogers
Fixes #19089
Closes #19090
Daniel Stenberg [Fri, 17 Oct 2025 07:31:55 +0000 (09:31 +0200)]
examples/websocket: fix use of uninitialized rlen
Pointed out by ZeroPath
Closes #19088
Daniel Stenberg [Thu, 16 Oct 2025 19:47:42 +0000 (21:47 +0200)]
openldap: limit max incoming size
Set the maximum allowed size of an incoming LDAP message, which to
OpenLDAP means that it allows malloc() up to this size. If not set,
there is no limit and we instead risk a malloc() failure.
The limit is arbitrarily set to 256K as I can't figure out what a
reasonable value should be.
OpenLDAP docs: https://openldap.org/software/man.cgi?query=lber-sockbuf&apropos=0&sektion=0&manpath=OpenLDAP+2.6-Release&arch=default&format=html
Bug: https://issues.oss-fuzz.com/issues/
432441303
Closes #19087
Viktor Szakats [Wed, 15 Oct 2025 17:46:35 +0000 (19:46 +0200)]
firefox-db2pem.sh: add macOS support, tidy-ups
Cherry-picked from #19076
Closes #19086
Viktor Szakats [Wed, 15 Oct 2025 17:45:48 +0000 (19:45 +0200)]
scripts: pass `--` before passing xargs
Also:
- GHA/checkdocs: escape `.` in -E regex expression.
Closes #19076
Viktor Szakats [Wed, 15 Oct 2025 18:22:20 +0000 (20:22 +0200)]
GHA: sync up `curl -V` step descriptions
Also to make it easier to recognize.
Also:
- GHA/linux-old: split steps to match other jobs.
- GHA: add `--disable` where missing.
Closes #19084
Viktor Szakats [Thu, 16 Oct 2025 16:10:41 +0000 (18:10 +0200)]
GHA/windows: stop installing Perl `Win32-Process*` modules
It's complex and did not help stabilizing CI runs.
Hard to say, but I'm suspicious it's related to the CI errors
-
1073741502 , 0xC0000142, seen in the 'build examples' and
'disk space used' steps.
Ref: #18526
Reverts
52775a7fb4ba63d66d60067dea4a5293fb7c55a1 #18296
Closes #19083
Viktor Szakats [Thu, 16 Oct 2025 14:29:56 +0000 (16:29 +0200)]
GHA: bump LibreSSL to 4.2.0
Also move back URLs to GitHub, sources are available there again.
Ref: https://github.com/libressl/portable/releases/tag/v4.2.0
Ref: https://ftp.openbsd.org/pub/OpenBSD/LibreSSL/libressl-4.2.0-relnotes.txt
Ref: #19050
Ref: #19081
Closes #19082
Viktor Szakats [Wed, 15 Oct 2025 13:06:08 +0000 (15:06 +0200)]
version: add GSS backend name and version
MIT Kerberos version detection is implemented for autotools and cmake.
Examples:
```
curl 8.17.0-DEV (x86_64-pc-linux-gnu) ... mbedTLS/3.6.4 libidn2/2.3.7 nghttp2/1.59.0 libgss/1.0.4 OpenLDAP/2.6.7
curl 8.17.0-DEV (x86_64-pc-linux-gnu) ... LibreSSL/4.1.1 libidn2/2.3.7 nghttp2/1.59.0 mit-krb5/1.20.1 OpenLDAP/2.6.7
curl 8.17.0-DEV (x86_64-pc-linux-gnu) ... LibreSSL/4.1.1 libidn2/2.3.7 nghttp2/1.59.0 mit-krb5 OpenLDAP/2.6.7
curl 8.17.0-DEV (x86_64-pc-linux-gnu) ... LibreSSL/4.1.1 nghttp2/1.59.0 mit-krb5/1.20.1 OpenLDAP/2.6.7
curl 8.17.0-DEV (aarch64e-apple-darwin24.6.0) ... GnuTLS/3.8.10 libidn2/2.3.8 libssh2/1.11.1 nghttp2/1.67.1 mit-krb5/1.22.1
```
Also:
- cmake/FindGSS: strip project name ("Kerberos 5 release") from
the version string when detected via `krb5-config`.
Closes #19073
Viktor Szakats [Thu, 16 Oct 2025 03:33:12 +0000 (05:33 +0200)]
autotools: drop detection of ancient OpenSSL libs `RSAglue` and `rsaref`
Closes #19078
Stefan Eissing [Tue, 14 Oct 2025 10:13:24 +0000 (12:13 +0200)]
apple sectrust: add to features
It should be visible in the feature list that libcurl is build with
Apple SecTrust enabled.
Closes #19057
Daniel Stenberg [Thu, 16 Oct 2025 07:01:17 +0000 (09:01 +0200)]
RELEASE-NOTES: synced
Daniel Stenberg [Thu, 16 Oct 2025 06:32:57 +0000 (08:32 +0200)]
connect: remove redundant condition in shutdown start
Pointed out by CodeSonar
Closes #19079
Daniel Stenberg [Wed, 15 Oct 2025 08:56:47 +0000 (10:56 +0200)]
urldata: make redirect counter 16 bit
Instead of long (up to 64-bit) as the maximum allowed value set since
b059f7deaf3 is 0x7fff. Saves 2 or 6 bytes.
Closes #19072
Daniel Stenberg [Wed, 15 Oct 2025 08:48:42 +0000 (10:48 +0200)]
urldata: make 'retrycount' a single byte
Since it only counts up to 5
Closes #19071
Daniel Stenberg [Tue, 14 Oct 2025 12:43:46 +0000 (14:43 +0200)]
lib: stop NULL-checking conn->passwd and ->user
They always point to a string. The string might be zero length.
Closes #19059
Emre Çalışkan [Tue, 14 Oct 2025 18:35:54 +0000 (21:35 +0300)]
transfer: reset retry count on each request
Reported-by: plv1313 on github
Fixes #18926
Closes #19066
Stefan Eissing [Tue, 14 Oct 2025 13:53:37 +0000 (15:53 +0200)]
quic: remove data_idle handling
The transfer loop used to check the socket and if no poll events
were seen, triggered a "DATA_IDLE" event into the filters to let
them schedule times/do things anyway.
Since we no longer check the socket, the filters have been called
already and the DATA_IDLE event is unnecessary work. Remove it.
Closes #19060
Daniel Stenberg [Wed, 15 Oct 2025 06:42:20 +0000 (08:42 +0200)]
tftp: check for trailing ";mode=" in URL without strstr
RFC 3617 defines two specific modes, "netascii" and "octet". This code
now checks only for those trailing ones - and not in the hostname since
they can't be there anymore.
Assisted-by: Jay Satiro
Closes #19070
Daniel Stenberg [Wed, 15 Oct 2025 06:27:48 +0000 (08:27 +0200)]
ftp: replace strstr() in ;type= handling
Since it needs to be a trailing piece of the path avoiding strstr() is
faster and more reliable.
Also stopped checking the host name since it cannot actually be there
since quite a long while back. The URL parser doesn't allow such a
hostname.
Moved the check into its own subfunction too.
Closes #19069
Daniel Stenberg [Tue, 14 Oct 2025 15:40:18 +0000 (17:40 +0200)]
lib: reduce use of data->conn->
If there are more than two of them in a function, use a local 'conn'
variable instead.
Closes #19063
Daniel Stenberg [Tue, 14 Oct 2025 16:08:27 +0000 (18:08 +0200)]
http: look for trailing 'type=' in ftp:// without strstr
- it could find a wrong string
- this is faster
Closes #19065
Jay Satiro [Sun, 12 Oct 2025 22:35:22 +0000 (18:35 -0400)]
ws: fix type conversion check
- Fix logic that checks whether a size_t will fit in a curl_off_t.
Reported-by: Viktor Szakats
Fixes https://github.com/curl/curl/issues/19017
Closes https://github.com/curl/curl/pull/19036
Jay Satiro [Fri, 10 Oct 2025 19:42:27 +0000 (15:42 -0400)]
tool_operate: retry on HTTP response codes 522 and 524
- Treat HTTP response codes 522 and 524 as a transient error since
Cloudflare may use them instead of 504 to signal timeout.
For example here is a 522 error message from Cloudflare:
"The initial connection between Cloudflare's network and the origin web
server timed out. As a result, the web page can not be displayed."
Prior to this change the curl tool did not retry on HTTP response codes
522 and 524 when --retry was used.
Fixes https://github.com/curl/curl/discussions/16143
Closes https://github.com/curl/curl/pull/19011
Viktor Szakats [Mon, 13 Oct 2025 20:57:01 +0000 (22:57 +0200)]
examples: check more errors, fix cleanups, scope variables
Inspired by Joshua's report on examples.
Closes #19055
Daniel Stenberg [Tue, 14 Oct 2025 08:49:53 +0000 (10:49 +0200)]
openldap: explain a const removing typecast
Closes #19056
Stefan Eissing [Tue, 14 Oct 2025 12:39:50 +0000 (14:39 +0200)]
scorecard: add perf support on linux
When calling scorecard with --flame to produce a flamegraph, use
"perf" on linux platforms to do the measurements. Update the scorecard
documentation about it.
Closes #19058
Viktor Szakats [Mon, 13 Oct 2025 14:30:18 +0000 (16:30 +0200)]
examples: improve global init, error checks and returning errors
- add `curl_global_init()` and `curl_global_cleanup()` where missing.
- check the result of `curl_global_init()` where missing.
- return the last curl error from `main()`.
- drop Win32-specific socket initialization in favor of `curl_global_init()`.
- rename some outliers to `res` for curl result code.
- fix cleanup in some error cases.
Inspired by Joshua's report on examples.
Closes #19053
Viktor Szakats [Mon, 13 Oct 2025 14:18:40 +0000 (16:18 +0200)]
examples: return `curl_easy_perform()` results
Where missing. Or explicitly `(void)` it where we ignore it on purpose.
Reported-by: Joshua Rogers (for `sepheaders.c`)
Closes #19052
Viktor Szakats [Mon, 13 Oct 2025 14:08:42 +0000 (16:08 +0200)]
examples: call `curl_global_cleanup()` where missing
Reported-by: Joshua Rogers (for `sepheaders.c`)
Closes #19051
Viktor Szakats [Mon, 13 Oct 2025 11:44:23 +0000 (13:44 +0200)]
checksrc: allow disabling warnings on FIXME/TODO comments
Follow-up to
71ace9f3c16a434385fc27b3e8bffb52deb6ccd1
Closes #19048
Daniel Stenberg [Mon, 13 Oct 2025 14:03:46 +0000 (16:03 +0200)]
RELEASE-NOTES: synced
Daniel Stenberg [Mon, 13 Oct 2025 07:40:38 +0000 (09:40 +0200)]
libssh2: avoid risking using an uninitialized local struct field
Reported-by: Joshua Rogers
Closes #19043
Daniel Stenberg [Mon, 13 Oct 2025 07:45:28 +0000 (09:45 +0200)]
libssh: catch a resume point larger than the size
As it would otherwise trigger broken math
Reported-by: Joshua Rogers
Closes #19044
Viktor Szakats [Sat, 11 Oct 2025 08:48:23 +0000 (10:48 +0200)]
REUSE: move copyright headers to `.checksrc`
To make it simpler to move them around, create and delete them without
syncing with `REUSE.toml`.
Also:
- checksrc: allow empty lines in `.checksrc`.
- comment on why curl printfs are disallowed in examples.
Closes #19024
Daniel Stenberg [Mon, 13 Oct 2025 07:36:08 +0000 (09:36 +0200)]
libssh2: fix EAGAIN return in ssh_state_auth_agent
Reported-by: Joshua Rogers
Closes #19042
Daniel Stenberg [Mon, 13 Oct 2025 07:24:57 +0000 (09:24 +0200)]
socks_sspi: remove the enforced mode clearing
Reported-by: Joshua Rogers
Closes #19040
Daniel Stenberg [Mon, 13 Oct 2025 06:41:38 +0000 (08:41 +0200)]
pop3: function could get the ->transfer field wrong
In pop3_perform(), pop3->transfer was derived from the old
data->req.no_body. Then, pop3_perform_command() re-computed
data->req.no_body.
Now we instead call pop3_perform_command() first.
Reported-by: Joshua Rogers
Closes #19039
Daniel Stenberg [Sun, 12 Oct 2025 13:58:43 +0000 (15:58 +0200)]
tool_getparam: make --fail and --fail-with-body override each other
This allows users to put one of them in their .curlrc and still easily
use the other one at will in command lines.
The --no-* versions disable both of them.
Reported-by: Mitchell Blank Jr
Fixes #19029
Closes #19034
Daniel Stenberg [Sun, 12 Oct 2025 21:48:14 +0000 (23:48 +0200)]
tool_msgs: make errorf() show if --show-error
Assisted-by: Mitchell Blank Jr
Ref: #19029
Closes #19035
Daniel Stenberg [Mon, 13 Oct 2025 06:27:01 +0000 (08:27 +0200)]
KNOWN_BUGS: We do not support auth-int for Digest using PUT or POST
Closes #19038
Daniel Stenberg [Sun, 12 Oct 2025 09:38:39 +0000 (11:38 +0200)]
tftp: error requests for blank filenames
Reported-by: Joshua Rogers
Closes #19033
Daniel Stenberg [Sun, 12 Oct 2025 09:24:07 +0000 (11:24 +0200)]
examples/synctime: fix null termination assumptions
bonus: dont parse argv[0] for options
Reported-by: Joshua Rogers
Closes #19032
Daniel Stenberg [Sun, 12 Oct 2025 09:15:08 +0000 (11:15 +0200)]
sws: pass in socket reference to allow function to close it
The function service_connection() now passes in a reference to the
socket instead of by value since the sub function http_connect() might
close it and set *infdp = CURL_SOCKET_BAD. This would previously not be
detected when service_connection() returned and potentially cause a
double close of the socket.
Reported-by: Joshua Rogers
Closes #19031
Daniel Stenberg [Sun, 12 Oct 2025 09:05:46 +0000 (11:05 +0200)]
libssh/libssh2: reject quote command lines with too much data
If there is lingering letters left on the right side after the paths
have been parsed, they are syntactically incorrect so returning error is
the safe thing to do.
Reported-by: Harry Sintonen
Closes #19030
Daniel Stenberg [Sat, 11 Oct 2025 20:44:30 +0000 (22:44 +0200)]
docs: expand on quoting rules for file names in SFTP quote
Reported-by: Harry Sintonen
Closes #19025
Stefan Eissing [Fri, 10 Oct 2025 12:33:36 +0000 (14:33 +0200)]
lib: SSL connection reuse
Protocol handlers not flagging PROTOPT_SSL that allow reuse of existing
SSL connections now need to carry the flag PROTOPT_SSL_REUSE.
Add PROTOPT_SSL_REUSE to imap, ldap, pop3, smtp and ftp.
Add tests the http: urls do not reuse https: connections and vice versa.
Reported-by: Sakthi SK
Fixes #19006
Closes #19007
Stefan Eissing [Fri, 10 Oct 2025 10:00:29 +0000 (12:00 +0200)]
libssh2: use sockindex consistently
Although the protocol should only run on index 0, there was a mix of
looked up sockindex and using constant 0 in tls send/recv.
Reported-by: Joshua Rogers
Closes #19004
Stefan Eissing [Fri, 10 Oct 2025 09:28:29 +0000 (11:28 +0200)]
smb: transfer debugassert to real check
That also works for non-debug builds.
Reported-by: Joshua Rogers
Cloes #19003
Daniel Stenberg [Sat, 11 Oct 2025 21:34:45 +0000 (23:34 +0200)]
lib1514: fix return code mixup
Reported-by: Joshua Rogers
Closes #19027
Daniel Stenberg [Sat, 11 Oct 2025 20:54:54 +0000 (22:54 +0200)]
socksd: remove --bindonly mention, there is no such option
Reported-by: Joshua Rogers
Closes #19026
Daniel Stenberg [Fri, 10 Oct 2025 21:54:15 +0000 (23:54 +0200)]
asyn-ares: remove wrong comment about the callback argument
Both the c-ares documentation and the c-ares source code contradict the
previous comment (and mentions/contains no such restriction).
Ref: #19001
Closes #19014
Daniel Stenberg [Fri, 10 Oct 2025 07:53:32 +0000 (09:53 +0200)]
doswin: CloseHandle the thread on shutdown
As this is in the tool shutdown the impact of it was nothing.
Also, move two global variables to local.
Follow-up to
9a2663322c330ff11275abafd612e9c
Reported-by: Joshua Rogers
Closes #18996
Daniel Stenberg [Sat, 11 Oct 2025 13:44:10 +0000 (15:44 +0200)]
RELEASE-NOTES: synced
Viktor Szakats [Fri, 10 Oct 2025 22:27:36 +0000 (00:27 +0200)]
examples/log_failed_transfers: make it build for WinCE
- include `windows.h` after `winsock2.h` via `curl/curl.h`.
- avoid `errno` for WinCE.
- avoid `_vscprintf` for WinCE.
Ref:
4535532ed36d2129b107ab357262072f82c2b34a #18843
Follow-up to
0780de2625bf8bb3bcb0f88bbbc401b2750ec1bb #18668
Closes #19016
Viktor Szakats [Fri, 10 Oct 2025 22:11:16 +0000 (00:11 +0200)]
curl_threads: delete WinCE fallback branch
Both WinCE and Windows use `CreateThread()` now, so the use of
`GetLastError()` works for both.
Follow-up to
03448f477a0cfa3868dfd15a7b9278dcecf944a2 #18998
Follow-up to
1c49f2f26d0f200bb9de61f795f06a1bc56845e9 #18451
Follow-up to
af0216251b94e751baa47146ac9609db70793b8e #1589
Closes #19015
Viktor Szakats [Fri, 10 Oct 2025 21:11:14 +0000 (23:11 +0200)]
cmake: pre-fill three more type sizes on Windows
Use `CMAKE_SIZEOF_VOID_P` to fill the size of three types that differ
on 32 and 64-bit Windows: `curl_socket_t`, `size_t`, and on mingw-w64:
`ssize_t`.
`time_t` remains the only type needing detection at configuration time,
with MSVC or mingw-w64.
Ref: https://cmake.org/cmake/help/v4.1/variable/CMAKE_SIZEOF_VOID_P.html
Closes #19013
Viktor Szakats [Fri, 10 Oct 2025 20:57:10 +0000 (22:57 +0200)]
cmake/FindGSS: simplify/de-dupe lib setup
- lib name is always `gss` with GNU GSS.
- move lib name assigments to the detection blocks.
Closes #19012
Patrick Monnerat [Fri, 10 Oct 2025 13:55:50 +0000 (15:55 +0200)]
os400: document threads handling in code.
This is to clarify threads unavaibility check and handling for security
bug busters unaware of OS400 specificities.
Fixes #18967
Closes #19009
Stefan Eissing [Fri, 10 Oct 2025 08:40:55 +0000 (10:40 +0200)]
c-ares: when resolving failed, persist error
Repeated calls to `Curl_async_is_resolved()` after a failure
returned OK and not the error that was the result of the resolve
fail.
Reported-by: Joshua Rogers
Closes #18999
Stefan Eissing [Fri, 10 Oct 2025 08:15:38 +0000 (10:15 +0200)]
thread: errno on thread creation
When thread creation fails, the code uses `errno` to remember the cause.
But pthread_create() never sets errno and gives the error as return value.
Fix that by setting the return value into errno on failure.
Windows: I think the ifdef was the wrong way around. Also set a generic
Windows Error code on CE systems.
Reported-by: Joshua Rogers
Closes #18998
Stefan Eissing [Fri, 10 Oct 2025 07:48:52 +0000 (09:48 +0200)]
conn: fix hostname move on connection reuse
When reusing a connection, the `host` and `conn_to_host` hostname
structs are moved from the template connection onto the existing one.
There was a NULLing of a tempplate member missing in `conn_to_host`
which could then lead to a double free.
Make this struct move into a static function, doing the correct
thing for both `struct hostname` in a connection.
Reported-by: Joshua Rogers
Closes #18995
Stefan Eissing [Thu, 9 Oct 2025 10:19:49 +0000 (12:19 +0200)]
cf-socket: set FD_CLOEXEC on all sockets opened
Removed TODO item
Reported-by: Joshua Rogers
Closes #18968
Stefan Eissing [Thu, 9 Oct 2025 09:23:42 +0000 (11:23 +0200)]
ws: fix some edge cases
Fix edge cases around handling of pending send frames and encoding
frames with size_t/curl_off_t possible flowy things.
Reported-by: Joshua Rogers
Closes #18965
Stefan Eissing [Thu, 9 Oct 2025 08:41:02 +0000 (10:41 +0200)]
gnutls: check conversion of peer cert chain
Check the result when converting the peer certificate chain
into gnutls internal x590 data structure for errors.
Reported-by: Joshua Rogers
Closes #18964
Stefan Eissing [Wed, 8 Oct 2025 12:29:54 +0000 (14:29 +0200)]
socks: advance iobuf instead of reset
During the SOCKS connect phase, the `iobuf` is used to receive repsonses
from the server. If the server sends more bytes than expected, the code
discarded them silently.
Fix this by advancing the iobuf only with the length consumed.
Reported-by: Joshua Rogers
Closes #18938
Viktor Szakats [Fri, 10 Oct 2025 04:43:41 +0000 (06:43 +0200)]
cmake/FindGSS: dedupe pkg-config module strings
Closes #18994
Viktor Szakats [Fri, 10 Oct 2025 04:37:45 +0000 (06:37 +0200)]
cmake/FindGSS: drop wrong header check for GNU GSS
GNU GSS offers `gss.h`; do not check for `gssapi.h`. `gssapi.h`
was originally published by Heimdal, and later MIT Kerberos also added it
for Heimdal compatibility.
Closes #18993
Viktor Szakats [Fri, 10 Oct 2025 01:09:16 +0000 (03:09 +0200)]
kerberos: bump minimum to 1.3 (2003-07-08), drop legacy logic
Previous minimum was: 1.2.4 (2002-02-28)
- assume `gssapi/gssapi.h` header for MIT Kerberos.
Drop logic detecting this header, and drop alternate logic including
a bare "gssapi.h". Bare `gssapi.h` is Heimdal-specific. MIT Kerberos
added support for it for Heimdal compatibility on 2006-11-09,
redirecting to `gssapi/gssapi.h`. MIT Kerberos supported the latter
header in the 1990s already.
Ref:
40e1a016f92903c731f07325bc1f9c6416ae1ac3 (2008-03-06)
Ref: https://github.com/krb5/krb5/commit/
d11935200186040132e05e2beaaba20a770ee3ef (2006-11-09)
- configure.ac: stop using `HAVE_GSSAPI_GSSAPI_H`.
Added in 2010 to support "ancient distros such as RHEL-3" where
`gssapi/gssapi_krb5.h` did not include `gssapi/gssapi.h`.
MIT Kerberos includes it since commit:
https://github.com/krb5/krb5/commit/
d9e959edfa8da7cab3bde96c9c4ca39beaf8db69 (2003-03-06)
Released in 1.3 (2003-07-08).
Bump minimum required version to avoid this issue.
Reverts
cca192e58f9ed7c4b33c1c991f69ff830c58b38f (2010-04-16)
Ref: https://web.mit.edu/kerberos/dist/historic.html
Ref: https://sources.debian.org/src/krb5/
Closes #18992
Viktor Szakats [Thu, 9 Oct 2025 11:34:56 +0000 (13:34 +0200)]
CI.md: refresh
Closes #18973
Viktor Szakats [Fri, 10 Oct 2025 12:37:41 +0000 (14:37 +0200)]
GHA/linux: test GNU GSS with autotools, cmake, valgrind and scan-build
The cmake build is running runtests with valgrind. The autotools one is
running scan-build.
Also:
- ignore two memleaks with GNU GSS detected by valgrind.
- add comment on support status of `GSS_C_DELEG_POLICY_FLAG`.
Closes #19008
Viktor Szakats [Fri, 10 Oct 2025 00:49:46 +0000 (02:49 +0200)]
kerberos: stop including `gssapi/gssapi_generic.h`
It's a legacy MIT Kerberos header that's no longer used by curl since:
355bf01c828af16c47ab52bccb9ade769f8bf158 (2015-01-09)
There were still mentions of it after this patch, when using versions
<1.2.3, but those versions aren't supported since:
99185417952da30c8ddd82ab962fb58da96260b2 (2008-06-12)
This header remains in use by autotools and cmake to detect MIT Kerberos
(vs. Heimdal, which doesn't have it.)
Ref: https://github.com/curl/curl/pull/18978#issuecomment-
3387414995
Closes #18990
Daniel Stenberg [Thu, 9 Oct 2025 21:21:37 +0000 (23:21 +0200)]
libssh: make atime and mtime cap the timestamp instead of wrap
The libssh API uses a 32 bit type for datestamp, so instead of just
force-typecast it, make sure it gets capped at UINT_MAX if the value is
larger.
Reported-by: Joshua Rogers
Closes #18989
Daniel Stenberg [Thu, 9 Oct 2025 21:03:18 +0000 (23:03 +0200)]
ldap: avoid null ptr deref on failure
ldap_get_dn() can return NULL on error
Reported-by: Joshua Rogers
Closes #18988
Viktor Szakats [Thu, 9 Oct 2025 15:15:02 +0000 (17:15 +0200)]
kerberos: drop logic for MIT Kerberos <1.2.3 (pre-2002) versions
curl requires 1.2.4 or newer.
Also:
- vms: stop defining `gss_nt_service_name`. Added in
f9cf3de70b3a494f627eda6cccf6607616eaf449 , symbol not used in curl code
since
355bf01c828af16c47ab52bccb9ade769f8bf158 .
Closes #18978
Daniel Stenberg [Thu, 9 Oct 2025 20:57:29 +0000 (22:57 +0200)]
openldap: fix memory-leak on oldap_do's exit path
On SSL sockbuf setup failure in `oldap_do`, the 'lud' data would not be
freed and instead leak.
Reported-by: Joshua Rogers
Closes #18986
Daniel Stenberg [Thu, 9 Oct 2025 20:50:01 +0000 (22:50 +0200)]
openldap: fix memory-leak in error path
The 'ber' pointer could escape a free if an early error occurred.
Reported-by: Joshua Rogers
Closes #18985
Daniel Stenberg [Thu, 9 Oct 2025 20:32:09 +0000 (22:32 +0200)]
ftp: simplify the 150/126 size scanner
The file size is weirdly returned in a 150 or 126 response as "XXX
bytes" mentioned somewhere in the response string. This is a rewrite of
the size scanner to replace the strange strstr() + backwards search from
before with a plain forward search until '[number] + " bytes"' is a
match.
Triggered by a report by Joshua Rogers about the previous parser.
Closes #18984
Daniel Stenberg [Thu, 9 Oct 2025 20:10:32 +0000 (22:10 +0200)]
RELEASE-NOTES: synced
Daniel Stenberg [Thu, 9 Oct 2025 15:23:56 +0000 (17:23 +0200)]
asyn-ares: use the duped hostname pointer for all calls
In one c-ares call the passed in pointer was used and not the new
duplicated one. This is probably fine but might as well use the new
pointer as all the other calls do, which will survive longer.
Reported-by: Joshua Rogers
Closes #18980
Daniel Stenberg [Thu, 9 Oct 2025 15:17:31 +0000 (17:17 +0200)]
Curl_resolv: fix comment. 'entry' argument is not optional
Reported-by: Joshua Rogers
Closes #18979
Daniel Stenberg [Thu, 9 Oct 2025 14:51:55 +0000 (16:51 +0200)]
krb5_gssapi: fix memory leak on error path
If a non-compliant amount of bytes is received, the function would
return error without free.
Reported-by: Joshua Rogers
Closes #18976
Daniel Stenberg [Thu, 9 Oct 2025 14:35:39 +0000 (16:35 +0200)]
vauth/digest: improve the digest parser
Previously, if for example the nonce would end with "realm=" etc it
would get the wrong piece, due to the naive parser.
Reported-by: Joshua Rogers
Closes #18975
Daniel Stenberg [Thu, 9 Oct 2025 11:18:05 +0000 (13:18 +0200)]
schannel_verify: fix mem-leak in Curl_verify_host
Reported-by: Stanislav Fort
Closes #18972
Jay Satiro [Sat, 20 Sep 2025 02:12:05 +0000 (22:12 -0400)]
examples: add an example for logging failed transfers
- Add an example that demonstrates per-transfer verbose logging to
memory.
The transfer's log is written to disk only if the transfer fails.
Closes https://github.com/curl/curl/pull/18668
Jay Satiro [Fri, 18 Oct 2024 18:12:31 +0000 (14:12 -0400)]
tool_operate: keep failed partial download for retry auto-resume
- Keep data from a failed download instead of discarding it on retry in
some limited cases when we know it's ok (currently only HTTP 200/206).
Prior to this change on failed transfer the tool truncated any outfile
data written before retrying the transfer. This change adds an exception
for HTTP downloads when the user requested auto-resume, because in that
case we can keep the outfile data and resume from the new position.
Reported-by: tkzv@users.noreply.github.com
Fixes https://github.com/curl/curl/issues/18035
Closes https://github.com/curl/curl/pull/18665
Stefan Eissing [Thu, 9 Oct 2025 08:26:30 +0000 (10:26 +0200)]
openssl: add comments regarding OCSP verification
To allow future reviewers of "security" reports to more easily find out
why code is this way.
Closes #18962
Stefan Eissing [Thu, 9 Oct 2025 09:42:43 +0000 (11:42 +0200)]
memdup0: handle edge case
When length is already SIZE_MAX, fail without allocating.
Reported-by: Joshua Rogers
Closes #18966
Stefan Eissing [Thu, 9 Oct 2025 07:55:38 +0000 (09:55 +0200)]
rustls: pass the correct result to rustls_failf
Reported-by: Joshua Rogers
Closes #18961
Viktor Szakats [Thu, 9 Oct 2025 11:48:06 +0000 (13:48 +0200)]
GHA: rename config files to match pyspelling
To make it more obvious what needs to be looked at when pyspelling is
reporting an issue.
Follow-up to
95e50ad69473d8229b85478a3f2138b7e632fbe8 #18756
Closes #18974
Viktor Szakats [Thu, 9 Oct 2025 10:54:17 +0000 (12:54 +0200)]
examples: update `.gitignore`
Follow-up to
f6f62933e917b8b5c9a9394907ce4b69600214b4 #18264
Closes #18971
Viktor Szakats [Thu, 9 Oct 2025 10:51:08 +0000 (12:51 +0200)]
examples: allow `vsnprintf` again
Ref: https://github.com/curl/curl/pull/18668#issuecomment-
3383422410
Follow-up to
b12da22db1f11da51082977dc21a7edee7858911 #18866
Closes #18970
Viktor Szakats [Thu, 9 Oct 2025 10:36:43 +0000 (12:36 +0200)]
examples/sessioninfo: do not disable security
Also make it return the curl result code.
Follow-up to
df70a68984308952dcacf33d11593cb22ad80464 #18909
Closes #18969
Daniel Stenberg [Thu, 9 Oct 2025 07:07:27 +0000 (09:07 +0200)]
hostip: don't store negative resolves due unrelated errors
Like for:
- OOM
- resolver_start() returns error
- DoH has problems
Fixes #18953
Fixes #18954
Reported-by: Joshua Rogers
Closes #18958
Daniel Stenberg [Thu, 9 Oct 2025 08:15:41 +0000 (10:15 +0200)]
mk-lib1521: verify the setopt options that accept -1
Daniel Stenberg [Thu, 9 Oct 2025 07:50:20 +0000 (09:50 +0200)]
setopt: allow CURLOPT_DNS_CACHE_TIMEOUT set to -1
It is documented as valid. Regression from commit
b059f7deaf3 shipped in
8.16.0
Reported-by: Andrei Kurushin
Fixes #18959
Closes #18960
Daniel Stenberg [Tue, 7 Oct 2025 07:22:05 +0000 (09:22 +0200)]
test1582: verify the TLS channel binding cert memory leak fix
Daniel Stenberg [Tue, 7 Oct 2025 05:49:00 +0000 (07:49 +0200)]
openssl: fix peer certificate leak in channel binding
Reported-by: Stanislav Fort
Bug: https://hackerone.com/reports/
3373640
Closes #18917
Stefan Eissing [Wed, 8 Oct 2025 11:06:48 +0000 (13:06 +0200)]
http2: ingress handling edge cases
Fix some edge cases around the `data_max_bytes` handling when
processing ingress.
Reported-by: Joshua Rogers
Closes #18933
Stefan Eissing [Wed, 8 Oct 2025 10:28:14 +0000 (12:28 +0200)]
http2: cleanup pushed newhandle on fail
When nghttp2_session_set_stream_user_data() fails, clean up the
new handle.
Reported-by: Joshua Rogers
Closes #18931