]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Arran Cudbard-Bell [Thu, 27 Jan 2022 17:23:16 +0000 (11:23 -0600)]
Fix ODR violation
Terry Burton [Thu, 27 Jan 2022 15:07:08 +0000 (15:07 +0000)]
eap-tls test: Builds using clang analyzers are slow (#4381)
Arran Cudbard-Bell [Thu, 27 Jan 2022 14:07:37 +0000 (08:07 -0600)]
Maybe the logging bio is stuck in an infinite loop?
Given we can only reproduce this issue in CI, this is the only way to check...
Arran Cudbard-Bell [Wed, 26 Jan 2022 19:33:42 +0000 (13:33 -0600)]
not a string
Terry Burton [Thu, 27 Jan 2022 14:00:10 +0000 (14:00 +0000)]
Scheduled fuzzing: Typo (#4380)
Alan T. DeKok [Wed, 26 Jan 2022 19:36:19 +0000 (14:36 -0500)]
add target-specific rules, and a "help" line
Arran Cudbard-Bell [Wed, 26 Jan 2022 19:29:49 +0000 (13:29 -0600)]
Minor documentation fixes. Closes #4301
Jorge Pereira [Wed, 26 Jan 2022 19:28:17 +0000 (16:28 -0300)]
rlm_python: Strip out -ffat-lto-objects (#4373)
Arran Cudbard-Bell [Wed, 26 Jan 2022 19:25:19 +0000 (13:25 -0600)]
Same deal with fr_value_box_memdup
James Jones [Wed, 26 Jan 2022 19:04:27 +0000 (13:04 -0600)]
Changes to make test_requeue_on_reconnect() work (#4372)
* Can't free preq until the end along with everything else, since
otherwise, a later demux will look at it at least to validate
talloc type and thus do a use after free, so we don't free
separately and give it the appropriate talloc context
* Setting preq->treq after fr_trunk_request_enqueue() was missing.
* In some cases, it takes longer for the trunk request to make it
to a stop state than the test author expected, because the trunk
is given a request_cancel_mux function. If we should instead be
hurrying test time along faster, we can make that change.
Arran Cudbard-Bell [Wed, 26 Jan 2022 18:55:37 +0000 (12:55 -0600)]
Explicitly document and test for NULL input case.
Arran Cudbard-Bell [Wed, 26 Jan 2022 18:50:16 +0000 (12:50 -0600)]
String input for fr_value_box_bstrndup maye be NULL
Arran Cudbard-Bell [Wed, 26 Jan 2022 17:15:50 +0000 (11:15 -0600)]
Check the result of fr_sbuff_out_aunescape_until correctly
Arran Cudbard-Bell [Wed, 26 Jan 2022 15:27:56 +0000 (09:27 -0600)]
Print functions return signed values
Terry Burton [Wed, 26 Jan 2022 14:47:47 +0000 (14:47 +0000)]
Scheduled fuzzing: Clone with greater depth to find last corpus update (#4377)
Arran Cudbard-Bell [Wed, 26 Jan 2022 14:45:25 +0000 (08:45 -0600)]
Fix more warnings
Arran Cudbard-Bell [Wed, 26 Jan 2022 01:24:53 +0000 (19:24 -0600)]
Mark more box functions up as non-null
Arran Cudbard-Bell [Wed, 26 Jan 2022 00:33:50 +0000 (18:33 -0600)]
More scan fixes
Arran Cudbard-Bell [Tue, 25 Jan 2022 23:19:28 +0000 (17:19 -0600)]
Print the error...
Arran Cudbard-Bell [Tue, 25 Jan 2022 23:11:03 +0000 (17:11 -0600)]
Remove bad asserts
Arran Cudbard-Bell [Tue, 25 Jan 2022 23:01:56 +0000 (17:01 -0600)]
Typo
Arran Cudbard-Bell [Tue, 25 Jan 2022 22:59:33 +0000 (16:59 -0600)]
Stupid compiler
Arran Cudbard-Bell [Tue, 25 Jan 2022 22:54:48 +0000 (16:54 -0600)]
Quiet compiler
Alan T. DeKok [Tue, 25 Jan 2022 22:12:09 +0000 (17:12 -0500)]
let's try to shut up clang scan even more
it would be greatly appreciated if the same version of clang
produced the same errors on different platforms.
Arran Cudbard-Bell [Tue, 25 Jan 2022 21:56:18 +0000 (15:56 -0600)]
Fix more build errors in sigtran
Arran Cudbard-Bell [Tue, 25 Jan 2022 15:49:05 +0000 (09:49 -0600)]
Break out MEMCMP_RETURN into an inline function
makes debugging easier
Alan T. DeKok [Tue, 25 Jan 2022 21:22:52 +0000 (16:22 -0500)]
try to quiet clang scan
we might not want to mark up each function with CC_HINT(nonnull),
but it might be worth it.
Alan T. DeKok [Tue, 25 Jan 2022 19:48:05 +0000 (14:48 -0500)]
allow for casts, even if we have type hints on input
and automatically upcase ipv4 / ipv6 addresses when doing network
comparisons
Alan T. DeKok [Tue, 25 Jan 2022 19:25:19 +0000 (14:25 -0500)]
parse enum names when they're known
until such time as the rest of the parser is fixed.
Alan T. DeKok [Tue, 25 Jan 2022 19:16:54 +0000 (14:16 -0500)]
Revert "remove unused function"
This reverts commit
8a5c44eaf76c5184874a0c63e330b33fbb226ab8 .
Alan T. DeKok [Tue, 25 Jan 2022 19:15:19 +0000 (14:15 -0500)]
typo
Alan T. DeKok [Tue, 25 Jan 2022 15:21:30 +0000 (10:21 -0500)]
re-enable casts in all expressions
Alan T. DeKok [Tue, 25 Jan 2022 15:20:51 +0000 (10:20 -0500)]
remove unused function
Alan T. DeKok [Tue, 25 Jan 2022 15:20:39 +0000 (10:20 -0500)]
try to shut up clang
Alan T. DeKok [Tue, 25 Jan 2022 15:20:29 +0000 (10:20 -0500)]
try to shut up clang
Alan T. DeKok [Tue, 25 Jan 2022 14:05:29 +0000 (09:05 -0500)]
add "has_pure_children" flag
so that when we're purifying functions, if the current node is not
pure, AND the node has no pure children, then we can short-circuit
the walk, and ignore this node.
Alan T. DeKok [Tue, 25 Jan 2022 14:03:40 +0000 (09:03 -0500)]
make xlat_expr use new tmpl functions
along with other cleanups, which removed some bad functionality
in preparation for rewrites. This means temporarily removing some
tests, too
Arran Cudbard-Bell [Tue, 25 Jan 2022 13:37:16 +0000 (07:37 -0600)]
Process casts during tokenization for single and double quoted strings
Arran Cudbard-Bell [Tue, 25 Jan 2022 13:27:36 +0000 (07:27 -0600)]
Explicitly print casts for both sides of the map
Arran Cudbard-Bell [Tue, 25 Jan 2022 01:22:04 +0000 (19:22 -0600)]
one more...
Alan T. DeKok [Tue, 25 Jan 2022 00:00:48 +0000 (19:00 -0500)]
return the value we parsed
Alan T. DeKok [Mon, 24 Jan 2022 23:59:51 +0000 (18:59 -0500)]
use the correct sbuff
Alan T. DeKok [Mon, 24 Jan 2022 23:45:54 +0000 (18:45 -0500)]
typo
Alan T. DeKok [Mon, 24 Jan 2022 23:42:02 +0000 (18:42 -0500)]
correct spelling mistake. Fixes #4374
Arran Cudbard-Bell [Mon, 24 Jan 2022 22:41:25 +0000 (16:41 -0600)]
Try attribute references and xlats first... oops
Arran Cudbard-Bell [Mon, 24 Jan 2022 22:39:56 +0000 (16:39 -0600)]
Add support for explicit casts for barewords
Arran Cudbard-Bell [Mon, 24 Jan 2022 22:20:54 +0000 (16:20 -0600)]
Use more type specific type printing parsing functions
Alan T. DeKok [Mon, 24 Jan 2022 22:02:07 +0000 (17:02 -0500)]
handle the cast of XLAT_TMPL being data
Alan T. DeKok [Mon, 24 Jan 2022 19:30:11 +0000 (14:30 -0500)]
allow (foo) in a cast, too.
Alan T. DeKok [Sun, 23 Jan 2022 14:02:16 +0000 (09:02 -0500)]
notes for DHCPv4 encodings
Arran Cudbard-Bell [Mon, 24 Jan 2022 22:01:21 +0000 (16:01 -0600)]
Add fr_type_from_str and fr_type_to_str
Which more than halves the length of the function call...
Arran Cudbard-Bell [Mon, 24 Jan 2022 21:26:53 +0000 (15:26 -0600)]
.git/HEAD doesn't exist for submodules
Arran Cudbard-Bell [Mon, 24 Jan 2022 21:25:41 +0000 (15:25 -0600)]
Apply validation to enum value names
Arran Cudbard-Bell [Mon, 24 Jan 2022 21:25:23 +0000 (15:25 -0600)]
Move sbuff talloc wrappers to using fr_slen_t
If something breaks then we'll need to fix the wrapped functions...
Arran Cudbard-Bell [Mon, 24 Jan 2022 20:14:34 +0000 (14:14 -0600)]
Fix build issue in rlm_sigtran
Seems like it was missed in the module call refactoring
Arran Cudbard-Bell [Mon, 24 Jan 2022 18:20:34 +0000 (12:20 -0600)]
Use new libkqueue EVFILT_LIBKQUEUE version interface
This lets us retrieve the real version at runtime.
Alan T. DeKok [Sun, 23 Jan 2022 13:53:38 +0000 (08:53 -0500)]
notes for option 119
Alan T. DeKok [Sat, 22 Jan 2022 13:48:28 +0000 (08:48 -0500)]
better handling of casts
Alan T. DeKok [Fri, 21 Jan 2022 15:49:03 +0000 (10:49 -0500)]
allow for non-socket listeners
which have no app_io->open, but do have app_io->event_list_set
Alan T. DeKok [Thu, 20 Jan 2022 21:25:57 +0000 (16:25 -0500)]
have only one return point from tokenize_expression()
which may make it easier to groupify things.
Alan T. DeKok [Thu, 20 Jan 2022 21:23:21 +0000 (16:23 -0500)]
groupify arguments to functions
for now, we create the arguments and then groupify them if necessary.
The alternative is to always create grouping nodes, and then
remove / reparent them if they're not necessary.
Alan T. DeKok [Thu, 20 Jan 2022 20:50:36 +0000 (15:50 -0500)]
skip work if we're already in the network thread
Nick Porter [Fri, 21 Jan 2022 14:49:55 +0000 (14:49 +0000)]
Remove Ubuntu 18.04 from build matrix (#4366)
Ubuntu 18.04 uses OpenSSL 1.1.1 which lacks the OPENSSL_INIT_NO_ATEXIT
option used to control cleanup sequence.
Also, 18.04 is likely to be end of life by the time FreeRADIUS v4 is
production ready.
Nick Porter [Fri, 21 Jan 2022 14:04:13 +0000 (14:04 +0000)]
Apply tmpl rules split to rlm_mruby (#4365)
Alan T. DeKok [Thu, 20 Jan 2022 17:30:54 +0000 (12:30 -0500)]
remove hacks
Alan T. DeKok [Thu, 20 Jan 2022 17:29:55 +0000 (12:29 -0500)]
all children are groups
don't bother doing casts here, as it's not necessary
Alan T. DeKok [Thu, 20 Jan 2022 16:34:14 +0000 (11:34 -0500)]
no need to pass "func" to xlat_purify_expr()
Alan T. DeKok [Thu, 20 Jan 2022 16:14:23 +0000 (11:14 -0500)]
more upcast fixes, and fixes for left / right shift
Alan T. DeKok [Thu, 20 Jan 2022 15:48:07 +0000 (10:48 -0500)]
one more attempt to shut up clang
Alan T. DeKok [Thu, 20 Jan 2022 14:12:17 +0000 (09:12 -0500)]
convert code to being table driven
Alan T. DeKok [Thu, 20 Jan 2022 14:09:54 +0000 (09:09 -0500)]
initialize the output box only as the last step before returning
Alan T. DeKok [Thu, 20 Jan 2022 13:59:57 +0000 (08:59 -0500)]
add upcasting from string/octets for operations, too
Alan T. DeKok [Thu, 20 Jan 2022 13:53:06 +0000 (08:53 -0500)]
add more casting for booleans
Alan T. DeKok [Thu, 20 Jan 2022 00:08:35 +0000 (19:08 -0500)]
add xlat_purify and tests
which for now only handle xlat_expr, but that's a good enough
test to catch things which had previously been missed.
Alan T. DeKok [Thu, 20 Jan 2022 00:01:33 +0000 (19:01 -0500)]
add automatic casting for comparisons, too.
because fr_value_box_cmp() only handles the case where the types
on LHS and RHS are exactly the same. As a result, we have to do
various kinds of promotion here.
Alan T. DeKok [Wed, 19 Jan 2022 20:41:36 +0000 (15:41 -0500)]
use the correct variable
Alan T. DeKok [Wed, 19 Jan 2022 20:39:38 +0000 (15:39 -0500)]
first pass of xlat_purify_expr()
which only works on expressions and comparisons, and then only
some of the time. And then only if it's asked to.
Alan T. DeKok [Wed, 19 Jan 2022 19:35:12 +0000 (14:35 -0500)]
fix for new tmpl_rules_t
Arran Cudbard-Bell [Wed, 19 Jan 2022 19:28:39 +0000 (14:28 -0500)]
Need to gether up more events before servicing the I/O callbacks
Arran Cudbard-Bell [Wed, 19 Jan 2022 19:23:58 +0000 (14:23 -0500)]
Reduce spurious rebalancing in the trunk
Arran Cudbard-Bell [Wed, 19 Jan 2022 19:14:26 +0000 (14:14 -0500)]
Don't crash if the first atexit handler is a global atexit handler
Alan T. DeKok [Wed, 19 Jan 2022 16:28:07 +0000 (11:28 -0500)]
add XOR
Alan T. DeKok [Wed, 19 Jan 2022 16:27:04 +0000 (11:27 -0500)]
more notes
Alan T. DeKok [Wed, 19 Jan 2022 16:23:33 +0000 (11:23 -0500)]
typo
Alan T. DeKok [Wed, 19 Jan 2022 16:18:16 +0000 (11:18 -0500)]
typo
Alan T. DeKok [Wed, 19 Jan 2022 16:14:00 +0000 (11:14 -0500)]
shut up clang analyzer
Alan T. DeKok [Wed, 19 Jan 2022 16:11:00 +0000 (11:11 -0500)]
remove string/octets prepend.
It's not really needed, and it interferes with the ability to
add XOR
Alan T. DeKok [Wed, 19 Jan 2022 15:48:24 +0000 (10:48 -0500)]
lots more notes
Arran Cudbard-Bell [Wed, 19 Jan 2022 16:02:11 +0000 (11:02 -0500)]
Remove duplicate definition
Arran Cudbard-Bell [Wed, 19 Jan 2022 15:20:01 +0000 (10:20 -0500)]
Split tmpl rules
Alan T. DeKok [Wed, 19 Jan 2022 15:06:30 +0000 (10:06 -0500)]
shut up clang, if possible
Alan T. DeKok [Wed, 19 Jan 2022 14:28:32 +0000 (09:28 -0500)]
rename XLAT_ATTRIBUTE to XLAT_TMPL
in preparation for making RHS of expressions into TMPLs, too
Alan T. DeKok [Wed, 19 Jan 2022 13:53:28 +0000 (08:53 -0500)]
quiet compiler
Alan T. DeKok [Wed, 19 Jan 2022 13:31:40 +0000 (08:31 -0500)]
various compiler fixes
Alan T. DeKok [Tue, 18 Jan 2022 22:23:24 +0000 (17:23 -0500)]
prototype
Alan T. DeKok [Tue, 18 Jan 2022 22:00:31 +0000 (17:00 -0500)]
let's handle -1, m'kay?
Alan T. DeKok [Tue, 18 Jan 2022 21:56:13 +0000 (16:56 -0500)]
add xlat expressions
currently not used anywhere other than tests.
Next steps are to change XLAT_ATTRIBUTE to XLAT_TMPL, so that
we can have enums, regexes, and unparsed strings in expressions
Alan T. DeKok [Tue, 18 Jan 2022 21:45:14 +0000 (16:45 -0500)]
minor rearrangement
Alan T. DeKok [Tue, 18 Jan 2022 21:44:47 +0000 (16:44 -0500)]
we need &Cast-Base for printing enum names for fr_type_t
Alan T. DeKok [Tue, 18 Jan 2022 21:43:48 +0000 (16:43 -0500)]
expose more functions in xlat_priv.h
in preparation for xlat expressions