| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
This quiets the compiler since it is not even returned anyway, and is a misleading variable name.
|
| |
|
|\
| |
| | |
Work around a bug in Npcap 1.00 in case of driver version mismatch
|
| | |
|
|\ \
| | |
| | | |
Use the intended date types for variables passed as or compared to other types
|
| | | |
|
|\ \ \
| | | |
| | | | |
build.sh: remove the check for Clang 15 and 16.
|
|/ / /
| | |
| | |
| | |
| | | |
We're working around the warning they produce for yynerrs in the parser
being set but not used.
|
|\ \ \
| |/ /
|/| | |
compiler: squelch set-but-not-used warning for yynerrs.
|
|/ / |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Valgrind does not like binaries produced by Clang 15:
$ valgrind --leak-check=full --error-exitcode=1 testprogs/findalldevstest
==23242== Memcheck, a memory error detector
==23242== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==23242== Using Valgrind-3.18.1 and LibVEX; rerun with -h for copyright info
==23242== Command: testprogs/findalldevstest
==23242==
==23242== Valgrind: debuginfo reader: ensure_valid failed:
==23242== Valgrind: during call to ML_(img_get)
==23242== Valgrind: request for range [70669806, +4) exceeds
==23242== Valgrind: valid image size of 1078392 for image:
==23242== Valgrind: "/tmp/cirrus-ci-build/testprogs/findalldevstest"
==23242==
==23242== Valgrind: debuginfo reader: Possibly corrupted debuginfo file.
==23242== Valgrind: I can't recover. Giving up. Sorry.
|
|\ \
| | |
| | | |
compiler: limit some Berkeley YACC workarounds to pre-2017-07-09 releases.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Those later releases only define yydebug if YYDEBUG is defined and only
declare yynerrs as global for non-reentrant parsers, so limit the
workarounds for those problems to releases before the 1.9 20170709
release.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Switch to FreeBSD 13.2 and size it down as in tcpdump much earlier.
Switch Clang to 15 in FreeBSD and Linux (for some reason there was no
pkg error in this project, but it makes sense to upgrade consistently
anyway).
|
| | |
| | |
| | |
| | |
| | |
| | | |
This reverts commit c7f7036094bf9c48fdfa803085857a35753e938f.
This is a follow-up to 5e6aead918140299fd5b12c74dd4c728732ab47d.
|
| | | |
|
| | |
| | |
| | |
| | |
| | |
| | | |
This should avoid, with autotools, when doing parallel builds:
make[2]: *** No rule to make target '../libpcap.a', needed by
'valgrindtest'. Stop.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
From https://www.gnu.org/software/make/manual/make.html#MAKE-Variable:
```
Recursive make commands should always use the variable MAKE, not the
explicit command name 'make', as shown here:
subsystem:
cd subdir && $(MAKE)
```
This should avoid the following warning when running 'make releasecheck'
in some cases:
make[1]: warning: jobserver unavailable: using -j1. Add '+' to parent
make rule.
|
|/ /
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The warnings were like:
perl: warning: Setting locale failed.
perl: warning: Please check that your locale settings:
LANGUAGE = (unset),
LC_ALL = (unset),
LANG = "en_US.UTF-8"
are supported and installed on your system.
perl: warning: Falling back to the standard locale ("C").
|
|\ \
| | |
| | | |
compiler: parse port numbers in ranges using the standard number parser.
|
|/ /
| |
| |
| |
| | |
Run them through the same code that's used for numbers in the lexical
analyzer, for consistency.
|
| |
| |
| |
| |
| | |
Apparently, the warning that Clang started to generate in release 15
still stands in Clang release 16.
|
| | |
|
| |
| |
| |
| |
| |
| | |
Report specific errors for various IP address and IP address+netmask
combinations when used as arguments to port, portrange, proto, and
protochain.
|
| |
| |
| |
| |
| | |
We don't support "<ipv6net> mask <ipv6mask>", so no need to pass an
<ipv6mask> argument.
|
| |
| |
| |
| |
| | |
Do the parsing ourselves; that avoids, for example, sscanf() mishandling
out-of-range numbers.
|
|\ \
| | |
| | | |
Make some static variables thread-local.
|
|/ /
| |
| |
| |
| |
| |
| |
| | |
This should make some functions thread-safe, even if it doesn't make
them fully-reentrant; at least some of them would requiree API changes
to full reentrancy.
This should fix #1174.
|
| |
| |
| |
| |
| |
| |
| |
| | |
Fix the cppcheck warning:
instrument-functions.c:77:15: warning: The scope of the variable
'instrument_type' can be reduced. [variableScope]
[skip ci]
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
In gencode.c instead of casting pointer types copy the data to squelch 4
previously known warnings from GCC and Clang. (Oddly enough, Haiku is
the only OS that unconditionally puts a 32-bit array into the union
inside struct in6_addr, yet the only OS where these warnings appeared.)
In pcap.c add a temporary workaround for ioctl() to squelch two other
Clang warnings:
pcap.c:1619:6: error: missing field 'size' initializer
[-Werror,-Wmissing-field-initializers]
pcap.c:1638:6: error: missing field 'size' initializer
[-Werror,-Wmissing-field-initializers]
With these changes GCC builds are now warning-free, so in build.sh set
LIBPCAP_TAINTED for Clang only and update the comment to show the
current remaining warnings.
|
| | |
|
| | |
|
| | |
|
| |
| |
| |
| |
| | |
Put the test for a particular named sanitizer into a macro, so it can be
used in the two loops that check sanitizers.
|
| |
| |
| |
| |
| |
| |
| |
| | |
This should keep silly tools that pass -DENABLE_SANITIZERS=OFF to CMake
if no sanitizers have been specified, rather than just not defining
ENABLE_SANITIZERS.
This should fix #1171.
|
|\ \
| | |
| | | |
npf: clean up handling of errors from setting promiscuous mode.
|
|/ /
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Handle both ERROR_NOT_SUPPORTED and "NDIS_STATUS_NOT_SUPPORTED with the
Customer bit set" as indications that an attempt to set promiscuous mode
was rejected because the adapter doesn't support it, in case the setting
of the Customer bit is dropped from the Npcap NPF driver in the future.
While we're at it, shuffle some #defines for status codes to put them
together and update comments.
|
| | |
|
| | |
|
| |
| |
| |
| | |
From repository git://git.savannah.gnu.org/config.git
|
| |
| |
| |
| | |
[skip ci]
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| | |
As Blake Easley points it out in GitHub bug report #1166, before commit
08a5870 in 2018 bpf_filter.c was a symlink to bpf/net/bpf_filter.c, but
now it is a regular file, so it is wrong to list it in .gitignore. Fix
that and reflect the file path change in DOS makefiles.
|
| |
| |
| |
| | |
Fix letter case and use bold/italics consistently with the other pages.
|
|/
|
|
|
|
| |
When "rpcapd" uses not a plain font, make that italics instead of bold.
Use bold for signal and file names, but not for service names. Get
newlines right for "-S" and "-X". Reword some prose for clarity.
|
| |
|
|
|
|
|
| |
The first 5.x release was 22 years ago, the vendor stopped supporting
this OS 10 years ago.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If the environment variable INSTRUMENT is
- unset or set to an empty string, print nothing, like with no
instrumentation
- set to "all" or "a", print all the functions names
- set to "global" or "g", print only the global functions names
The configuration with --enable-instrument-functions remains.
Note that before this change, the default was to print all functions.
Now it is to print nothing.
This allows to run with tcpdump, tcpslice or any program linked
to libpcap:
$ INSTRUMENT=a ./my_program ...
$ INSTRUMENT=g ./my_program ...
$ INSTRUMENT= ./my_program ...
or
$ export INSTRUMENT=global
$ ./my_program ...
This also allows to run the statically compiled binary on another host
after copying it.
Moreover:
Reduce the scope of a variable.
Rename a variable.
Remove '\n' in the perror() call.
Remove 2 spaces in function calls (style).
(Same as in tcpdump and tcpslice.)
[skip ci]
|