Builder libpcap-solaris10-i386 Build #3114

Results:

Build successful

SourceStamp:

BuildSlave:

Reason:

The SingleBranchScheduler scheduler named 'schedule-libpcap-solaris10-i386' triggered this build

Steps and Logfiles:

1. git update ( 4 secs )
   1. stdio
2. shell Set developer mode ( 0 secs )
   1. stdio
3. shell_1 './autogen.sh' ( 5 secs )
   1. stdio
4. shell_2 './configure' ( 43 secs )
   1. stdio
   2. config.log
5. shell_3 'gmake' ( 23 secs )
   1. stdio
6. shell_4 'gmake check' ( 4 mins, 48 secs )
   1. stdio
7. shell_5 'rm -rf ...' ( 0 secs )
   1. stdio
8. shell_6 'gmake install ...' ( 3 secs )
   1. stdio
9. trigger triggered trigger-tcpdump-solaris10-i386 ( 5 mins, 39 secs )
   1. - no logs -
   2. tcpdump-solaris10-i386 #5799

Build Properties:

Forced Build Properties:

Responsible Users:

1. Denis Ovsienko

   denisohnoyoudont@ovsienko.info

Timing:

All Changes:

:

1. Change #227915

   Category	libpcap
   Changed by	Denis Ovsienko <denisohnoyoudont@ovsienko.info>
   Changed at	Mon 21 Apr 2025 22:33:40
   Repository	https://git.tcpdump.org/libpcap
   Project	libpcap
   Branch	master
   Revision	3c704d19c086c3e8259d60df1c94b39e47f995d4

   Comments

   filtertest: Plug a few resource leaks.
   Make the variables that need to be managed global, add a function to
   deallocate all resources if required and call it before exit() both in
   main() and in error().  This improves Valgrind SNR and resolves the
   following in particular:
   
   $ valgrind --leak-check=full --show-leak-kinds=all --quiet \
     ./testprogs/filtertest EN10MB 'invalid syntax'
   filtertest: can't parse filter expression: syntax error
   ==411153== 15 bytes in 1 blocks are still reachable in loss record 1 of 2
   ==411153==    at 0x48417B4: malloc (vg_replace_malloc.c:381)
   ==411153==    by 0x10EFCB: copy_argv (filtertest.c:229)
   ==411153==    by 0x10EFCB: main (filtertest.c:419)
   ==411153==
   ==411153== 648 bytes in 1 blocks are still reachable in loss record 2 of 2
   ==411153==    at 0x48465EF: calloc (vg_replace_malloc.c:1328)
   ==411153==    by 0x117651: pcap_open_dead_with_tstamp_precision (pcap.c:4476)
   ==411153==    by 0x10EE7B: main (filtertest.c:392)
   ==411153==

   Changed files

   • testprogs/filtertest.c

2. Change #227916

   Category	libpcap
   Changed by	Denis Ovsienko <denisohnoyoudont@ovsienko.info>
   Changed at	Mon 21 Apr 2025 22:35:11
   Repository	https://git.tcpdump.org/libpcap
   Project	libpcap
   Branch	master
   Revision	c92b5ec094d8c1fb991c381d12e72e367f73e72e

   Comments

   filtertest: Simplify usage().

   Changed files

   • testprogs/filtertest.c

3. Change #227917

   Category	libpcap
   Changed by	Denis Ovsienko <denisohnoyoudont@ovsienko.info>
   Changed at	Mon 21 Apr 2025 22:35:11
   Repository	https://git.tcpdump.org/libpcap
   Project	libpcap
   Branch	master
   Revision	50d9ca2fa2e16d9bcee079199378c0d1020395fa

   Comments

   TESTrun: Factor two helper functions out.

   Changed files

   • testprogs/TESTrun

4. Change #227918

   Category	libpcap
   Changed by	Denis Ovsienko <denisohnoyoudont@ovsienko.info>
   Changed at	Mon 21 Apr 2025 22:35:11
   Repository	https://git.tcpdump.org/libpcap
   Project	libpcap
   Branch	master
   Revision	a5090c63f88ade0525170f7bd4e0c941aaaf2f2e

   Comments

   TESTrun: Add more reject tests with pcap_ether_hostton().
   pcap_ether_hostton() returns an allocated 48-bit binary MAC address,
   which ought to be freed no matter if the attempt to use it was
   successful or not.  The latter case stands for several code paths that
   start at gen_scode(), invoke pcap_ether_hostton() and eventually
   bpf_error(); each of these ought to arrange a means to free the memory.
   Some of these code paths are already covered with tests, add tests to
   cover the remaining code paths and make the problem space completely
   visible to Valgrind.

   Changed files

   • testprogs/TESTrun

5. Change #227919

   Category	libpcap
   Changed by	Denis Ovsienko <denisohnoyoudont@ovsienko.info>
   Changed at	Mon 21 Apr 2025 22:35:11
   Repository	https://git.tcpdump.org/libpcap
   Project	libpcap
   Branch	master
   Revision	ad56fa246d80deeca6f632726eedde395162eb7b

   Comments

   Plug memory leaks from pcap_ether_hostton().
   pcap_ether_hostton() returns a pointer to allocated memory, which needs
   to be freed, which gen_scode() does for the successful case only.  To
   have the memory freed in any case, copy the value into a local variable
   and free the allocated memory immediately thereafter -- this suffices
   since the data size is fixed and the value needs to be used only once
   per code path.  In other words, apply the solution from commit 2b01095
   to the problem from commit bcbef22.

   Changed files

   • gencode.c

6. Change #227920

   Category	libpcap
   Changed by	Denis Ovsienko <denisohnoyoudont@ovsienko.info>
   Changed at	Mon 21 Apr 2025 22:35:11
   Repository	https://git.tcpdump.org/libpcap
   Project	libpcap
   Branch	master
   Revision	8f99c57221636f18f73ac026c9f9f7a9c9594dce

   Comments

   Consolidate MAC-48 address matching code.
   gen_gateway(), gen_scode(), gen_ecode() and gen_broadcast() all have a
   switch block with 11 or 13 DLTs and a default case.
   
   The purpose of the latter is to avoid any calls to pcap_ether_hostton()
   in gen_scode() case Q_HOST proto Q_LINK if the DLT does not use MAC-48
   addresses, so add a new is_mac48_linktype() function to support this
   behaviour; for consistency add the same guard to gen_scode() case
   Q_GATEWAY and to gen_ecode().
   
   The purpose of the latter is to avoid any calls to pcap_ether_hostton()
   in gen_scode() if the DLT does not use MAC-48 addresses, so add a new
   is_mac48_linktype() function to support this behaviour; add similar
   guards to gen_scode() case Q_GATEWAY and to gen_ecode() for consistency.
   
   These changes replace a few custom messages with templates, so update,
   add and coalesce the tests as required.

   Changed files

   • gencode.c
   • testprogs/TESTrun