Builder libpcap-solaris11-sparc Build #127

Results:

Build successful

SourceStamp:

BuildSlave:

Reason:

The SingleBranchScheduler scheduler named 'schedule-libpcap-solaris11-sparc' triggered this build

Steps and Logfiles:

1. git update ( 3 secs )
   1. stdio
2. shell './build_matrix.sh' ( 4 hrs, 27 mins, 28 secs )
   1. stdio
   2. config.log
3. trigger triggered trigger-tcpdump-solaris11-sparc ( 1 hrs, 14 mins, 57 secs )
   1. - no logs -
   2. tcpdump-solaris11-sparc #97

Build Properties:

Forced Build Properties:

Responsible Users:

1. Denis Ovsienko

   denisohnoyoudont@ovsienko.info

Timing:

All Changes:

:

1. Change #258445

   Category	libpcap
   Changed by	Denis Ovsienko <denisohnoyoudont@ovsienko.info>
   Changed at	Sat 21 Feb 2026 16:49:46
   Repository	https://git.tcpdump.org/libpcap
   Project	libpcap
   Branch	master
   Revision	69ca891986d5ddac184df5ce1b3d81ff406c7758

   Comments

   Untangle the ANDing in gen_relation_internal().

   Changed files

   • gencode.c

2. Change #258446

   Category	libpcap
   Changed by	Denis Ovsienko <denisohnoyoudont@ovsienko.info>
   Changed at	Sat 21 Feb 2026 16:49:55
   Repository	https://git.tcpdump.org/libpcap
   Project	libpcap
   Branch	master
   Revision	6fe3ac3b76f1a9bcfdc57743c5319367e2f03381

   Comments

   In "net <n> mask <m>" catch ENOMEM for the "m" too.
   In the grammar any of the two HIDs in "HID NETMASK HID" can be NULL if
   in the lexer sdup() failed because newchunk_nolongjmp() failed because
   calloc() failed because the process ran out of memory.  Since commit
   dfafa1f the grammar has a NULL guard for the first HID, but not for the
   second one.  When the latter only is NULL, the grammar does not skip the
   invocation of gen_mcode(), which then gets the NULL as the netmask
   string argument and a placeholder 0 value as the netmask length unsigned
   argument.
   
   In gen_mcode() this activates the code path for the "HID '/' NUM"
   syntax, so the network number 0.0.0.0 "passes" the bitwise check and
   eventually fails with a seemingly correct "out of memory" error message,
   but further downstream, and any other network number fails the bitwise
   check with a "non-network bits set" error message, which is misleading.
   
   In the grammar add a NULL guard for the second HID to fail early with
   "out of memory" in all cases.  Do not add any tests: even though the
   problem is reproducible with some effort, there is no reliable way to
   manage the external effects that define the failure scenario.

   Changed files

   • CHANGES
   • grammar.y.in

3. Change #258447

   Category	libpcap
   Changed by	Denis Ovsienko <denisohnoyoudont@ovsienko.info>
   Changed at	Sat 21 Feb 2026 17:46:34
   Repository	https://git.tcpdump.org/libpcap
   Project	libpcap
   Branch	master
   Revision	251867b8902d280281f2d33aec14d85389db077e

   Comments

   pcap-filter(7): Refine ext. headers and accessors.
   In the "port <name or number>" section document IPv6 extension headers
   as not supported.  In the "ip proto <name or number" section make it
   clear that IPv4 fragmentation is not a factor.  In the "ip6 proto <name
   or number>" section document the degree of support for extension
   headers.
   
   In the "protocol data accessors" section give an exhaustive list of
   protocols that libpcap currently implements on top of IPv4, for "icmp6"
   document extension headers as not supported, relate "igrp" with the same
   abbreviation.

   Changed files

   • pcap-filter.manmisc.in

4. Change #258448

   Category	libpcap
   Changed by	Denis Ovsienko <denisohnoyoudont@ovsienko.info>
   Changed at	Sat 21 Feb 2026 17:50:40
   Repository	https://git.tcpdump.org/libpcap
   Project	libpcap
   Branch	master
   Revision	d69f902f0a2efa25ef32766303262962a84ea7e9

   Comments

   Never return NULL from gen_bcmp().
   Simplify ANDing of the comparison blocks while at it.

   Changed files

   • gencode.c