Update to 3.12.1. From the changelog:

* Fix calling opmlexport without arguments

doc: Updated mail/rss2email to 3.12.1

Update to 3.12. From the changelog:

* Drop support for EOL Python 3.4, add support for Python 3.8
* Add List_ID and List_Post headers to the generated emails
* Add a new `reply-changes` setting
* Improve configurability of text wrapping for the emails
* Use `platform.node()` instead of "dev.null.invalid" in
  "Message-ID" header
* Improve locking support for when multiple rss2email instances are run
  in parallel
* Fix handling of __VERSION__ and __URL__ in user-agent strings
* Fix opmlexport

doc: Updated mail/rss2email to 3.12

Update to 0.11.5 (as 0.999.11.5). From the changelog:

- Added pollmax limit to sslserver and sslhandle.
- sslclient streamlined with tcpclient. fehQlibs-15 are required.

Update to 1.11.5. From the changelog:

- Recovered tcpclient IPv4 connectivity in case of a given IPv4
  address to connect to employing fehQlibs-15 and the enhanced
  dns_ip_qualify routine.

Update to 36c. From the changelog:

- Streamlined with fehQlibs-15.

Update to fehQlibs-15 (as fehqlibs-0.9.15). From the changelog:

- Added the Guninski patch for alloc from Qualys (CVE-2005-1513).
- dns_ipq (IP qualify) now with unified IPv4/IPv6 handling and evaluation
  of DNS well-know names 'localhost' and 'ipv[4|6]-loopback' (RFC 6761).
- Fixed DNS lookup for IPv6 addresses. Adjusted DNS man pages.

Bump default BUILDLINK_API_DEPENDS, as all fehware requires this version.

doc: Updated net/djbdnscurve6 to 36c

doc: Updated net/fehqlibs to 0.9.15

doc: Updated net/ucspi-ssl to 0.999.11.5

doc: Updated net/ucspi-tcp6 to 1.11.5

Define PATH_MAX on Illumos. Respect CFLAGS.

Remove outdated MESSAGE. Add local copies of online manpages. Bump PKGREVISION.

doc: Updated mail/qtools to 0.56nb1

Strip -lcrypt on macOS, which (at least in recent versions) doesn't have it.

- Specify ISO-8859-1 encoding on a couple Texinfo docs that aren't UTF8.
  Fixes `utf8 "\\xF6" does not map to Unicode` that was breaking SmartOS bulk builds.
- Use afplay(1) on macOS (included since 2007) instead of deprecated audio/qtplay.
  Other platforms still need to define default audio players.
- Include librsvg to get the gdk-pixbuf SVG loader for solvege.svg.
- Bump PKGREVISION.

doc: Updated audio/solfege to 3.22.2nb5

Update to 2.9.2.1. From the changelog:

- Bugfixes.
- New bigkv_* set of functions in libdatastruct.

Update to 2020.05.04. From the changelog:

- The setstate command from idtools is now included in fdtools.
- Compatibility with the latest skalibs.
- Silencing modern compiler warnings.

doc: Updated devel/skalibs to 2.9.2.1

doc: Updated sysutils/fdtools to 2020.05.04

Update to 4.0. From the changelog:

New functionality:

* Added MemoryAccountant
* Added SimpleStringCache that also removed the memory leak caused by longjmp in C
* Thread-safe memory leak detector overloads
* New command-line options:
  * -h help option
  * -s shuffle (random) option
  * -t run a specific test option
  * -vv extra verbose option
  * -k add a package name to junit output
* Added new asserts: CHECK_COMPARE, and improved C macros
* Support for newer compilers and address sanitizer

Small improvements:

* Fixed problems with gdb
* More 16-bit support
* Added Makefile for making the examples with an installed CppUTest
* Small mock improvements
* Removed more compiler warnings
* Support for C++14, C++17, and C++2x (added to automated build)

Improved maintainability:

* Docker builds
* Vastly improved the automated build with more platforms and variants
* Continuously releasing the passing build
* MS-DOS support (added to automated build)

doc: Updated devel/cpputest to 4.0

Update to 1.1.0. From the changelog:

- Fixed Perl Pod errors, thanks to a pull request from Mohammad S Anwar
  (#470).
- Fixed test failures when running with the localization set to German
  or Italian. Thanks to Slaven Rezić for the report (#472).
- Fixed an issue when the full name of the current user is not set, so
  that it properly falls back on the username. Thanks to Slaven Rezić and
  Matthieu Foucault for the report and testing various fixes (#473).
- Eliminated an error when using the `-t` option to specify a target, due
  to a missing dependency declaration in the Target class. Thanks to
  Clinton Adams for the fix (#509)!
- Updated the IPC::System::Simple Win32 workaround added in 0.9999 to
  properly support released versions of IPC::System::Simple. This fixes
  errors running the database command-line clients on Windows (#503).
- Sqitch now only passes the `--password` option to the MySQL client if
  it was not read from the `.my.cnf` file, as it's more secure to let
  the client use `.my.cnf`, and it eliminates a warning from recent
  versions of the client. Thanks to Kiel R Stirling for the fix (#484)!
- Added a note to the tutorials to skip setting the `engine.$engine.client`
  config when using the Docker image.
- Added the new `check` command, which compares the SHA1 hashes of the
  deploy scripts to the database, and reports if any have been modified
  since they were deployed. Thanks to Matthieu Foucault for the pull
  request and diligent work on this feature (#477)!
- Added the `--modified` option to the `rebase` and `revert` commands, to
  revert to the earliest change with a modified deploy script. Handy for
  rapid rebasing during iterative development cycles. Thanks to Matthieu
  Foucault for this feature (#477)!
- Fixed an issue where the Snowflake engine would complain about not
  finding the account name even for commands that don't need them, such
  as `init`. Thanks to Stack Overflow user vbp13 for the report (#502).

doc: Updated databases/p5-App-Sqitch to 1.1.0

Simplify GitHub declarations. I can never remember how to do this.
Thanks, leot@!

Add Leah Neukirchen's implementation of DJB's redo build system in
portable C with zero dependencies.

Add and enable leahneukirchen-redo.

Avoid CONFLICTS with other redo implementations. Bump PKGREVISION.

Update to 0.42c. From the changelog:

- jobserver.py: _try_read()'s alarm timeout needs to throw an exception.
- Documentation fixes.

pkgsrc changes:

- Avoid CONFLICTS with other redo implementations.

doc: Added devel/leahneukirchen-redo version 0.2

doc: Updated devel/apenwarr-redo to 0.42c

doc: Updated devel/jdebp-redo to 1.4nb1

Skip building and installing API docs with Doxygen, as suggested
by the author. Bump PKGREVISION.

doc: Updated mail/qsmtp to 0.37nb2

Explicitly disable g-ir-doc-tool, since it's not in PLIST.

Build with OpenSSL 1.1.x. While here, wrap long line in MESSAGE.

Add sysutils to CATEGORIES to match ucspi-tcp.

Modernize patch filenames, adding patch comments. NFCI.

Modernize patch filenames, adding patch comments. NFCI.

Modernize patch filenames, adding patch comments. NFCI.

Modernize patch filenames, adding patch comments. NFCI.

Modernize patch filenames. NFCI.

Modernize patch filenames. NFCI.

Modernize patch filenames. NFCI.

Reorder CATEGORIES to mollify pkglint.

Fix Illumos build.

Assign non-default SUBST_FILES.djberrno with =, not += (because it
doesn't actually append).

Retire 'djbware-errno-hack' and associated options.mk cleverness.
Instead:

1. Package makefiles including their own options.mk
2. Packages say "SUBST_CLASSES+=djberrno" to get the hack, if needed
3. Packages adjust SUBST_FILES.djberrno, if needed

Should fix bulk build failures due to multiple inclusions of options.mk
and/or incorrect definitions of DJB_ERRNO_HACK.

Approved during the freeze by wiz@.

Append .md suffix to text docs. Adjust references.

Rationale, in some sort of order:

- Plain text is already valid Markdown (and we can improve it as we go)
- Markdown is already plain text, readable in all the same ways
- GitHub also renders Markdown nicely in web browsers
- Changing installed names for docs is a non-zero compatibility break,
  but we consider it small enough to tolerate:
    1. Programs are unaffected
    2. People will press <tab> and find what they're seeking

Supports goals: Being easily packaged by OS integrators
Broaches non-goals: Breaking compatibility (but tolerably so)
Risks: Users won't find needed info; patches to hier.c may get skewed

Fix Markdown formatting in examples.

Format unordered lists more consistently.

Improve README Markdown.

Make CHANGES an unordered list.

Make THANKS an unordered list.

Replace ```foo``` with `foo` for one-liners.

Replace ``foo'' with "foo".

Replace --- with ### for headings.

Replace --- with -- when it's simulating em-dash.

Tabs to spaces.

Wrap URLs in <>.

Update to 1.4.16. From the changelog:

indexers:

* Fix handling of XML empty tag syntax when there's a quoted parameter right
  before the closing `/>`.  This caused `<title xml:lang="en-US"/>` to treat
  the body text as the document title.  Spotted by Gaurav Arora.

* omindex: Fix killing of filter child process if the parent process receives a
  signal.  Spotted by Gaurav Arora.

omega:

* Reject $setrelevant without an argument list.  This has never been documented
  as allowed, and previously crashed with a segfault.  Fixes #802, reported by
  Gaurav Arora.

* If there's an error opening the databases we now close any we managed to open
  successfully before the error so that things like $dbsize can't end up
  reporting values for a subset of the specified databases.

portability:

* Use our own autoconf cache variable namespace (xo_cv_ prefix instead of
  ac_cv_) to avoid colliding with standard autoconf macro use if config.site or
  a shared config.cache is used.  The former case caused a build failure for
  the OpenBSD port with 1.4.15, reported by Lucas R.

Update to 1.4.16. From the changelog:

API:

* MSet::snippet(): The snippet now includes trailing punctuation which carries
  meaning or gives useful context.  See
  https://github.com/xapian/xapian/pull/180, reported by Robert Stepanek.

* MSet::snippet(): Fix segfault generating snippet from default-constructed
  MSet.  This probably isn't something you'd typically do, but it shouldn't
  crash.  Found during extended testing of #803 (which only affected git
  master) which was reported by Robert Stepanek.

* Remove trailing full stop from exception messages.  We conventionally don't
  include one, but a few cases didn't follow that convention.

testsuite:

* Replace direct use of ftime() which gives deprecation warnings with recent
  mingw.  Reported by srinivasyadav22.

matcher:

* Fix segfault in rare cases in the query optimiser.  We keep a pointer to the
  most recent posting list to use as a hint for opening the next posting list,
  but the existing mechanism to take ownership of this hint had a flaw.  We now
  invalidate the hint in situations where it might be indirectly deleted which
  is safe, but somewhat conservative.

* Improve the optimisation of an always-matching OP_VALUE_GE to also take
  effect when the value slot's lower bound is equal to the limit of the
  OP_VALUE_GE.  Patch from boda sadalla.

glass backend:

* Report the correct errno value if commit() fails.  We were potentially
  reporting ENOENT from an unlink() call cleaning up a temporary file prior to
  throwing the exception instead.

documentation:

* Fix missing menus in API documentation.  Newer doxygen generates .js files
  which we also need to distribute and install.  Reported by sec^nd on #xapian.

* Note OP_FILTER ignored subquery bug fixed in 1.4.15 as present in 1.4.14 and
  older.

portability:

* Use our own autoconf cache variable namespace (xo_cv_ prefix instead of
  ac_cv_) to avoid colliding with standard autoconf macro use if config.site or
  a shared config.cache is used.  The former case caused a build failure for
  the OpenBSD port with 1.4.15, reported by Lucas R.

* Use clock_gettime() and nanosleep() under modern mingw as these allow higher
  precision than what we previously used.

Bindings:

* Remove code to support SVN snapshots since we stopped using SVN more than 5
  years ago.

* Ignore overloads for logical ops, *, /.  These were already ignored for
  several languages, and aren't actually usefully wrapped for any of the other
  languages.

CSharp:

* Work around mono terminfo parsing bug in more cases.  With this, "make",
  "make check", "make install" and "make uninstall" all work on Ubuntu 18.10.
  Patch from Dipanshu Garg, fixes https://github.com/xapian/xapian/pull/287 and
  #801.

Lua:

* Allow passing a Lua function as a MatchSpy.  This was supposed to be
  supported already, but the typemaps weren't set up.

* On platforms where sizeof(long) is 4, SWIG was wrapping Xapian::BAD_VALUENO
  as a negative constant in Lua, which was then rejected by a check which
  disallows passing negative values for unsigned C++ types.  We now direct SWIG
  to handle Xapian::valueno as double (which is what numbers in Lua usually
  actually are) which gives us an unsigned constant, and also eliminates the
  negative value check.

* Correct documentation - get_description() is wrapped as tostring() in Lua,
  not str() as we previously claimed.

* Add test coverage for passing Lua function for a Stopper.

Perl:

* Resolve the remaining issues and remove the "experimental" marker:

  + Add search_xapian_compat() function which sets up aliases in the
    Search::Xapian namespace to aid writing code which uses either
    Search::Xapian or this module.

  + Allow passing Perl sub for simpler Xapian functor classes.  This fills in a
    missing feature compared to Search::Xapian.  See #523.

  + Remove useless PerlStopper class which was an incomplete copy of the
    apparently non-functional Search::Xapian::PerlStopper.  We now support
    passing a Perl sub for a Stopper object.

  + Adjust some method names to match Search::Xapian.  Iterators now support
    inc() (and dec() where the C++ class supports operator--) like
    Search::Xapian, rather than increment() and prev().  Reported by Eric Wong
    in #523.

  + Drop undocumented and unexpected extra equals() method.

  + Provide compatibility with ENQ_ASCENDING, etc constants.  SWIG wraps these
    as $Xapian::Enquire::ASCENDING, which better matches the C++ API, but
    Search::Xapian wraps this as Search::Xapian::ENQ_ASCENDING, etc so provide
    those too for compatibility.  Reported by Eric Wong in #523.

  + Drop stringification and int conversion overloads.  These seem more
    confusing than helpful, and overloading stringification works badly
    with SWIG-generated bindings.

  + Document remaining known differences from Search::Xapian.

* Update recently tested versions in README.

* Improve documentation.

* Fix t/02pod.t to look for files in right directory.

Ruby:

* Don't print iterator sizes to stdout.  This was some debugging accidentally
  left in as part of a change in 1.4.12.  Patch from Dan Callaghan.

doc: Updated textproc/xapian to 1.4.16

doc: Updated textproc/xapian-omega to 1.4.16

These run under -rcptcheck, not -qfilter.

Add and enable py-approvaltests.

Add and enable texttest.

Define rcvar so goaccess=NO works. Fix a spell-o. Bump PKGREVISION.

Explicitly ulimit qmailsend. Bump PKGREVISION.

Initial import of TextTest, an open source tool for text-based
functional testing. This means running a program in lots of different
ways from the command line, and using the text output produced as a
means of controlling the behaviour of that application.

Initial import of py-approvaltests, the Python port of an
assertion/verification library to aid testing.

You can use ApprovalTests to verify objects that require more than a
simple assert including long strings, large arrays, and complex hash
structures and objects. ApprovalTests really shines when you need a more
granular look at the test failure. Sometimes, trying to find a small
difference in a long string printed to STDOUT is just too hard!
ApprovalTests solves this problem by providing reporters which let you
view the test results in one of many popular diff utilities.

Note highlight update.

Restore missing .include of options.mk.

Support PKG_SYSCONFDIR. Bump PKGREVISION.

Update to 0.70. From the changelog:

[Fixed]
- Due to hash key randomization, incompatible column sets were reported
  where in fact the sets are equal (with tests)
- Multiple examples would always report incompatible column sets
- Multiple examples trigger parser error due to accessing a
  deprecated scenario attribute
- Too much code was running with $/ bound to `undef` when parsing
  Gherkin from file
- NAME section missing in two modules causing Dist::Zilla to fail
  insertion of VERSION section

[Changed]
- Some improved error messages

[Added]
- New --strict option for `pherkin` which causes an exit value of 1
  when there are 'pending' or 'missing' steps

Update to 0.71. From the changelog:

[Fixed]
- Feature file parser crashes on empty files
- 'prove' plugin doesn't run 'post_execute' hooks

Update to 3.56. From the changelog:

- added support for Sequence Alignment Maps (SAM files)
- added empty-file mode to `--no-trailing-nl`
  (https://gitlab.com/saalen/highlight/issues/147)
- fixed issue with `--syntax-by-name` waiting for stdin
  (https://gitlab.com/saalen/highlight/-/issues/151)
- fixed issue with `--syntax` reading matching files in the current
  working directory
  (https://gitlab.com/saalen/highlight/-/issues/151)
- fixed string parsing in lisp.lang
  (https://gitlab.com/saalen/highlight/-/issues/150)
- fixed output of UTF-8 text in xterm256 or truecolor output
  (https://gitlab.com/saalen/highlight/-/issues/152)
- fixed regex in js.lang (thanks to Jens Schleusener)
- fixed calculation of testcase markers with UTF-8 input
- allowed number literals with underscores in Java, Scala, D, Julia,
  C#, Perl and Ada definitions
- added Nord theme
  (https://gitlab.com/saalen/highlight/-/merge_requests/125)

Update to notqmail 1.08 (as qmail-1.03nb49). From the changelog:

## Fix bugs

- Vulnerabilities we've inherited from qmail 1.03, reported by Qualys.
    - CVE-2005-1515: fix signedness wraparound in `substdio_{put,bput}()`.
    - CVE-2005-1514: fix possible signed integer overflow in `commands()`.
    - CVE-2005-1513: fix integer overflow in `stralloc_readyplus()`.
    - Fix several other places where variables could overflow.
- `qmail-pop3d`: instead of running as root if root authenticates (and
  being a vector for a dictionary attack on the root password), exit 1
  to look just like a failed `checkpassword` login.
- `qmail-inject`: do not parse header recipients if `-a` is given.
- Correctly detect multiple IP addresses on the same interface.
- Remove workaround for ancient DNS servers that do not properly
  support CNAME.
- Fix possible integer overflow in `alloc()`.

## Reduce bug likelihood

- Remove `dnscname` and `dnsmxip` programs that were being built but not
  installed.
- Remove `systype` and related platform detection.
- Remove unused variable in `maildir.c`.
- Reduce variable scope in `tcpto.c`.
- Avoid local variables shadowing same-named globals.
- Avoid needing `exit.h` in named-pipe bug check.
- Add a `test` target and some unit tests, using Check.
- Add missing function declarations in `cdbmss.h`, `scan.h`.
- Add missing return types to `main()`.
- Add `hier.h` for inclusion in `instcheck.c`, `instchown.c`, `instpackage.c`.
- Use system headers and types instead of the `HASSHORTSETGROUPS` check.
- Use system headers instead of redeclaring `exit()`, `read()`,
  `write()`, `malloc()`, `free()`, `fork()`, `uint32_t`.
- Use C89 function signatures for code we've touched so far.
- Automated builds:
    - TravisCI: move setting `MAKEFLAGS` out of the script and into
      the matrix.
    - Add FreeBSD builds with CirrusCI.
    - Add a GitHub Actions build.

## Other changes

- Remove DJB's TODO.
- Replace many `pobox.com` URLs.
- Acknowledge Erik Sjölund's `qmail-local.c` bugfix that we've
  inherited from netqmail.
- Avoid generating catted manpages by building with `NROFF=true`.
- Optionally create a `systemd` service file.
- Run an alternate `qmail-remote` by setting `QMAILREMOTE` in
  `qmail-send`'s environment.

## Intent to remove

In the course of developing this release, we found programs that we
intend to remove in the next release. We believe none of these remains
necessary or useful enough to be worth the cost of maintaining. If you
disagree, please let us know!

- Remove `qsmhook`, long since replaced by `preline`.
- Remove inefficient `maildirwatch`.
- Remove obsolete mail client wrappers.
- Remove `qmail-pop3d`, since Maildir is well supported by actively
  maintained POP3 servers.

doc: Added devel/py-approvaltests version 0.2.6

doc: Added devel/texttest version 4.0.0

doc: Updated devel/p5-Test-BDD-Cucumber to 0.70

doc: Updated devel/p5-Test-BDD-Cucumber to 0.71

doc: Updated mail/qmail to 1.03nb49

doc: Updated mail/qmail-run to 20200519

doc: Updated sysutils/tarsnap to 1.0.39nb2

CI: make the NROFF tests to actually override nroff

These things just set environment variables, but since they were note exported
they were not visible to make, so nothing actually changed.

Make: make catman dependency rule portable

The previous version only worked with GNU make. It was not caught because of
broken CI rules.

add back accidentially removed typecast

Fixes: 2fe3ac71d35e68d1d42273a3925b9f7dc8020742

fmt.h: use prototypes, remove missing functions

fmt_*.c: convert to prototypes

remove obsolete mail client wrappers

These 3 scripts call maildir2mbox before calling one of 3 mail programs. Both
elq and pine are obsolete and are not maintained anymore, and I have no idea
what "Mail" was. Noone uses any of these today anymore, and most people don't
do on local mail stores either.

remove the "alloc" parameter from cdbmake_add() and cdbmake_split()

Both of these functions are only called from a single place, where both pass in
the alloc() function. The parameter was also named alloc, which can cause
compiler warnings when -Wshadow is enabled. Simply remove the parameter, which
makes everything work exactly as before.

remove the custom allocator

 * it's as simple as possible, probably intentional. This makes things like
   alloc_re() (i.e. realloc()) inefficient as it does not look if the thing it
   reallocates is already the last thing in the batch, and simply extends that,
   or if alloc_free() has the last block and adds things back to the pool
   (ideally with zeroing for security reasons).
 * it makes it basically impossible to detect overruns on early small
   allocations with tools like valgrind, address sanitizer or friends as
   those allocations never hit malloc(), so are never traced by any custom
   tools. Any overrun will just go into the next variable, untracable.
 * allocators do their own buffering, alignment, and so on. If they are broken,
   your system is severely screwed. And if we need more money^H^H^H^Hemory we
   will hit that buggy allocator anyway, making things harder to detect as it
   will only happen on "heavy" memory usage. This is not TruOS, SunOS 1.x or
   something like that anymore. We expect a reasonably sane OS. Everything else
   will get us screwed in much more severe ways.
 * it has shown issues when coming close to 4GiB allocations (#37, #109, older
   reports). DJB correctly says that an allocation of that size in your mailer
   shows that you have already done something wrong, but still: 64 bit
   platforms are the default, so using a 32 bit type to pass allocation sizes
   is a bug of it's own.
 * when any additional patch accidentially mixes calls to this functions with
   normal realloc()/free() this can lead to random crashes
 * why bother at all with having a custom allocator?

remove unused substdio_bget()

Spotted by Qualsys during their audit.

scan.h: use prototypes, remove missing functions

scan_*.c: convert to prototypes

use POSIX compliant calling of head and tail

use default strdup() instead of open coding it