Changes:
* improve reporting of invalid diff command arguments (reported by Dan Shearer)
* accept libfossil global -V|--verbose option for all commands
* plug memleak when using the p keymap in blame view
* fix fnc blame -r and make it behave like fossil blame -o for familiarity
* improve fnc timeline <path> lookup of repository paths not in the work tree
* make tree navigation more intuitive when opening repository subtrees
* implement [ & ] keymaps to navigate to the previous/next hunk in the diff
* allow fnc stash get to take an optional <stash_id> argument
* document fnc(1) missing W diff view keymap to ignore end-of-line whitespace
* document optional <stash_id> argument of fnc stash pop [stash_id]
* fix typo producing gcc 12.1.0 compiler warning (comparison always true)
* improve scope matching for function prototype in diff hunk headers
* ifdef out A_BOLD attr on OpenBSD to fix diff view line highlight in xterm(1)
* display checkout hash and build date with fnc -v like fossil(1)
* miscellaneous diff API implementation improvements
* protect against out-of-bound reads now line_type array is optionally populated
* add -b|--brief option to fnc diff and b diff view keymap
* general fnc(1) manual page markup, syntax, and documentation improvements
* rename FNC_DIFF_COLOUR_CHUNK to FNC_DIFF_COLOUR_HUNK to match internal naming
* (breaking change: existing FNC_DIFF_COLOUR_CHUNK will be silently ignored)
* ensure diff parser does not internally mark modified files as renamed
* improve const correctness with fsl_uuid_cstr where possible
* fix contrived empty diff case when work tree root is passed to fnc diff
* improve path handling of files on disk vs. files in the repository database
* produce less verbose output on error by not presumptuously dumping help
* zap unnecessary allocation and minor refactor in the path parser
* don't display empty diff in contrived modified checkout edge case
* catch missed fsl_content_get() fsl_unlink_file() & fsl_mprintf() return codes
* miscellaneous style(9) fixes
* fix missed format specifier argument in blame path error message
* ensure fnc stash (get|pop) doesn't report updated files as renames
* use "(checkout)" label in diff header when diffing or stashing the work tree
* internal stash refactor in preparation to move code into libfossil
* catched missed SIGINT and SIGTERM signals
* use CHAR_BIT for bitstring manipulation macros
* fix assumption of invalid command argument as a path to pass to fnc timeline
* merge upstream libfossil with various changes
OK op@
if setuptools is present (see site-packages/distutils-precedence.pth);
enforce a dependency to avoid the case where setuptools might be removed
mid-run, causing g-ir-scanner to fail. ok aja@, who reported sporadic
build failures which must have been a result of this (I think it may
relate to runtime patching of distutils function relating to the MSVC
compiler).
Initially ported to avoid the bundled copy in net/tg_owt, but other ports
can probably pick it up as well.
Feedback OK sthen tb
---
Collection of CRC32C implementations that dispatches to one suitable to
the hardware. CRC32C is the cyclic redundancy check that uses the iSCSI
polynomial specified in RFC 3720.
Various fixes to normal form handling in GVariant
These fixe cover various issues encountered by calling
g_variant_get_normal_form() on untrusted arrays and tuples.
OK sthen@ Init port from Brad
Comment:
minimal and cross platform C++ GUID library
Description:
CrossGuid is a minimal, cross platform, C++ GUID library. It uses the
best native GUID/UUID generator on the given platform and had a generic
class for parsing, stringifying, and comparing IDs.
Maintainer: The OpenBSD ports mailing-list <ports@openbsd.org>
WWW: https://github.com/graeme-hill/crossguid
leave MODPY_VERSION= ${MODPY_DEFAULT_VERSION_2} in individual Makefiles for
now as I'll try to remove some, but running tests for these ports takes
forever and that's not finished yet
The build now involves various C++11 features like trailing commas in
enums. And an upstream commit attempts to fix macros by enabling tests
depending on C++11 or gcc >= 4.8 everywhere. Switch compiler on sparc64
to ports-gcc while waiting for upstream to clean this mess up.
c19904d6e8
ok aja
it was integrated into python core as the "sqlite3" module; nmap's zenmap
subpackage listed a dependency, but didn't really use this
(the module is confusingly named; it is for sqlite3 not sqlite2!)
Uses getpwnam(3) and crypt(3) to check the user passwords, but our
getpwnam(3) doesn't return password hashes since 2016.
Problem spotted by William Orr while toying with this port. William
provided a fix but it's more reasonable to drop this ancient piece of
software. ok tb@
zenmap is a python 2 gui supposed to be run as root, ndiff diffs nmap
outputs. ndiff from the nmap package requires the nmap-zenmap package to
be installed, but doesn't depend on it and won't work unless you actually
install/use zenmap... which noone ever complained about.
Upstream has some active python 3 pull requests, but nothing concrete and
nmap port updates are overdue anyway.
Completely remove all python (2) bits from this port.
Feedback jca sthen
OK jca (on zenmap removal, ndiff removal probably as well)
OK sthen
no maintainer, python 2 only, just a few old distros still package it,
no upstream activity since years, i don't even see our 1.1b1 version
from 2015 on their homepage which still takes about python 2.2.2,
outlook 2000 and the alpha stage this software is in.
OK jca
no maintainer, python 2 only, no consumers, last port update in 2003,
upstream has newer versions (don't know since when), no big distro
packages it, homepage is reachable but clicking on "Download" yields
a python 2.4 backtrace on EACCES failure...
OK jca
Unmaintained since 2010, dead upstream since 2014, a single python 2 script
without build phase, only a tiny patch for serial console convenience.
Users can just grab the single file from github and run it themselves,
no need to keep py2 package around for this.
homepage times out for me, no maintainer, python 2 only, no consumers,
no update since 2009, we lag behind the last upstream version from 2018.
There is plenty of maintained irc bot software to chose from.
OK sthen
no maintainer, no consumers, python 2 only, no update since 2016,
py3 net/py-pcapy exists (albeit broken currently).
Not a good idea to parse network packets with dead python 2 code.
OK sthen
Unmaintained since its only update in 2011, tb already fixed up LibreSSL
breakage, upstream URL times out, noone but AIX and Arch AUR packages it.
maintainer timeout
OK tb
pytest-asyncio is an Apache2 licensed library, written in Python,
for testing asyncio code with pytest.
asyncio code is usually written in the form of coroutines, which
makes it slightly more difficult to test using normal testing tools.
pytest-asyncio provides useful fixtures and markers to make testing
easier.
Autocommand turns a function into a command-line program. It converts
the function's parameter signature into command-line arguments, and
automatically runs the function if the module was called as __main__.
In effect, it lets you create a smart main function.
setuptools (it's used as a package locator but importlib.metadata in
newer Python core or the external importlib_metadata are preferred).
So drop the RDEP in that case (it's still kept for py27) and bump
revisions.
Previously, gccmakedep(1) was incidentally enabled by way of including
<xorg.cf> on those architectures for which an old configuration
existed for building the X11 server. Other archs would fall back
to using makedepend(1) for Makefile dependency generation.
ok tb@ phessler@
Mypy is an optional static type checker for Python that aims to combine
the benefits of dynamic (or "duck") typing and static typing.
Mypy combines the expressive power and convenience of Python with a
powerful type system and compile-time type checking.
Mypy type-checks standard Python programs; run them using any Python VM
with basically no runtime overhead.
This port remains unmaintained since import in 2011, only on life support by
tb and his LibreSSL efforts.
Previous changes saw no response maintainer and this port is already a
maintainence burden for porters.
maintainer timeout (again)
OK tb