Age | Commit message (Collapse) | Author | Files | Lines |
|
Simple DirectMedia Layer - SDL version 2.0.14 (stable)
In addition to lots of bug fixes, here are the major changes in this
release:
General:
* Added support for PS5 DualSense and Xbox Series X controllers to the
HIDAPI controller driver
* Added game controller button constants for paddles and new buttons
* Added game controller functions to get additional information:
* SDL_GameControllerGetSerial()
* SDL_GameControllerHasAxis()
* SDL_GameControllerHasButton()
* SDL_GameControllerGetNumTouchpads()
* SDL_GameControllerGetNumTouchpadFingers()
* SDL_GameControllerGetTouchpadFinger()
* SDL_GameControllerHasSensor()
* SDL_GameControllerSetSensorEnabled()
* SDL_GameControllerIsSensorEnabled()
* SDL_GameControllerGetSensorData()
* SDL_GameControllerRumbleTriggers()
* SDL_GameControllerHasLED()
* SDL_GameControllerSetLED()
* Added the hint SDL_HINT_JOYSTICK_HIDAPI_PS5 to control whether the
HIDAPI driver for PS5 controllers should be used.
* Added joystick functions to get additional information:
* SDL_JoystickGetSerial()
* SDL_JoystickRumbleTriggers()
* SDL_JoystickHasLED()
* SDL_JoystickSetLED()
* Added an API to allow the application to create virtual joysticks:
* SDL_JoystickAttachVirtual()
* SDL_JoystickDetachVirtual()
* SDL_JoystickIsVirtual()
* SDL_JoystickSetVirtualAxis()
* SDL_JoystickSetVirtualButton()
* SDL_JoystickSetVirtualHat()
* Added SDL_LockSensors() and SDL_UnlockSensors() to guarantee exclusive
access to the sensor list
* Added SDL_HAPTIC_STEERING_AXIS to play an effect on the steering wheel
* Added the hint SDL_HINT_MOUSE_RELATIVE_SCALING to control whether
relative motion is scaled by the screen DPI or renderer logical size
* The default value for SDL_HINT_VIDEO_MINIMIZE_ON_FOCUS_LOSS is now
false for better compatibility with modern window managers
* Added SDL_GetPreferredLocales() to get the application’s current
locale setting
* Added the hint SDL_HINT_PREFERRED_LOCALES to override your
application’s default locale setting
* Added SDL_OpenURL() to open a URL in the system’s default browser
* Added SDL_HasSurfaceRLE() to tell whether a surface is currently using
RLE encoding
* Added SDL_SIMDRealloc() to reallocate memory obtained from
SDL_SIMDAlloc()
* Added SDL_GetErrorMsg() to get the last error in a thread-safe way
* Added SDL_crc32(), SDL_wcscasecmp(), SDL_wcsncasecmp(), SDL_trunc(),
SDL_truncf()
* Added clearer names for RGB pixel formats, e.g.
SDL_PIXELFORMAT_XRGB8888, SDL_PIXELFORMAT_XBGR8888, etc.
Windows:
* Added the RAWINPUT controller driver to support more than 4 Xbox
controllers simultaneously
* Added the hint SDL_HINT_JOYSTICK_RAWINPUT to control whether the
RAWINPUT driver should be used
* Added the hint SDL_HINT_JOYSTICK_HIDAPI_CORRELATE_XINPUT to control
whether XInput and WGI should be used to for complete controller
functionality with the RAWINPUT driver.
macOS:
* Added the SDL_WINDOW_METAL flag to specify that a window should be
created with a Metal view
* Added SDL_Metal_GetLayer() to get the CAMetalLayer backing a Metal
view
* Added SDL_Metal_GetDrawableSize() to get the size of a window’s
drawable, in pixels
Linux:
* Added Vulkan support to the KMSDRM video driver
* Added the hint SDL_HINT_AUDIO_DEVICE_APP_NAME to specify the name that
shows up in PulseAudio for your application
* Added the hint SDL_HINT_AUDIO_DEVICE_STREAM_NAME to specify the name
that shows up in PulseAudio associated with your audio stream
* Added the hint SDL_HINT_LINUX_JOYSTICK_DEADZONES to control whether
HID defined dead zones should be respected on Linux
* Added the hint SDL_HINT_THREAD_PRIORITY_POLICY to specify the thread
scheduler policy
* Added the hint SDL_HINT_THREAD_FORCE_REALTIME_TIME_CRITICAL to allow
time critical threads to use a realtime scheduling policy
Android:
* Added SDL_AndroidRequestPermission() to request a specific system
permission
* Added the hint SDL_HINT_ANDROID_BLOCK_ON_PAUSE_PAUSEAUDIO to control
whether audio will pause when the application goes intot he background
OS/2:
* Added support for OS/2, see docs/README-os2.md for details
Emscripten (running in a web browser):
* Added the hint SDL_HINT_EMSCRIPTEN_ASYNCIFY to control whether SDL
should call emscripten_sleep internally
|
|
|
|
Place an exclusive or shared lock on a file. It uses 'LockFile' on
Windows and 'fcntl' locks on Unix-like systems.
|
|
More bulk build unbreaking...
|
|
|
|
Upstream changes:
0.30 2020-10-26
- Added function File::MimeInfo::has_mimeinfo_database() to test if
mimeinfo database is found
- mimetype and mimeopen now exit with an error code if no mimeinfo database
is found
- Added explicit dependency on Encode::Locale
- Properly decode parameters for `mimeopen` and `mimetype`.
Patch provided by Nikos Skalkotos
- Fix for undefined defaults (fixes #36).
Reported & fix provided by Martin Vassor
- Improved installation instructions, specifically on macOS, by
Oleg Kostyuk & Michiel Beijen
|
|
Upstream changes:
6.14 2020-11-30 05:48:07 -0700
- Production version identical to 6.12_01
6.12_01 2020-11-27 16:13:15 -0700
- Handle default apache indexes (gh#20)
- Ignore certain known navigation links in apache indexes (gh#20)
- Support Win32 OpenSSH listing (gh#21 simon04++)
6.11 2020-10-06 03:00:45 -0600
- Remove accidental recursive req on LWP::Simple
6.10 2020-10-05 15:18:30 -0600
- Production version identical to 6.09_01
6.09_01 2020-10-05 08:17:32 -0600
- Remove TABs from source files and tests (gh#17)
- Restores (theoretical, untested) support for 5.6, which was removed when
the tests added modern file tests (gh#17)
- Fixes or silences several minor nits reported by PerlCritic (gh#17)
6.08_01 2020-10-01 17:28:35 -0600
- We now only call Time::Local::timelocal with four digit years (gh#5, gh#14)
6.07 2020-09-30 08:47:18 -0600
- Production version identical to 6.06_01
6.06_01 2020-09-30 06:34:08 -0600
- Add examples to documentation.
6.05_01 2020-09-29 21:25:35 -0600
- New maintainer (Graham Ollis / plicease)
- New GitHub repository https://github.com/PerlAlien/File-Listing
- Include Perl core modules in prereqs (rt#85100 / gh#2)
- Converted Changes file to standard format (the original
Changes file is available as Changes.original)
- Migrate Test.pm tests to use Test::More, all tests
use strict and warnings now.
|
|
Upstream changes:
version 2.06 at 2020-12-23 17:32:27 +0000
-----------------------------------------
Change: 3d44f83c8ccc5eba3ad3f449bab2a985464b7726
Author: Chris 'BinGOs' Williams <chris@bingosnet.co.uk>
Date : 2020-12-23 17:32:27 +0000
Fixes for Big Sur
Dynaloader fix:
https://github.com/perl/perl5/commit/d296ead16762852ec34d173616a271c8
56711f77
ExtUtils::MakeMaker fix:
https://github.com/Perl-Toolchain-Gang/ExtUtils-MakeMaker/commit/21a4
41da515d55654bca4971f21ad2550de069d2
Made adjustments to the check_patchperl tool
-----------------------------------------
version 2.04 at 2020-11-20 12:36:29 +0000
-----------------------------------------
Change: 771e4fd9595bcc90f482f454203d84c568814966
Author: Chris 'BinGOs' Williams <chris@bingosnet.co.uk>
Date : 2020-11-20 12:36:29 +0000
Tweaked hints patching
Hints file patching now occurs before other patching and has an
independent certified perl version.
Updated included hints files to v5.33.4 level.
-----------------------------------------
version 2.02 at 2020-11-08 19:28:50 +0000
-----------------------------------------
Change: f02eab39c2f0073e6aceff6bb0b73f4d03e39b8f
Author: Chris 'BinGOs' Williams <chris@bingosnet.co.uk>
Date : 2020-11-08 19:28:50 +0000
This is 2.02
Change: fa458343e6ae7f70412c466eb604633db44bf24b
Author: Chris 'BinGOs' Williams <chris@bingosnet.co.uk>
Date : 2020-11-08 19:19:39 +0000
Finesse the check_patchperl script
Suppress the git output and only show patchperl output if there is a
problem with patching
Change: d4fe0e9fa23d0fd78b368a5edb34550120e2c233
Author: Ichinose Shogo <shogo82148@gmail.com>
Date : 2020-11-08 19:18:21 +0000
fix env command syntax
Change: 20692277283d6eb1616eab987583b33cfb612198
Author: Ichinose Shogo <shogo82148@gmail.com>
Date : 2020-11-08 19:18:21 +0000
update deprecated GitHub Actions commands
add-path and set-env are deprecated for security reason.
https://github.blog/changelog/2020-10-01-github-actions-deprecating-s
et-env-and-add-path-commands/ update them to use the new environment
files.
https://docs.github.com/en/free-pro-team@latest/actions/reference/wor
kflow-commands-for-github-actions#environment-files
Change: 175ccb4bc8c88a0e3e3f4600fe1db09f023a9999
Author: Ichinose Shogo <shogo82148@gmail.com>
Date : 2020-11-07 09:17:00 +0000
fix -Duseshrplib builds on darwin
port of https://github.com/Perl/perl5/issues/15057
Change: 7c806f542ffd0b5e75d724404012fd8bc9afc4e4
Author: Ichinose Shogo <shogo82148@gmail.com>
Date : 2020-11-07 09:12:34 +0000
Patch Time::HiRes in 5.24.2
Time::HiRes is broken on threaded darwin in 5.24.2, not only 5.24.1
and 5.24.0. https://github.com/bingos/devel-patchperl/pull/26 is need
in Perl 5.24.2.
Change: eaf8ec2500acde369ad78aea12d1de0011480a0e
Author: Ichinose Shogo <shogo82148@gmail.com>
Date : 2020-06-23 15:26:13 +0000
Perl 5.32.0 is released
Change: 0b60d5d7d06de737cc4e033f658aad8cff906657
Author: Chris 'BinGOs' Williams <chris@bingosnet.co.uk>
Date : 2020-06-09 16:26:25 +0000
Fix make depend problems with v5.9.[0-4]
Change: df4f5009c804675bf97160490a95af8a67ded395
Author: Chris 'BinGOs' Williams <chris@bingosnet.co.uk>
Date : 2020-06-09 16:00:56 +0000
Fix patching the v5.7.x series
Change: 3dce408aaf97a3b245b29303f20b30a7584fb362
Author: Chris 'BinGOs' Williams <chris@bingosnet.co.uk>
Date : 2020-06-08 21:13:06 +0000
Added a tool to check patchperl from the repository
It checkouts the perl5 git repository and runs patchperl for all of
the supported perl versions
Change: 3e819851d4e28806c629ed783ef56b02f6e5ad12
Author: Chris 'BinGOs' Williams <chris@bingosnet.co.uk>
Date : 2020-06-08 19:09:04 +0000
Fixed utils/h2ph.PL patching
v5.21.9 actually already had the fix
Change: 647e88747c8469d9c9b418a78a5b999984f17acb
Author: Ichinose Shogo <shogo82148@gmail.com>
Date : 2020-06-08 12:28:52 +0000
introduce GitHub Actions
|
|
Upstream changes:
Changes for version 6.83 - 2020-12-01
Time zone fixes No new zoneinfo data so no release on 2020-09-01.
Newest zoneinfo data (tzdata 2020d). Requested by Shawn Carroll (GitHub #32)
|
|
- use github links
- use unsigned char patches from original author
- patch configure.ac to disable doxygen since original is troublesome on netbsd9 system
|
|
|
|
Originally packaged by Nikita in wip.
Tokei is a program that displays statistics about your code. It will
show number of files, total lines within those files and code,
comments, and blanks grouped by language. Some of Tokei's features
are:
- Tokei is fast
- Tokei is accurate, it correctly handles multi line comments,
nested comments, and not counting comments that are in
strings. Providing an accurate code statistics.
- It supports a huge range of languages, supporting over 150
languages, and their various extensions.
- It can output in multiple formats(CBOR, JSON, TOML, YAML)
allowing Tokei's output to be easily stored, and reused. These can
also be reused in Tokei combining a previous run's statistics with
another set.
|
|
* Fix HOMEPAGE.
Changelog:
* Remove artificial restriction on lifetime parameters on enums (#604)
* Add an option for converting usize/isize into size_t/ptrdiff_t. (#606)
* Allow controlling the cargo profile used for expansion. (#607)
* Support wider range of expressions in enum discriminants (#614)
* Support generation of Cython bindings (#590)
* Fixed some issues with style=tag and recursive structs (#615)
* Default C style to Both (as specified in docs) (#615)
* Fix resolution of path dependencies from certain modules. (#629)
* Support inlined definitions for tuple variants with a single field in C (#631)
|
|
An easy way to determine which directories on the users computer you
should use to save data, caches and logs. A port of Python's 'Appdirs'
(\url{https://github.com/ActiveState/appdirs}) to R.
|
|
# CHANGES IN xfun VERSION 0.19
## NEW FEATURES
- Added functions `bg_process()` to run a command in a background
process, and `proc_kill()` to kill a process.
- Added a function `del_empty_dir()` to delete a directory if it is
empty.
- Added functions `is_abs_path()` and `is_rel_path()` to check if
paths are absolute or relative.
- Added a function `is_sub_path()` to test if a path is under a
directory.
- Added a function `is_web_path()` to test if a path is a web path
that starts with `http://` or `https://` or `ftp://` or `ftps://`.
- Documented and exported the previously internal functions
`dir_exists()` and `file_exists()` (thanks, @cderv, #36).
- Added a function `dir_create()` to create a directory recursively by
default when it does not exist.
- Added an argument `fail` to `Rscript_call()` to allow users to
customize the error message when an error occurred in calling the
function in a new R session.
## MINOR CHANGES
- `file_ext()`, `sans_ext()`, and `with_ext()` no longer use
`tools::file_ext()` or `tools::file_path_sans_ext()`, but provide a
slightly different implementation. They treat `tar.(gz|bz2|xz)` and
`nb.html` as file extensions, and also allow extensions to contain a
trailing `~` or `#`.
|
|
|
|
The goal of 'pak' is to make package installation faster and more
reliable. In particular, it performs all HTTP operations in parallel,
so metadata resolution and package downloads are fast. Metadata and
package files are cached on the local disk as well. 'pak' has a
dependency solver, so it finds version conflicts before performing the
installation. This version of 'pak' supports CRAN, 'Bioconductor' and
'GitHub' packages as well.
|
|
|
|
Lp_solve is freely available (under LGPL 2) software for solving
linear, integer and mixed integer programs. In this implementation we
supply a "wrapper" function in C and some R functions that solve
general linear/integer problems, assignment problems, and
transportation problems. This version calls lp_solve version 5.5.
|
|
|
|
Metadata and package cache for CRAN-like repositories. This is a
utility package to be used by package management tools that want to
take advantage of caching.
|
|
Provides two convenience functions assert() and test_pkg() to
facilitate testing R packages.
|
|
See also:
https://www.tidyverse.org/blog/2020/12/usethis-2-0-0/
# usethis 2.0.0
This version is anticipated to be released as usethis v2.0.0.
## Adoption of gert and changes to Git/GitHub credential handling
Usethis has various functions that help with Git-related tasks, which
break down into two categories:
1. Git tasks, such as clone, push, and pull. These are things you
could do with command line Git.
1. GitHub tasks, such as fork, release, and open an issue or pull
request. These are things you could do in the browser or with the
GitHub API.
We've switched from git2r to the gert package for Git operations
(<https://docs.ropensci.org/gert/>). We continue to use the gh package
for GitHub API work (<https://gh.r-lib.org>).
The big news in this area is that these lower-level dependencies are
getting better at finding Git credentials, finding the same
credentials as command line Git (and, therefore, the same as RStudio),
and finding the same credentials as each other. This allows usethis to
shed some of the workarounds we have needed in the past, to serve as a
remedial "credential valet".
Under the hood, both gert and gh are now consulting your local Git
credential store, when they need credentials. At the time of writing,
they are using two different even-lower-level packages to do this:
* gert uses the credentials package (<https://docs.ropensci.org/credentials/>)
* gh uses the gitcreds package (<https://gitcreds.r-lib.org/>)
Even now, gert and gh should discover the same credentials, at least
for github.com. In the future, these two packages may merge into one.
Git/GitHub credential management is covered in a new article:
[Managing Git(Hub)
Credentials](https://usethis.r-lib.org/articles/articles/git-credentials.html)
The main user-facing changes in usethis are:
* usethis should discover and use the same credentials as command line Git.
* usethis should be able to work with any GitHub deployment. While
github.com is the default, GitHub Enterprise deployments are fully
supported. The target GitHub host is determined from the current
project's configured GitHub remotes, whenever possible.
As a result, several functions are deprecated and several other
functions have some deprecated arguments.
* Deprecated functions:
- `use_git_credentials()`
- `git_credentials()`
- `github_token()`
* Functions with (deprecated arguments):
- `create_from_github()` (`auth_token`, `credentials`)
- `use_github()` (`auth_token`, `credentials`)
- `use_github_links()` (`host`, `auth_token`)
- `use_github_labels()` (`repo_spec`, `host`, `auth_token`)
- `use_tidy_labels()` (`repo_spec`, `host`, `auth_token`)
- `use_github_release()` (`host`, `auth_token`)
The switch to gert + credentials should eliminate most
credential-finding fiascos. Gert also takes a different approach to
wrapping libgit2, the underlying C library that does Git
operations. The result is more consistent support for SSH and TLS,
across all operating systems, without requiring special effort at
install time. More users should enjoy Git remote operations that "just
work", for both SSH and HTTPS remotes. There should be fewer
"unsupported protocol" errors.
## GitHub remote configuration
Usethis gains a more formal framework for characterizing a GitHub
remote configuration. We look at:
* Which GitHub repositories `origin` and `upstream` point to
* Whether you can push to them
* How they relate to each other, e.g. fork-parent relationship
This is an internal matter, but users will notice that usethis is more
clear about which configurations are supported by various functions
and which are not. The most common configurations are reviewed in a
[section of Happy
Git](https://happygitwithr.com/common-remote-setups.html).
When working in a fork, there is sometimes a question whether to
target the fork or its parent repository. For example,
`use_github_links()` adds GitHub links to the URL and BugReports
fields of DESCRIPTION. If someone calls `use_github_links()` when
working in a fork, they probably want those links to refer to the
*parent* or *source* repo, not to their fork, because the user is
probably preparing a pull request. Usethis should now have better
default behaviour in these situations and, in some cases, will present
an interactive choice.
## Default branch
There is increasing interest in making the name of a repo's default
branch configurable. Specifically, `main` is emerging as a popular
alternative to `master`. Usethis now discovers the current repo's
default branch and uses that everywhere that, previously, we had
hard-wired `master`.
`git_branch_default()` is a newly exported function that is also what's used internally.
`use_course()`, `use_zip()`, and `create_download_url()` all have some
support for forming the URL to download a `.zip` archive of a repo,
based on a repo specification (e.g. `OWNER/REPO`) or a browser
URL. These helpers now form a URL that targets `HEAD` of the repo,
i.e. the default branch.
## Changes to Git/GitHub functionality
The default Git protocol is now "https" and we no longer provide an
interactive choice, by default, in interactive sessions. As always, a
user can express a preference for "ssh" in individual function calls,
for an R session via `use_git_protocol()`, and for all R sessions via
the `usethis.protocol` option (#1262).
`pr_resume()` is a new function for resuming work on an existing local
PR branch. It can be called argument-less, to select a branch
interactively.
`pr_fetch()` can also be called with no arguments, to select a PR
interactively. The `owner` argument is replaced by `target`, with a
choice of the source (default) or primary repo.
`pr_forget()` is a new function for abandoning a PR you initiated
locally or fetched from GitHub. It only does local clean up and, for
example, doesn't delete a remote branch or close a PR (#1263).
`pr_view()` can now be called with no arguments. If the current branch
is associated with an open PR, we target that and, otherwise, we offer
an interactive selection.
`pr_finish()` deletes the remote PR branch if the PR has been merged
and the current user has the power to do so, i.e. an external
contributor deleting their own branch or a maintainer deleting a
branch associated with an internal PR (#1150). It no longer errors if
the PR branch has already been deleted (#1196).
`pr_pull_upstream()` is renamed to `pr_merge_main()` to emphasize that
it merges the **main** line of development into the current branch,
where the main line of development is taken to mean the default
branch, as reported by `git_branch_default()`, of the source repo,
which could be either `upstream` or `origin`, depending on the
situation.
`create_from_github()` will only create a read-only clone, due to lack
of a GitHub personal access token, if explicitly directed to do so via
`fork = FALSE`.
`create_from_github()` and `use_tidy_thanks()` accept browser and Git
URLs as the `repo_spec` argument, to be friendlier to copy/paste. When
a URL is passed, the `host` is also extracted from it.
`create_github_token()` is a new name for the function previously
known as `browse_github_token()` and `browse_github_pat()`.
`issue_close_community()` and `issue_reprex_needed()` are two new
functions for maintainers who process lots of GitHub issues. They
automate canned replies and actions, e.g. labelling or closing (#940).
GitHub Actions is the preferred platform for continuous integration,
because that is what the tidyverse team currently uses and
maintains. Functions related to Travis-CI and AppVeyor are
soft-deprecated to raise awareness about this change and to make it
clear that, if substantial maintenance becomes necessary, we may elect
to retire the function (#1169).
`browse_github_actions()` is a new function to open the Actions page
of the respective repo on GitHub, similar to existing `browse_*()`
functions (@pat-s, #1102).
`use_github_pages()` is a new function to activate or reconfigure the
GitHub Pages site associated with a repository (#224).
`use_tidy_pkgdown()` implements the complete pkgdown configuration
used by the tidyverse team (#224).
`pr_sync()` is defunct and can be replicated by calling `pr_pull()`,
`pr_merge_main()`, then `pr_push()`.
## Licensing improvements
All `use_*_license()` functions now work for projects, not just
packages.
`use_apl2_license()` (not `use_apache_license()`) and
`use_gpl3_license()` no longer modify the license text (#1198).
`use_mit_license()` now sets the default copyright holder to
"{package} authors". This makes it more clear that the copyright
holders are the contributors to the package; unless you are using a
CLA there is no one copyright holder of a package (#1207).
New `use_gpl_license()` and `use_agpl_license()` make it easier to
pick specific versions of the GPL and AGPL licenses, and to choose
whether or not you include future versions of the license. Both
default to version 3 (and above).
New `use_proprietary_license()` allows your package to pass R CMD
check while making it clear that your code is not open source
(#1163). Thanks to @atheriel for the blog post suggesting the wording:
https://unconj.ca/blog/copyright-in-closed-source-r-packages-the-right-way.html
`use_lgpl_license()` now uses version 3 (and above), and gains new
`version` and `include_future` argument to control which version is
used.
`use_gpl3_license()`, `use_agpl3_license()` and `use_apl2_license()`
have been deprecated in favour of the new `version` argument to
`use_gpl_license()`, `use_agpl_license()` and `use_apache_license()`.
The `name` argument to `use_mit_license()` has been changed to
`copyright_holder` to make the purpose more clear. The `name` argument
has been removed from all other license functions because it is not
needed; no other license makes an assertion about who the copyright
holder is.
## RStudio preferences
usethis is now fully cognizant of the [changes to RStudio
preferences](https://blog.rstudio.com/2020/02/18/rstudio-1-3-preview-configuration/)
in RStudio 1.3:
`edit_rstudio_snippets()` looks in the new location, and if you have
snippets in the old location, will automatically copy them to the new
location (#1204)
New `edit_rstudio_prefs()` opens RStudio preferences file for editing
(#1148).
`use_blank_slate()` can now configure your global, i.e. user-level,
RStudio preference, in addition to project-level (#1018).
## Other changes
`browse_package()` and `browse_project()` are new functions that let
the user choose from a list of URLs derived from local Git remotes and
DESCRIPTION (local or possibly on CRAN) (#1113).
The legacy `"devtools.desc"` option is no longer consulted when
populating a new DESCRIPTION file. You must use the
`"usethis.description"` now (#1069).
`use_dev_package()` gains a `remote` parameter to allow you to specify
the remote. The existing behaviour, which adds an `OWNER/REPO` GitHub
remote, remains the default (#918, @ijlyttle).
`use_cpp11()` is a new function to set up an R package to use cpp11.
`create_package(roxygen = FALSE)` once again writes a valid NAMESPACE
file (and also has no Roxygen* fields in DESCRIPTION) (#1120).
`create_package()`, `create_project()`, `create_from_github()`, and
`proj_activate()` work better with relative paths, inside and outside
of RStudio (#1122, #954).
`use_testthat()` gains an edition argument to support testthat v3.0.0
(#1185)
`use_version()` now updates `src/version.c` if it exists and contains
a line matching `PKG_version = "x.y.z";`.
usethis has been re-licensed as MIT (#1252, #1253).
## Dependency changes
New Imports: gert, jsonlite (was already an indirect dependency),
lifecycle, rappdirs
No longer in Imports: git2r, rematch2
|
|
|
|
|
|
|
|
Changelog:
1.75.0
New Libraries
* JSON: JSON parsing, serialization, and DOM in C++11, from Vinnie Falco and
Krystian Stasiowski.
+ Fast compilation requiring only C++11
+ Easy and safe modern API with allocator support
+ Compile without Boost, define BOOST_JSON_STANDALONE
+ Optional header-only, without linking to a library
* LEAF: A lightweight error-handling library for C++11, from Emil Dotchevski.
+ Small single-header format, no dependencies.
+ Designed for maximum efficiency ("happy" path and "sad" path).
+ No dynamic memory allocations, even with heavy payloads.
+ O(1) transport of arbitrary error types (independent of call stack
depth).
+ Can be used with or without exception handling.
* PFR: Basic reflection without macro or boilerplate code for user defined
types, from Antony Polukhin.
Updated Libraries
* Asio:
+ Enabled support for UNIX domain sockets on Windows.
+ Added executor-converting construction and assignment to ip::
basic_resolver.
+ Added compatibility between polymorphic executors and the (deprecated)
handler invocation hook.
+ Added the experimental::as_single completion token adapter.
+ Added support for MSG_NOSIGNAL on more platforms by using
_POSIX_VERSION to detect whether it is supported.
+ Added the ability to compile using libpthread on Windows.
+ Added workarounds for the Intel C++ compiler.
+ Added more support for detecting and optimising for handlers that have
no custom executor.
+ Reduced lock contention for timer cancellation on Windows.
+ Reinstated a previously removed null-pointer check, as it had a
measurable impact on performance.
+ Fixed the executor concept to test for a const-qualified execute().
+ Fixed any_executor support for builds without RTTI support.
+ Fixed the thread_pool unit test to work without RTTI support.
+ Fixed C++20 coroutines compatibility with clang on Windows.
+ Fixed some compatibility issues with Windows Runtime.
+ Fixed shadow name warnings caused by addition of asio::query.
+ Fixed a "logical ‘or’ of equal expressions" warning on linux.
+ Fixed a benign switch fallthrough warning.
+ Added missing push/pop_options.hpp includes.
+ Suppressed zero-as-null-pointer-constant warnings.
+ Fixed a comma-operator warning.
+ Updated the documentation to clarify when the select reactor is used on
Windows.
+ Fixed potential ambiguity caused by any_executor comparisons and
conversion.
+ Added detection of non-experimental C++20 coroutines on MSVC 19.8.
+ Fixed compatibility with uClibc.
+ Fixed strand<> adaptation of Networking TS executors when targeting
older C++ versions or less conformant compilers.
+ Consult the Revision History for further details.
* Atomic:
+ Implemented SSE2 and SSE4.1 versions of address lookup algorithm, which
is used in the internal lock pool implementation. This may improve
performance of waiting and notifying operations in heavily contended
cases.
+ Fixed a possible compilation error on AArch64 targets caused by
incorrect instructions generated for bitwise (logical) operations with
immediate constants. (#41)
* Beast:
+ This update brings bug fixes and support for the
BOOST_ASIO_ENBALE_HANDLER_TRACKING compile flag from Boost.Asio:
+ We'd love to know how you or your company use Beast, consider adding an
entry to the Companies and Individuals Using Beast list.
+ See the full Release Notes for a complete list of changes.
* Container:
+ New devector container.
+ Fixed bugs/issues:
o #152 Tree-based containers have troubles with move-only types.
o #156 Compile error with vector.
o PR#157 Add missing include.
o #159: pmr::monotonic_buffer_resource crashes on large single
allocations.
o #160: Usage of uses_allocator needs a remove_cvref_t.
o #162: small_vector on MSVC x86 call-by-value crash.
o #161: polymorphic_allocator(memory_resource*) non-standard
extension causes headache.
o PR#163: container_rebind for small_vector with options.
o #165: Link error with shared library and memory_resource inline
members.
o PR#166: Fix encoding error in copyright headers.
o PR#167: error: the address of 'msg' will always evaluate as 'true'
warning with GCC 4.4.
o #169: flood of warnings building dlmalloc_ext_2_8_6.c on clang11.
* Endian:
+ endian_arithmetic no longer inherits from endian_buffer
+ When BOOST_ENDIAN_NO_CTORS is defined, the unaligned endian_buffer and
endian_arithmetic are C++03 PODs, to enable use of __attribute__((
packed))
* Filesystem:
+ New: Added creation_time operation, which allows to obtain file
creation time. (Inspired by PR#134)
+ The returned value of last_write_time(p, ec) operation in case of
failure has been changed to a minimal value representable by std::
time_t instead of -1.
+ The returned value of hard_link_count(p, ec) operation in case of
failure has been changed to static_cast<uintmax_t>(-1) instead of 0.
+ On POSIX systems, file_size will now indicate error code errc::
function_not_supported if the path resolves to a non-regular file.
Previously, errc::operation_not_permitted was reported.
+ On Linux, many operations now use statx system call internally, when
possible, which allows to reduce the amount of information queried from
the filesystem and potentially improve performance. The statx system
call was introduced in Linux kernel 4.11.
+ Removed const-qualification from return types of some path methods.
This could prevent move construction and move assignment at the call
site in some cases. (#160)
+ On OpenBSD 4.4 and newer, use statvfs system call to obtain filesystem
space information. (Inspired by PR#162)
+ On Windows, space now returns with an error if the provided path does
not idendify an existing file. (#167)
* GIL:
+ BREAKING: In next release, we are going to drop support for GCC 5. We
may also change the required minimum C++ version from C++11 to C++14.
* Histogram:
+ This update brings
o Bug-fixes for corner-cases
o Small documentation improvements
o Fixes for new warnings from latest compilers and when compiling
against the C++20 standard
+ See the full Release Notes for a complete list of changes.
* Interprocess:
+ Fixed bugs:
o #127: static assertion failure with boost interprocess 1.74 and
basic_managed_shared_memory.
* Intrusive:
+ Fixed bugs:
o PR#48: MSVC "assignment within conditional" warning fix.
o PR#49: Fix empty control statement warnings.
o #52: Invalid casting in BOOST_INTRUSIVE_BSR_INTRINSIC.
* Log:
+ Bug fixes:
o Corrected the file counter that would be used in text_file_backend
when generating the target file name (based on the pattern set by
set_target_file_name_pattern method) when the log file is rotated.
(#125)
o Replaced a volatile version counter in basic_sink_frontend with an
atomic. (#128)
o In the asynchronous_sink frontend, resolved a possible conflict
between flush and run methods, if run is called from a user's
thread instead of the internal dedicated thread spawned by the
frontend. (#131)
+ See changelog for more details.
* Move:
+ Fixed bugs:
o #30: (void) C-cast is a non-portable way of suppressing compiler
warnings.
* Mp11:
+ Added mp_pairwise_fold (suggested by Barry Revzin)
+ Removed mp_invoke (use mp_invoke_q)
* Optional:
+ boost::none is constexpr-declared.
+ Fixed issue #78.
* Outcome:
+ Announcements:
o After a year and three major Boost releases announcing this
upcoming change, this is the FINAL RELEASE of the v2.1 branch. From
Boost 1.76 onwards, the v2.2 branch becomes the default. This
branch has a number of major breaking changes to Outcome v2.1, see
the documentation for details.
+ Enhancements:
o The ADL discovered event hooks have been replaced with
policy-specified event hooks instead. This is due to brittleness
(where hooks would quietly self-disable if somebody changed
something), compiler bugs (a difference in compiler settings causes
the wrong hooks, or some but not all hooks, to get discovered), and
end user difficulty in using them at all. The policy-specified
event hooks can be told to default to ADL discovered hooks for
backwards compatibility: set OUTCOME_ENABLE_LEGACY_SUPPORT_FOR to
less than 220 to enable emulation.
o Improve configuring OUTCOME_GCC6_CONCEPT_BOOL. Older GCCs had
boolean based concepts syntax, whereas newer GCCs are standards
conforming. However the precise logic of when to use legacy and
conforming syntax was not well understood, which caused Outcome to
fail to compile depending on what options you pass to GCC. The new
logic always uses the legacy syntax if on GCC 8 or older, otherwise
we use conforming syntax if and only if GCC is in C++ 20 mode or
later. This hopefully will resolve the corner case build failures
on GCC.
+ Bug fixes:
o Boost.Outcome should now compile with BOOST_NO_EXCEPTIONS defined.
Thanks to Emil, maintainer of Boost.Exception, making a change for
me, Boost.Outcome should now compile with C++ exceptions globally
disabled. You won't be able to use boost::exception_ptr as it can't
be included if C++ exceptions are globally disabled.
o #236 In the Coroutine support the final_suspend() was not noexcept,
despite being required to be so in the C++ 20 standard. This has
been fixed, but only if your compiler implements noop_coroutine.
Additionally, if noop_coroutine is available, we use the much more
efficient coroutine handle returning variant of await_suspend()
which should significantly improve codegen and context switching
performance.
* Polygon:
+ C++20 fixes for event_comparison_type, vertex_equality_predicate_type,
and voronoi_predicates. (Glen Fernandes)
* Preprocessor:
+ When variadic data is empty in C++20 mode with __VA_OPT__ support the
variadic size has been corrected to be 0. This also means that in this
C++20 mode it is now valid to convert to and from empty arrays and
lists and variadic data. The end-user can read the "C++20 Support For
Variadic Macros" part of the "variadic macros" topic for more
information about empty variadic data in the library.
+ The macro BOOST_PP_IS_STANDARD() has been added for identifying if the
currently used preprocessor is a C++ standard conforming preprocessor.
A number of preprocessors which generally work correctly with the
library but need various internal workarounds, including the currently
default VC++ preprocessor, are not considered C++ standard conforming
preprocessors. However most preprocessors, including among others gcc,
clang, and the new but currently non-default VC++ preprocessor in
VS2019, are C++ standard conforming preprocessors.
+ For C++ standard conforming preprocessors a number of the limits
defined in the config/limits.hpp can now be changed to higher amounts
for a TU. The end-user should read the "limitations" topic to
understand how and which limits can be changed.
+ For C++ standard conforming preprocessors, in order to allow the
maximum number of FOR and WHILE iterations, the beginning 'r' and 'd'
iteration numbers in the user-defined macros start at 1 and not 2, as
it did in previous releases. This could be a breaking change if these
iteration numbers are used in the user-defined macros ( they probably
would not be ), but the change was necessary to fix some arcane bugs
when dealing with numerical/logical operations with maximum numbers as
well to allow the user-defined macros to be called the correct possible
maximum number of times. For non-C++ conforming preprocessors, this
change was not made because those non-conforming C++ preprocessors
generally have limitations which disallow the maximum number of looping
constructs to be run, and it was felt not to introduce a possible
breaking change to those more fragile preprocessors would be better. It
was also felt that besides fixing some arcane preprocessor bugs and
providing the possible maximum number of user-defined macro
invocations, this change could be made because it has never been
documented what the starting 'r' and 'd' iteration numbers actually are
but only that these numbers are incremented for each iteration.
+ The library has been upgraded to assume variadic macro support for any
compiler working with the library. Ostensibly this means that the
library is now a C++11 on up library, yet most of the major compilers,
including gcc, clang, and VC++, also support variadic macros in C++98/
C++03 mode as long as strict compliance to C++98/C++03 is not turned on
when using one of those compilers.
* Rational:
+ Fix Rational operators to not break under new C++20 operator==
rewriting rules. (Glen Fernandes)
* Signals2:
+ Correct C++ allocator model support to fix compilation in C++20
standards mode. (Glen Fernandes)
* System:
+ The platform-specific headers windows_error.hpp, linux_error.hpp, and
cygwin_error.hpp emit deprecation messages and are slated for removal.
+ The old names for generic_category() and system_category() emit
deprecation messages and are slated for removal.
+ error_condition::failed is deprecated and is slated for removal.
operator bool() for error_condition has been reverted to its old
meaning of value() != 0. This is done for compatibility with std::
error_condition as the next release is expected to improve
interoperability with <system_error> even further. Note that this does
not affect error_code::failed, which is still alive and well.
+ The overload of error_condition::message that takes a buffer is
deprecated and is slated for removal, for the same reasons. Note that
this does not affect error_code::message.
* uBLAS:
+ Correct C++ allocator model support to fix compilation in C++20
standards mode. (Glen Fernandes and Conrad Poelman)
* VMD:
+ The VMD number parsing has been upgraded to support the ability for the
end-user to change the number limits in the Preprocessor library.
+ The macro BOOST_VMD_IS_GENERAL_IDENTIFIER has been added to support the
parsing of input that represents a preprocessor token which matches the
VMD identifier syntax, without having to register the identifier as a
specific identifier.
* Wave:
+ Added new C++20 tokens, including the spaceship operator <=>
+ Fixed bugs:
o #94: fix incorrect behavior of __LINE__ and __FILE__ under
rescanning
1.74.0
New Libraries
* STLInterfaces: A library of CRTP bases to ease the writing of STL views,
iterators, and sequence containers, from Zach Laine.
Updated Libraries
* Asio:
+ Added an implementation of the proposed standard executors (P0443r13,
P1348r0, and P1393r0).
+ Added support for the proposed standard executors to Asio's I/O
facilities.
o The supplied executors now meet the requirements for the proposed
standard executors. These classes also continue to meet the
existing requirements for the Networking TS model of executors.
o All I/O objects, asynchronous operations, and utilities will
interoperate with both new proposed standard executors, and with
existing Networking TS executors.
o The any_io_executor type alias has been introduced as the default
runtime-polymorphic executor for all I/O objects. This defaults to
the execution::any_executor<> template. If required for backward
compatibility, BOOST_ASIO_USE_TS_EXECUTOR_AS_DEFAULT can be defined
to use the old asio::executor polymorphic wrapper instead.
o Support for the existing Networking TS model of executors can be
disabled by defining BOOST_ASIO_NO_TS_EXECUTORS.
+ Added converting move construction and assignment to
basic_waitable_timer.
+ Enabled C++20 coroutine support when using gcc 10.
+ Added overloads of co_spawn that launch an awaitable.
+ Added a new constructor overload to use_awaitable_t's default executor
adapter, to enable conversion between executor types.
+ Added support for using detached_t as a default completion token, by
adding members as_default_on() and as_default_on_t<>.
+ Added a move constructor to ssl::stream<>.
+ Changed ssl::stream<> write operations to linearise gather-write buffer
sequences.
+ Added compile-time detection of the deprecated asio_handler_invoke,
asio_handler_allocate, and asio_handler_deallocate hooks, when
BOOST_ASIO_NO_DEPRECATED is defined.
+ Implemented a number of performance optimisations.
+ Added source location support to handler tracking.
+ Implemented various improvements to the handlerviz.pl tool.
+ Added the handlerlive.pl tool, which processes handler tracking output
to produce a list of "live" handlers.
+ Added the handlertree.pl tool, which filters handler tracking output to
include only those events in the tree that produced the nominated
handlers.
+ Added changes for clang-based Embarcadero C++ compilers.
+ Fixed a deadlock that can occur when multiple threads concurrently
initialise the Windows I/O completion port backend.
+ Fixed async_compose to work with copyable handlers when passed by
lvalue.
+ Fixed completion signature deduction in co_spawn.
+ Removed a spurious Executor base class from the executor_binder
implementation.
+ Various fixes and improvements in the documentation and examples.
+ Consult the Revision History for further details.
* Atomic:
+ Added missing const qualifiers to some operations in atomic_ref.
+ Added support for yield instruction on ARMv8-A. The instruction is used
internally in spin loops to reduce CPU power consumption.
+ Added support for C++20 waiting and notifying operations. The
implementation includes generic backend that involves the internal lock
pool, as well as specialized backends for Windows, Linux, FreeBSD,
DragonFly BSD and NetBSD. Atomic types provide a new method
has_native_wait_notify, a static boolean constant
always_has_native_wait_notify and a set of capability macros that allow
to detect if the implementation supports native waiting and notifying
operations for a given type.
+ Changed internal representation of atomic_flag to use 32-bit storage.
This allows for more efficient waiting and notifying operations on
atomic_flag on some platforms.
+ Added support for build-time configuration of the internal lock pool
size. The user can define the BOOST_ATOMIC_LOCK_POOL_SIZE_LOG2 macro to
specify binary logarithm of the size of the lock pool. The default
value is 8, meaning that the size of the lock pool is 256, up from 64
used in the previous release.
+ Added support for a new set of atomic types dedicated for inter-process
communication: ipc_atomic_flag, ipc_atomic and ipc_atomic_ref. Users
are recommended to port their code using non-IPC types for
inter-process communication to the new types. The new types provide the
same set of operations as their non-IPC counterparts, with the
following differences:
o Most operations have an added precondition that is_lock_free
returns true for the given atomic object. The library will issue a
compile time error if this precondition is known to be not
satisfied at compile time.
o All provided operations are address-free, meaning that the atomic
object (in case of ipc_atomic_ref - the referenced object) may be
located in process-shared memory or mapped into the same process at
multiple different addresses.
o The new has_native_wait_notify operation and
always_has_native_wait_notify constant indicate support for native
inter-process waiting and notifying operations. When that support
is not present, the operations are implemented with a busy loop,
which is less efficient, but still is address-free. A separate set
of capability macros is also provided to indicate this support.
+ Added new atomic_unsigned_lock_free and atomic_signed_lock_free types
introduced in C++20. The types indicate the atomic object type for an
unsigned or signed integer, respectively, that is lock-free and
preferably has native support for waiting and notifying operations.
+ Added new gcc assembler backends for ARMv8-A (for both AArch32 and
AArch64). The new backends are used to implement operations not
supported by compiler intrinsics (including 128-bit operations on
AArch64) and can also be used when compiler intrinsics are not
available. Both little and big endian targets are supported. AArch64
backend supports extensions defined in ARMv8.1 and ARMv8.3.
+ Added support for big endian targets in the legacy ARM backend based on
gcc assembler blocks (this backend is used on ARMv7 and older targets).
Previously, the backend assumed little endian memory layout, which is
significant for 64-bit operations.
+ Improved performance of seq_cst stores and thread fences on x86 by
using lock-prefixed instructions instead of mfence. This means that the
operations no longer affect non-temporal stores, which was also not
guaranteed before. Use specialized instructions and intrinsics to order
non-temporal memory accesses.
+ Fixed capability macros for 80-bit long double on x86 targets not
indicating lock-free operations even if 128-bit atomic operations were
available.
+ Fixed compilation of gcc asm blocks on Alpha targets.
+ In the gcc __sync* intrinsics backend, fixed that store and load
operations of large objects (larger than a pointer size) could be
non-atomic. The implementation currently assumes that small objects can
be stored with a single instruction atomically on all modern
architectures.
* Beast:
+ This update brings bug fixes and support for the following changes
changes in Boost.Asio:
+ Beast supports BOOST_ASIO_NO_DEPRECATED. Define this to help identify
areas of your Beast and Asio code which use deprecated Asio interfaces.
+ Beast also supports BOOST_ASIO_NO_TS_EXECUTORS. Define this to identify
uses of executors from the now potentially outdated Networking TS
+ Asio will use the Standard Executors model by default. You can prevent
this behaviour by defining BOOST_ASIO_USE_TS_EXECUTOR_AS_DEFAULT in
which the Networking TS model will be used by default. Setting this
flag does not prevent a program from using executors from the Standard
Executors model explicitly.
+ We'd love to know how you or your company use Beast, consider adding an
entry to the Companies and Individuals Using Beast list.
+ See the full Release Notes for a complete list of changes.
* Bimap:
+ Correct allocator usage (fixes C++20 compilation). (Glen Fernandes)
* Config:
+ Implement BOOST_NO_CXX11_OVERRIDE and BOOST_OVERRIDE. (Glen Fernandes)
* Core:
+ Implemented the allocator access utilities which provide a replacement
for allocator_traits with individual traits and functions for each
facility. They support the C++11 allocator model when possible and
provide a fallback for C++98 compatibility. These are now used in
Circular_Buffer, Wave, Lockfree, Heap, Smart_Ptr, Dynamic_Bitset,
Format, Bimap and more. (Glen Fernandes)
* DLL:
+ Multiple fixes for the library_info work on empty shared objects.
+ Compilation fixes for C++98 and C++11 modes (#28).
+ Fixes for smart_library manglings (thanks to XiaLiChao82 #37).
* Endian:
+ Enabled scoped enumeration types in endian_reverse.
+ Enabled bool, enum, float, double in endian_reverse_inplace.
+ Added an overload of endian_reverse_inplace for arrays.
* Filesystem:
+ Removed compile-time checks for support for symlinks and hardlink on
Windows. Instead, a runtime check is used. (PR#142)
+ Fixed handling of reparse points in canonical and read_symlink on
Windows. This also affects other algorithms that involve canonical and
read_symlink in their implementation. (PR#100, #85, #99, #123, #125)
+ Fixed that read_symlink on Windows could potentially fail or cause
failures elsewhere with a sharing violation error, if the same symlink
was opened concurrently. (#138)
+ Fixed that is_symlink(directory_entry) would always return false, even
if the directory entry actually referred to a symlink. (PR#148)
+ Added missing status inspection operation overloads for directory_entry
and error_code (e.g. is_directory(directory_entry, error_code&)).
Removed incorrect noexcept specifications for the overloads not taking
the error_code arguments.
+ copy_file implementation has been updated to perform checks on the
source and target files, as required by C++20 ([fs.op.copy.file]/4.1).
In particular, the operation will fail if the source or target file is
not a regular file or the source and target paths identify the same
file.
+ copy_file on POSIX systems will now also copy the source file
permissions to the target file, if the target file is overwritten.
+ New: Added copy_file implementations based on sendfile and
copy_file_range system calls on Linux, which may improve file copying
performance, especially on network filesystems.
+ Deprecated: The copy_option enumeration that is used with the copy_file
operation is deprecated. As a replacement, the new enum copy_options
(note the trailing 's') has been added. The new enum contains values
similar to the copy_options enum from C++20. The old enum values are
mapped onto the new enum. The old enum will be removed in a future
release.
+ New: Added copy_options::skip_existing option, which allows copy_file
operation to succeed without overwriting the target file, if it exists.
+ New: Added copy_options::update_existing option, which allows copy_file
operation to conditionally overwrite the target file, if it exists, if
its last write time is older than that of the replacement file.
+ New: copy_file now returns bool, which indicates whether the file was
copied.
+ New, breaking change: copy operation has been extended and reworked to
implement behavior specified in C++20 [fs.op.copy]. This includes
support for copy_options::recursive, copy_options::copy_symlinks,
copy_options::skip_symlinks, copy_options::directories_only,
copy_options::create_symlinks and copy_options::create_hard_links
options. The operation performs additional checks based on the
specified options. Applying copy to a directory with default
copy_options will now also copy files residing in that directory (but
not nested directories or files in those directories).
+ New: Added create_directory overload taking two paths. The second path
is a path to an existing directory, which is used as a source of
permission attributes to use in the directory to create.
+ Deprecated: copy_directory operation has been deprecated in favor of
the new create_directory overload. Note that the two operations have
reversed order of the path arguments.
+ equivalent on POSIX systems now returns the actual error code from the
OS if one of the paths does not resolve to a file. Previously the
function would return an error code of 1. (#141)
+ equivalent no longer considers file size and last modification time in
order to test whether the two paths refer to the same file. These
checks could result in a false negative if the file was modified during
the equivalent call.
+ New: Added absolute overloads taking error_code argument.
+ Operations that have current_path() as the default value of their
arguments and also have an error_code argument will use the
current_path(error_code& ec) overload to obtain the current path, so
that its failure is reported via the error_code argument instead of an
exception.
+ space now initializes the space_info structure members to -1 values on
error, as required by C++20 ([fs.op.space]/1).
+ space on Windows now accepts paths referring to arbitrary files, not
only directories. This is similar to POSIX systems and corresponds to
the operation description in C++20. (#73)
+ New: Added implementation of temp_directory_path for Windows CE. (PR#25
)
+ New: Improved compatibility with WASI platform. (PR#144)
+ New: Improved support for Embarcadero compilers. (PR#130)
+ New: Added implementations of unique_path operation based on getrandom
(Linux), arc4random_buf (OpenBSD/FreeBSD/CloudABI) and BCrypt (Windows)
system APIs.
+ Deprecated: Auto-linking against system libraries on Windows with
MSVC-compatible compilers is deprecated and will be removed in a future
release. This affects users linking against static library of
Boost.Filesystem. Users are advised to update their project build
systems to either use a shared library of Boost.Filesystem, or
explicitly specify the dependencies of Boost.Filesystem in the linker
command line. Users of shared library of Boost.Filesystem are not
affected.
* Flyweight:
+ Maintenance work.
* Format:
+ Correct allocator usage (fixes C++20 compilation). (Glen Fernandes)
* Geometry:
+ Improvements
o PR#720 Additional R-tree constructors (thanks to Caian Benedicto).
o Various improvements in buffer, set and relational operations.
+ Solved issues
o #709 memcpy called for object with no trivial copy-assignment.
o #721 Compilation error in bgi::detail::rtree::visitors::insert.
o #727 MSVC warning: conditional expression is constant.
+ Bugfixes
o PR#700 Missing cases for default strategies in distance algorithm.
o PR#738 Longitudes out of range in direct geodesic formulas.
* GIL:
+ Added
o Added new constructor initializing any_image from r-value reference
to any image (PR#486).
o Implemented mechanism to reverse kernel_2d (PR#489).
+ Changed
o BREAKING: Replace Boost.Variant with Boost.Variant2 (PR#474) which
completes removal on uses of Boost.MPL (missing from Boost 1.72.0
change added PR#274).
o Use perfect forwarding from apply_operation to visit (PR#491).
+ Removed
o BREAKING: Removed dependency on Boost.Variant
+ Fixed
o Fixed invalid conversion from RGB8 to CMYK32 due to overflow (PR#
470).
o Fixed image constructor from other image (PR#477).
o Fixed error plane_view_t is not a class or namespace name (PR#481).
o Fixed interleaved_view factory using point<std::ptrdiff_t> for
dimension (PR#487).
o Fixed documentation replacing uses MPL with MP11 in tutorial (PR#
494).
o Fixed missing header in numeric/kernel.hpp to make it
self-contained (PR#502).
+ Acknowledgements
o Samuel Debionne, Pranam Lashkari, Mateusz Loskot, Debabrata Mandal
* Heap:
+ Correct destruction of top node in skew_heap. (Glen Fernandes)
+ Correct and simplify allocator use. (Glen Fernandes)
* Integer:
+ Fixed compilation of gcd in C++20 mode with clang 10.
+ Improved support for Embarcadero compilers. (PR#21)
* Iterator:
+ boost/function_output_iterator.hpp header is now deprecated. Users
should replace its inclusion with boost/iterator/
function_output_iterator.hpp. (PR#51)
+ Improved support for Embarcadero compilers. (PR#55)
* LexicalCast:
+ Fixed warnings on missing override (thanks to EugeneZelenko #35, #34).
+ Fixes for the the Embarcadero compilers (thanks to Edward Diener).
* Log:
+ Bug fixes:
o The syslog sink backend now verifies the IP version of the local
and target addresses set by user. The addresses must have the same
IP version as was specified in the ip_version named parameter on
the sink backend construction (by default, IPv4 is assumed). When
an address is obtained as a result of host name resolution, only
addresses with matching IP version are considered. (#119)
+ New Features:
o Move constructors and assignment operators of various components
were marked noexcept.
o Added a new range_manip stream manipulator that can be used for
outputting elements of a range, optionally separated by a
delimiter.
o Added a new tuple_manip stream manipulator that can be used for
outputting elements of a tuple or any other heterogeneous sequence,
optionally separated by a delimiter.
o Added a new optional_manip stream manipulator that can be used for
outputting optionally present values.
+ See changelog for more details.
* Mp11:
+ Improved compilation performance of mp_with_index<N> for large N.
+ Added tuple_transform (contributed by Hans Dembinski.)
* Multi-index Containers:
+ Added node extraction and insertion following the analogous interface
of associative containers as introduced in C++17. This feature has also
been extended to non key-based indices, in contrast to C++ standard
library sequence containers, which do not provide such functionality.
+ Clarified documentation on read/write key extractors (issue #32).
+ Maintenance work.
* Nowide:
+ The library now requires a C++11-compliant compiler and stdlib
+ LFS: Add support for files > 2 GB where the underlying system supports
it
+ Generic UTF conversion functions are now available in the boost::nowide
::utf namespace
+ Add support for stat with UTF-8 paths
* Outcome:
+ Announcements:
o The v2.1 branch is expected to be retired end of 2020, with the
v2.2 branch becoming the default. You can use the future v2.2
branch now using better_optimisation. This branch has a number of
major breaking changes to Outcome v2.1, see the front page for
details.
+ Enhancements:
o BREAKING CHANGE void results and outcomes no longer default
construct types during explicit construction. Previously if you
explicitly constructed a result<T> from a non-errored result<void>,
it default constructed T. This was found to cause unhelpful
surprise, so it has been disabled.
o New macro OUTCOME_ENABLE_LEGACY_SUPPORT_FOR. The macro
OUTCOME_ENABLE_LEGACY_SUPPORT_FOR can be used to enable aliasing of
older naming and features to newer naming and features when using a
newer version of Outcome.
o Concepts now have snake case style naming instead of camel case
style. When Outcome was first implemented, it was thought that C++
20 concepts were going to have camel case style. This was changed
before the C++ 20 release, and Outcome's concepts have been renamed
similarly. This won't break any code in Outcome v2.1, as
compatibility aliases are provided. However code compiled against
Outcome v2.2 will need to be upgraded, unless
OUTCOME_ENABLE_LEGACY_SUPPORT_FOR is set to 210 or lower.
o Concepts now live in OUTCOME_V2_NAMESPACE::concepts namespace.
Previously concepts lived in the convert namespace, now they live
in their own namespace.
o New concepts basic_result<T> and basic_outcome<T> added. End users
were finding an unhelpful gap in between is_basic_result<T> and
value_or_error<T> where they wanted a concept that matched types
which were basic_result, but not exactly one of those. Concepts
filling that gap were added.
o Operation TRY works differently from Outcome v2.2 onwards. This is
a severely code breaking change which change the syntax of how one
uses OUTCOME_TRY(). A regular expression suitable for upgrading
code can be found in the list of changes between Outcome v2.1 and
v2.2.
+ Bug fixes:
o #224 The clang Apple ships in Xcode 11.4 (currently the latest) has
not been patched with the fixes to LLVM clang that fix noexcept(std
::is_constructible<T, void>) failing to compile which I originally
submitted years ago. So give up waiting on Apple to fix their
clang, add a workaround to Outcome.
o Spare storage could not be used from within no-value policy
classes. Due to an obvious brain fart when writing the code at the
time, the spare storage APIs had the wrong prototype which
prevented them working from within policy classes. Sorry.
* PolyCollection:
+ Fixed internal ambiguity problem between boost::type_erasure::any and
boost::any (issue #17).
+ Maintenance work.
* SmartPtr:
+ Added owner_equals to shared_ptr, weak_ptr, local_shared_ptr.
+ Added owner_hash_value to shared_ptr, weak_ptr.
+ Added owner_equal_to, owner_hash.
+ Added std::hash specializations for shared_ptr, local_shared_ptr.
+ Added boost::hash support to, and std::hash, std::equal_to
specializations for, weak_ptr.
* Stacktrace:
+ Fixed a build error when compiled with -fno-exceptions (thanks to
Jeremiah Rodriguez #91).
* System:
+ operator bool() now returns failed() instead of value() != 0.
* Type_Traits:
+ Implemented conjunction, disjunction, negation, is_trivially_copyable,
is_scoped_enum, and is_unscoped_enum. (Glen Fernandes)
* Variant:
+ Fixed warnings on missing override (thanks to EugeneZelenko #78).
+ Fixes for the the Embarcadero compilers (thanks to Edward Diener #79).
+ Updated header locations to avoid warnings about using deprecated
headers (thanks to Andrey Semashev #80)
* Variant2:
+ Added support for derived types in visit.
+ Improved compilation performance for many (hundreds of) alternatives.
+ Added support for visit<R>.
* Wave:
+ Implement C++20 features for variadics, including __VA_OPT__ (PR#75)
+ Implement __has_include (PR#102)
+ Introduce new sample: check_macro_naming, useful with Boost itself (PR#
97)
+ Fix compilation issue caused by std::allocator member removal in C++20
(PR#72)
+ Repair Xpressive lexer and token_statistics sample (PR#79)
+ Repair lexertl lexer (PR#78)
+ Ensure hooks are run on predefined macros as well (PR#87)
+ Various minor bug fixes
+ C++98/03 support is now deprecated and will be removed in 1.77
* YAP:
+ Fixed compilation errors for placeholders; they now work in the general
case, and in particular work with yap::print().
+ constexpr all the YAP.
+ Fix printing of correct value category in yap::print().
+ Doc clarification.
Updated Tools
* Boostbook:
+ Change encoding of generated documentation from US-ASCII to UTF-8.
(Glen Fernandes)
|
|
|
|
Updated devel/R-pkgbuild to 1.2.0
Added devel/R-gert version 1.0.2
|
|
Simple git client for R based on 'libgit2' with support for SSH and
HTTPS remotes. All functions in 'gert' use basic R data types (such as
vectors and data-frames) for their arguments and return values. User
credentials are shared with command line 'git' through the
git-credential store and ssh keys stored on disk or ssh-agent. On
Linux, a somewhat recent version of 'libgit2' is required; we provide
a 'PPA' for older Ubuntu 'LTS' versions.
|
|
(pkgsrc)
- LICENSE converted gnu-gpl-v3 to mit (as described below)
(upstream changes)
# pkgbuild 1.2.0
* pkgbuild is now licensed as MIT (#106)
* `compile_dll()` gains a `debug` argument for more control over the compile options used (@richfitz, #100)
* `pkgbuild_process()` and `build()` now use colored compiler diagnostics if supported (#102)
* Avoid documentation link ambiguity in R 4.1 (#105)
|
|
# testthat 3.0.1
* New `testthat.progress.verbose_skips` option. Set to `FALSE` to stop
reporting skips as they occur; they will still appear in the summary
(#1209, @krlmlr).
* `CheckReporter` results have been tweaked based on experiences from running
R CMD check on many packages. Hopefully it should now be easier to see
the biggest problems (i.e. failures and errors) while still having
skips and warnings available to check if needed (#1274). And now the full
test name is always shown, no matter how long (#1268).
* Catch C++ tests are no longer reported multiple times (#1237) and
are automatically skipped on Solaris since Catch is not supported (#1257).
`use_catch()` makes it more clear that your package needs to suggest
xml2 (#1235).
* `auto_test_package()` works once again (@mbojan, #1211, #1214).
* `expect_snapshot()` gains new `error` argument which controls whether or not
an error is expected. If an unexpected error is thrown, or an expected error
is not thrown, `expect_snapshot()` will fail (even on CRAN) (#1200).
* `expect_snapshot_value(style = "deparse")` handles more common R data
structures.
* `expect_snapshot_value()` now passes `...` on to `waldo::compare()` (#1222).
* `expect_snapshot_file()` gives a hint as to next steps when a failure
occurs in non-interactive environments (with help from @maelle, #1179).
`expect_snapshot_*()` gives a more informative hint when you're running
tests interactively (#1226).
* `expect_snapshot_*()` automatically removes the `_snaps` directory if
it's empty (#1180). It also warns if snapshots are discarded because tests
have duplicated names (#1278, @krlmlr).
* `local_reproducible_output()` now sets the LANGUAGE env var to "en". This
matches the behaviour of R CMD check in interactive settings (#1213).
It also now unsets RSTUDIO envvar, instead of setting it to 0 (#1225).
* `RstudioReporter` has been renamed to `RStudioReporter`.
* `skip_if_not()` no longer appends "is not TRUE" to custom messages
(@dpprdan, #1247).
* `test_that()` now warns (3e only) if code doesn't have braces, since
that makes it hard to track the source of an error (#1280, @krlmlr).
|
|
# cli 2.2.0
* New `style_hyperlink()` function to add hyperlinks, on terminals that
support them.
* `cli_format_method()` now works properly in knitr, and other environments
that catch message conditions (#159).
* ANSI strings created by `col_*`, `bg_*` and `style_*` now also add the
`character` class to the result. This fixes issues with code that
expect `character` objects.
* New functions to manipulate ANSI strings: `ansi_aling()`,
`ansi_has_any()`, `ansi_nchar()`, `ansi_regex()`, `ansi_strip()`,
`ansi_strsplit()`, `ansi_substr()`, `ansi_substring()`.
# cli 2.1.0
* New `cli_vec()` function to allow easier formatting of collapsed
vectors. It is now also possible to use styling to set the collapsing
parameters (#129).
* New `pluralize()` function to perform pluralization without generating
cli output (#155).
* `console_width()` works better now in RStudio, and also in terminals.
* Styling of verbatim text work properly now (#147, @tzakharko).
* Messages (ie. `message` conditions) coming from cli now have the
`cliMessage` class, so you can easily suppress them without suppressing
other messages (#156).
* cli prints the output to `stderr()` now, if there is an output or
message sink. This is to make interactive and non-interactive sessions
consistent (#153).
* Pluralization works corrently now if the last alternative is the
empty string (#158).
* cli now caches the result of the dark background detection in iTerm on
macOS. Reload cli to delete the cache (#131).
* The `is_dynamic_tty()`, `is_ansi_tty()` and `ansi_hide_cursor()` and
releted functions now default to the `"auto"` stream, which is
automatically selected to be either `stdout()` or `stderr()`.
See the manual for details (#144).
* The default theme now quotes file names, paths, email addresses if they
don't start or end with an alphanumeric character or a slash. This is
to make it easier to spot names that start or end with a space (#167).
* `make_spinner()` clears the line properly now (@tzakharko, #164).
* Semantic cli functions now automatically replace Unicode non-breaking
space characters (`\u00a0`) with regular space characters, right before
output. They are still used to calculate the line breaks, but not
outputted (#161).
* Progress bars now respect `is_dynamic_tty()` and do not output `\r` when this
is false (@jimhester, #177)
|
|
# rlang 0.4.9
## Breaking changes
* Dropped support for the R 3.2 series.
## New features
* `inject()` evaluates its argument with `!!`, `!!!`, and `{{`
support.
* New `enquo0()` and `enquos0()` operators for defusing function
arguments without automatic injection (unquotation).
* `format_error_bullets()` is no longer experimental. The `message`
arguments of `abort()`, `warn()`, and `inform()` are automatically
passed to that function to make it easy to create messages with
regular, info, and error bullets. See `?format_error_bullets` for
more information.
* New `zap_srcref()` function to recursively remove source references
from functions and calls.
* A new compat file for the zeallot operator `%<-%` is now available
in the rlang repository.
* New `%<~%` operator to define a variable lazily.
* New `env_browse()` and `env_is_browsed()` functions. `env_browse()`
is equivalent to evaluating `browser()` within an environment. It
sets the environment to be persistently browsable (or unsets it if
`value = FALSE` is supplied).
* Functions created from quosures with `as_function()` now print in a
more user friendly way.
* New `rlang_print_backtrace` C callable for debugging from C
interpreters (#1059).
## Bugfixes and improvements
* The `.data` pronoun no longer skips functions (#1061). This solves a
dplyr issue involving rowwise data frames and list-columns of
functions (tidyverse/dplyr#5608).
* `as_data_mask()` now intialises environments of the correct size to
improve efficiency (#1048).
* `eval_bare()`, `eval_tidy()` (#961), and `with_handlers()` (#518)
now propagate visibility.
* `cnd_signal()` now ignores `NULL` inputs.
* Fixed bug that prevented splicing a named empty vector with the
`!!!` operator (#1045).
* The exit status of is now preserved in non-interactive sessions when
`entrace()` is used as an `options(error = )` handler (#1052,
rstudio/bookdown#920).
* `next` and `break` are now properly deparsed as nullary operators.
# rlang 0.4.8
* Backtraces now include native stacks (e.g. from C code) when the
[winch](https://r-prof.github.io/winch/) package is installed and
`rlang_trace_use_winch` is set to `TRUE` (@krlmlr).
* Compatibility with upcoming testthat 3 and magrittr 2 releases.
* `get_env()` now returns the proper environment with primitive
functions, i.e. the base namespace rather than the base environment
(r-lib/downlit#32).
* `entrace()` no longer handles non-rlang errors that carry a
backtrace. This improves compatibility with packages like callr.
* Backtraces of unhandled errors are now displayed without truncation
in non-interactive sessions (#856).
* `is_interactive()` no longer consults "rstudio.notebook.executing"
option (#1031).
|
|
git2r 0.27.1
------------
CHANGES
* Fixed the CITATION file to pass 'R CMD check' without a NOTE.
git2r 0.27.0
------------
IMPROVEMENTS
* Updated the bundled libgit2 source code to version '0.28.5'.
* Updated the build configuration script to be able to build git2r
with a system installation of libgit2 version >= 1.0.
* Updated to use libgit2 version 1.0.0 on Windows.
* The build configuration script checks for minimum required version
of libssh2 (version >= 1.8). Issue #420.
* Updated to use roxygen2 version 7.1.0 to build the documentation.
* Make it easier to view and change the timezone (John Blischak in
#408).
* Fixed 'ls_tree' to handle content in subfolder, see description in
PR #402.
* The 'branch_create' function has been changed to use the
'last_commit()' function as default to determine the commit to which
the new branch should point.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Reflect Python version support noted in setup.py.
|
|
|
|
[Fixed]
- Passing multiple tags arguments to prove correctly intersects
the sets;
e.g. '--feature-option tags=@wip --feature-option tags=@daily' now
correctly runs stricttly the scenarios matching both @wip and @daily
- Tutorial.pod incorrectly stated --tags=@tag1,~@tag2 runs scenarios
tagged '@tag1' except those tagged '@tag2': it runs all tagged
'@tag1' and all *not* tagged '@tag2'.
- Clarified difference between step definitions, step models and step
execution contexts
[Added]
- Explanation in 'pherkin' how to pass tag patterns
- Expanded explanation in Tuturial.pod how to pass tag patterns
Updating during the freeze for bugfixes and documentation improvements.
|
|
|
|
|
|
|
|
|
|
|