+ Don't allow mounting of already mounted sandboxes, or unmounting of
unmounted sandboxes.
+ Also, add an extra "ismounted" case label to be able to tell if a
sandbox is mounted. A df(1) invocation should continue to work just
as well :)
No objections on tech-pkg@
Addresses the first part of PR pkg/51992 from Paul Goyette
nih remove:
- now this command works correctly without external package
repository
nih leaf:
- new option "-u" for printing user-installed leaf packages
nih status:
- option -s has been renamed to "-p"
Primary extension for pkg_summary(5) is now "gz"
Updates in man.1
nih.default.conf:
- update default URLs to package repositories
root Makefile:
- +PKG_DBDIR defaults to ${PREFIX}/pkgdb
pkg_summary2leaves:
- new option -u for printing packages installed by user
- option -r does not implies -a any longer
pkg_grep_summary:
- new option -E and "-t nonempty" for matching non-empty fields
Fixes in tests
Minor code clean-ups and coding style correction in grep_summary.c
COPTS defaults to "-O3 -g"
root Makefile: BMAKE variables defaults to "bmake"
encouragement from David Holland, and also the younger generation -
apply the fix from the PR.
Thanks for the PR, the fix, and the nudges, folks, and sorry it took so long!
Changes since 23.2.3:
Do not warn about buildlink3 identifiers that are only known when the
package includes a buildlink3.mk file from the infrastructure.
Changes since 23.2.2:
Warn about platform-specific variables in parameterized variables,
except for OPSYS-specific variables in conditions for that particular
operating system.
Allow the tilde character in lists of Go modules.
Use correct plural form in messages about invalid characters.
Pull in changes from revision 1.13 of NetBSD sha2.c from 14 years ago to
fix type punning issues seen with newer GCCs.
Fixes "pkg_admin digest" on SmartOS with GCC 12, where the output was
completely wrong, causing bulk builds to rebuild every package every time
now that USE_PKG_ADMIN_DIGEST=yes is the default.
Changes since 23.2.1:
Do not wrongly warn when adding a tool that is only available on
${OPSYS} to CONFIGURE_ENV.${OPSYS}.
Use proper capitalization for 'makefile' in a few places, when referring
to the general file format instead of a file named exactly 'Makefile'.
Changes since 23.2.0:
Warn about language features that are listed in USE_LANGUAGES but rather
belong in USE_CC_FEATURES or USE_CXX_FEATURES.
Check each addition to GITHUB_SUBMODULES to have the correct number of
words.
## Version 23.8.1 (2023-08-22)
* Fix regression introduced in 23.8.0 where only the first argument to
"keep" and "unkeep" commands was processed. Reported by gdt.
## Version 23.8.0 (2023-08-16)
This is the biggest overhaul of pkgin yet! As well as fixing many bugs, pkgin
23.8.0 runs "pkgin -n upgrade" over 2,000x faster on a test system with 12,663
packages installed compared to 22.10.0 (1.4 seconds vs over 54 minutes).
* Large overhaul of dependency resolution code, fixing many upgrade issues,
providing significant performance improvements, and improving code and
logic readability.
* Add support for SUPERSEDES.
* Fix install and removal ordering.
* Improve and fix handling of "keep" packages, using "pkg_add -A" to
correctly install automatic packages rather than applying keep or unkeep
flags afterwards.
* Output improvements, sorting packages alphabetically and providing a
simple progress counter. Fix unwanted output that could end up mixed
with package paths in "pkgin export".
* Correct handling of +REQUIRED_BY for reverse dependencies, fixing
DEPENDS that use alternate matches.
* Many internal code improvements, memory leaks plugged, and compiler
warnings eradicated.
Fixes at least the following issues:
* NetBSDfr/pkgin#10
* NetBSDfr/pkgin#44
* NetBSDfr/pkgin#79
* NetBSDfr/pkgin#95
* NetBSDfr/pkgin#97
* NetBSDfr/pkgin#110
* NetBSDfr/pkgin#113
* NetBSDfr/pkgin#128
* NetBSD PR#51465
* NetBSD PR#56683
The first 'bmake extract' is done only to look at the files from the
package, to determine the dependencies. In that stage, the package
cannot be built successfully. To make the whole workflow less
surprising, install the auto-detected dependencies before leaving the
rest of the work to the human package maintainer. This means that
running url2pkg produces more output and may take significantly more
time, but that time would be spent anyway later, so it's a net win.
Suggested by gdt@.
Changes since 23.1.13:
Warn about attempts to set CONFIGURE_ARGS that are already handled by
mk/configure/gnu-configure.mk.
Warn about possible incomplete expressions that are missing the initial
'$'.