* SHA-256 has replaced SHA-1 as a hash function used for verifying
integrity of the ports tree.
* The RSA signature verification is now performed using OpenSSL
rather than "borrowing" code from FreeBSD Update.
* In addition to distributing the ports tree, portsnap now also
produces up-to-date INDEX, INDEX-4, and INDEX-5 files.
* When fetching a series of patches, portsnap now prints a simple
progress indicator.
PR: ports/79044
Submitted by: maintainer
1) Added check for really old ports in PMGRrAddDependencies that
aborts portmanager and provides suggestion on how to update manually.
This is to fix a seg fault when @comment DEPORIGIN: is missing in
a port's +CONTENTS file
2) Incorporated Alistair Sutton's <alistair.sutton@gmail.com>
code that adds interactive update option "portmanager -ui".
3) Slowly switching licience from GNU to BSD in the sources
PR: ports/78839
Submitted by: Maintainer
1) Incorporated Florent Thoumie's <flz@xbsd.org> suggestion
of adding pkg-message so pkg users may also see it.
2) added handling of ports that conflict with ports that
are being upgraded, inspired by "Brian John" <brianjohn at fusemail.com>
3) Added percent done message to xterm title bar as requested by
<mark.magiera at gmail.com>
4) port is repackaged because too many changes to configure.ac
for a patch to be feasable
PR: 77942
Submitted by: Michael C. Shultz <ringworm01 at gmail.com>
- Change email address from Verizon to gmail so I may hear from my european
friends
PR: ports/77575
Submitted by: Michael C. Shultz <ringworm01@gmail.com> (maintainer)
1) Some code clean up using valgrind
2) Added flushing of stdout/stderr so messages are
not lost when redirecting portmanager output to a file
PR: ports/77525
Submitted by: Michael C. Shultz <reso3w83@verizon.net> (maintainer)
* Do not warn on direct command use if the command is preceeded by "--" [1]
* Add a check for ports installing charset.alias or locale.alias. In most
cases these ports should depend on libiconv or gettext, and not install
their own copies of these files [2]
PR: 77410 [2]
Requested by: gerald [1]
Submitted by: leeym [2]
* Finally found and fixed the bug haunting portmanager for the past few
revisions. Portmanager now "flushes" after droping new records into it's
data bases :)
* Improvement of database code to handle wider variety of strings/character
codes merged from another project that uses libMG.
PR: ports/77409
Submitted by: Michael C. Shultz <reso3w83@verizon.net> (maintainer)
There may be a bug in edge cases where a port name originates from multiple
usr/ports directories. Patch allows printing of useful information if this
happens and hopefully traps the bug with assert statements.
PR: ports/77276
Submitted by: Michael C. Shultz <reso3w83@verizon.net> (maintainer)
1) Fixed bug where ports were being updated for old dependencies
when they did not require it.
2) Added -sl option: show leaf ports (ports with no other ports
depending on them)
3) Added -slid option: interactivly delete leaf ports
PR: 77238
Submitted by: Michael C. Shultz <reso3w83 at verizon.net>
1) Add handling of identical portnames from differing origins in port tree
2) Minor bug fix involving portmanager's handling of missing dependencies
PR: ports/76872
Submitted by: maintainer
1) added -ip "ignore port" command line option and the same for pm-020.conf
2) if a port fails to build, port manager now instead of failing, adds it to ignore.db then continues
PR: ports/76529
Submitted by: maintainer
1) Made MGmDbArrayFree and MGmDbArray source in /libMG/src/libMG.h
more readable. Fixed a few bad mallocs in MGmDbArrayFree as well.
2) As per suggestions from John E Hein <jhein@timing.com> in
libPMGR/src/libPMGR.h removed hard coded paths and changed so
ports infrastructure files are accessed by using
$PORTSDIR & $PKG_DBDIR environment vars and fixed hard coded
path references in /portmanager/portmanager.c
3) In pmupgrade.c added an abort if error during clean in an attempt
to halt looping when a port conflict occurs, this is untested so
temporary for now.
4) libPMGR/src/PMGRrMakeDescribe.c fixed hard coded paths
5) pmupgrade.c fixed hard coded paths
6) pmupgrade.c Major cleanup of code, seperated upgrade routine into
upgrade, checkForOldDepencies, and PMGRrDoUpgrade routines.
7) fixed portmanager seg faulting when it self updates. (thanks to
"Noah" <admin2@enabled.com> for reporting
8) Added "HINTS AND TIPS" section to portmanager(1) man page.
PR: ports/75596
Submitted by: maintainer
package and lets you decide (in a beauty dialog interface) for each one
if you want to keep it or delete it.
This is similar to pkg_cutleaves, but it does not require any dependencies
and is done in visual, dialog(1) interface instead of command line.
PR: ports/75517
Submitted by: Timothy Redaelli <drizzt@gufi.org>
- Change maintainer email
- Removed all html docs and replaced with man page
- Add WITH_DEBUG knob
- Cleaned up a few compile warnings
PR: ports/75413
Submitted by: Michael C. Shultz (maintainer)
portsdb(1):
- Add a -F flag, which calls make fetchindex internally to
update the ports index file.
portupgrade(1):
- Make -fP properly reinstall an installed package using a binary
package.
- Do not use a binary package when custom MAKE_ARGS or -m option is
defined, unless -PP is specified.
portsclean(1):
- -L: Make sure that a library does not shadow itself.
misc.:
- RcNG'ify the startup script.
- Do not use Object#class which is obsolete in Ruby >=1.8.
1. The working directory and ports directory can be specified
in the portsnap.conf configuration file.
2. Parts of the ports tree (down to the level of individual ports)
can be extracted (e.g., "portsnap extract security/freebsd-update").
This also adds an empty file, $PREFIX/portsnap/.package.this.directory,
which (as the name suggests) ensures that $PREFIX/portsnap/ will be
generated when a package is installed using pkg_add.
PR: 74835
Submitted by: maintainer
* Ignore comments when checking for absolute pathname usage
* Check DISTNAME for uses of PORTREVISION and PORTEPOCH [1]
* Make sure to quote meta-characters in PORTNAME before using it as a
regular expression [1]
* Fix bug when matching OPTIONS in ".if defined(WITH_OPT1) &&
defined(WITHOUT_OPT2)" [2]
* Update OMF checks to account for the new INSTALLS_OMF macro
* Update GConf checks to account for the new GCONF_SCHEMAS macro
PR: 73536 [2]
Submitted by: mat [1]
leeym [2]
