Copy head (r256279) to stable/10 as part of the 10.0-RELEASE cycle.

Approved by: re (implicit)
Sponsored by: The FreeBSD Foundation

sysconf(3) returns a long, not an int.

sysconf(3) returns a long, not an int.

Fix compilation errors.

Use snprintf instead of strc* functions and add bounds checking when creating
pkgngpath

Submitted by: sbz, gahr

Provide an option to users to shoot themselves in the foot.

Reviewed by: bapt
Approved by: cperciva
MFC after: 1 month

Use modern license instead of being consistent with the other pkg_ tools
Use a more informative message
Fix some style(9) nits.
Bump version number
In pkg_add only warn users after the chroot is performed.

Submitted by: gavin, jilles
Reviewed by: beat, bapt
Approved by: cperciva
MFC after: 1 month

Warn users when using pkg tools if it looks like they
be be pkgng users.

Reviewed by: bapt (earlier version)
Reviewed by: kwm
Approved by: cperciva
MFC after: 3 days

Make sure that each va_start has one and only one matching va_end,
especially in error cases.

if a file in plist starts with / then do not prefix it with "prefix" [1]
pkg info -g returns 1 if a file mismatch [2]
flush stdout in pkg info -g [3]
clean up quiet mode (-q | --quiet) output of pkg_version(1) [4]
fix missing error call in uname check added to pkg_version(1) [5]
fix pkg_add(1) fails to install with -C from bad path [6]
only resolve path from pkg_add(1) -p if the given prefix do not start with a '/' [7]

PR: bin/13128 [1]
bin/139015 [2]
bin/113702 [3]
bin/142570 [4]
bin/146857 [5]
bin/157543 [6]
Submitted by: cy [1]
Anton Yuzhaninov <citrin@citrin.ru> [2]
Ighighi <ighighi@gmail.com> [3]
"N.J. Mann" <njm@njm.me.uk> [4]
gcooper [5]
David Naylor <naylor.b.david@gmail.com> [6]
netchild [7]
MFC after: 2 weeks

Do not change owner, group, or mode when package database directory and
its contents are created with pkg_add(1). It may happen when the packing
list contains @owner, @group, or @mode.

Reported by: Donald Bostrom (dbostrom at niksun dot com)
Reviewed by: portmgr (bapt)

Bump PKG_INSTALL_VERSION which should've been done in r236233

MFC after: 3 days
Pointed out by: bapt

Let pkg_add use the ENV variable PACKAGESUFFIX. This
can be used to override the default .tbz package extension to
.txz .tgz or .tar

Submitted by: kmoore
Obtained from: PC-BSD
Sponsored by: iXsystems
MFC after: 3 days

Make sure that each va_start has one and only one matching va_end,
especially in error cases.

- Print package name in case an empty pkgdep line is found.

PR: bin/164378
Submitted by: Yuri <yuri AT tsoft.com> and many others
Approved by: flz
MFC after: 2 weeks

Touch up some more small typos missed in the previous round.

Reported by: Ben Kaduk <minimarmot@gmail.com> et al.

Spelling fixes for usr.sbin/

- add a missing "be" and "in"
- fix other errors introduced when committing r226436
- add 'function' to a sentence where it makes sense

Submitted by: delphij
Submitted by: dougb
Submitted by: jhb
Approved by: dougb
Approved by: jhb

- change "is is" to "is" or "it is"
- change "the the" to "the"

Approved by: lstewart
Approved by: sahil (mentor)
MFC after: 3 days

Backout libinstall.a -> libpkg commit.

Discussed with: erwin, brooks, bapt

Bump PKG_INSTALL_VERSION to 20100403.

Fix pkg_delete, check if the file we're trying to delete is a
symlink before complaining that it doesn't exist. Typical case
would be a leftover library symlink that's left over after the
actual library has been removed.

Reported by: tabthorpe

Various fixes.

- Replace hardcoded INDEX version. [1]
- Fix a buffer overlap. [2]
- Remove empty package when fetching fails and -K is used. [3]
- Remove useless chmod2() after mkdtemp(3). [4]
- Replace mkdir(1) call with mkdir(2). [5]
- Get rid of some vsystem() calls.
- Switch from lstat(2) to open(2) in fexists().
- Try rename(2) in move_file() first.
- Bump PKG_INSTALL_VERSION to 20100401.

PR: bin/145101 [1], bin/139492 [2], bin/144919 [3]
bin/144920 [4], bin/144921 [5]
Submitted by: gcooper [1,2,3,4,5]

- Add support for UPDATING remote fetching.
- Reorganize EXAMPLES section in pkg_updating(1).
- Style fixes.
- Bump PKG_INSTALL_VERSION to 20100122.

Submitted by: beat
MFC after: 1 week

- Bump PKG_INSTALL_VERSION to 20090902 after dougb's changes.
- Change the comment to say that version must be changed when a
non-cosmetic change is made.

Add support for INDEX-9 [1]

While I'm here, strip off support for FreeBSD 5.x.

Submitted by: Alexey Shuvaev <shuvaev@physik.uni-wuerzburg.de> [1]

Don't imply that only FTP urls are supported when we can't fetch
a package.

Approved by: re (kib)
MFC after: 3 weeks

When running pkg_add -r, check & install our dependencies for each
package rather than expecting our top level package to get all of
the dependencies correct.

Previously, the code depended on the top level package having all
of the pkgdep lines in +CONTENTS correct and in the right order,
but that doesn't always happen due to code such as this (in
security/gnutls/Makefile):

.if (defined(WITH_LZO) || exists(${LOCALBASE}/lib/liblzo2.so)) && !defined(WITHOUT_LZO)
LIB_DEPENDS+= lzo2:${PORTSDIR}/archivers/lzo2
....

With such conditional dependencies, my 'sophox-packages' package won't
install. The dependency tree looks like this:

sophox-packages
...
x11/gnome2
x11/gnome-applets
net/libgweather
devel/libsoup
security/gnutls
security/libgcrypt
security/libgpg-error
...
x11/gnome2
archivers/file-roller
archivers/gtar
archivers/lzop
archivers/lzo2
...

gnutls doesn't depend on lzo2 initially, but lzo2 is dragged into the
mix via other dependencies and is built by the initial 'make'. The
subsequent package generation for gnutls adds a pkgdep line for lzo2
to gnutls' +CONTENTS but the pkgdeps in sophox-packages' +CONTENTS
has gnutls *before* lzo2.

As a result, sophox-packages cannot install; gnutls fails because lzo2
is missing, 82 more packages fail because gnutls is missing and the
whole thing spirals into a super-confusing mess!

MFC after: 3 weeks

Skip @pkgdep if there's no argument.

Submitted by: pav
MFC after: 1 week

- Backout latest changes (follow symlinks: r186496, r186518).
- Bump PKG_INSTALL_VER to 20090106.

Fix memory leaks introduced in last commit.
Bump version to 20081227.

Reported by: gcooper
Submitted by: Andrea Barberio <insomniac@slackware.it>
MFC after: 1 month

Follow symlinks when deleting directories.
Bump PKG_INSTALL_VER to 20081225 (Merry Christmas \o/).

PR: bin/54446
Submitted by: Andrea Barberio <insomniac@slackware.it>
MFC after: 1 month

Use humanize_number to report pen-sizes so people don't have to count
the digits when trying to install openoffice.

Enable the -n option of pkg_create(1).

The changes to make the option work are already in place, but I missed
the patch hunk that adds it to the getopt() option-handling loop.

Pointy hat: keramida
Approved by: flz
MFC after: 1 week

Remove support for RELENG_4 (__FreeBSD_version < 500039).

MFC after: 1 day

- add: Keep dependent packages too if -K is specified.
- updating: terminating '\n' is not part of the package origin.
- bump PKG_INSTALL_VERSION to 20080612.

PR: bin/119368 [1], bin/124459 [2]
Submitted by: gcooper [1], Beat Gatzi <beat@chruetertee.ch> [2]
MFC after: 3 days

- Add long options to pkg_install.
- Remove check for '-?' as it's not listed in authorized options.
- Bump PKG_INSTALL_VERSION to 20080530.

Bump PKG_INSTALL_VERSION to 20080528.

X-MFC after: latest HEAD changes have been MFC'ed to RELENG_[67]

- Restore functionality broken in previous commit; we need to be able to report
multiple installed packages with the same PKGORIGIN.

Reported by: marcus
MFC after: 1 month

Optimize package registration/deregistration. Previously, when looking up the
package name for the origin of a dependency, all entries in /var/db/pkg were
traversed for each dependency of added/removed package. Now, gather all the
origins first, then do the lookup in a single pass over /var/db/pkg.

This should provide a major speedup for packages with hundreds of dependencies.

Submitted by: rdivacky (earlier version)
MFC after: 1 month

Some code cleanups, including removing of stale getopt argument and
hardcoded +CONTENTS defines.

PR: bin/117855
Submitted by: Beat Gaetzi <beat@chruetertee.ch>
MFC after: 14 days

Momoize the results of isinstalledpkg()

PR: bin/116452
Submitted by: wollmann
MFC after: 7 days

Add whitespace.

Update for INDEX-8 support.

- Replace rather inefficient bubble sort with a recursive depth-first search.
This speeds up registration of packages considerably.
- style(9) police welcome!

PR: bin/112630
Submitted by: Stephen Montgomery-Smith <stephen@cauchy.math.missouri.edu>
Tested by: bento i386 experimental run
MFC after: 14 days

Make consistent with style(msmith). 4 spaces indent, tab for each 2 indents.

Invoke tar(1) with the -p option when installing a package
from an URL (i.e., do it the same way as when installing
from a file). This fixes the lossage of the setuid bits.
It wasn't a problem before because GNU tar(1) implied the
-p option for root, but BSD tar(1) doesn't do that.

Discussed with: tobez and some advanced users :)

Cosmetic: fix indentation.

Turn on libfetch debugging if -v was specified on the command line.

MFC after: 2 weeks

Allow variable amounts of verbosity.

Add INDEX-7 and remove trailing spaces.

MFC after: 5 days

Add -K option to pkg_add and pkg_info.
This option saves packages to PKGDIR (if defined or current directory by default)
as they are downloaded.
Silent a warning when -n is used and package has a +DISPLAY file.

Approved by: krion
MFC after: 1 week

When using @cwd %%FOO%%, we must ensure to return in the original
prefix later, but doing so with @cwd %%OLDPREFIX%% (having
PLIST_SUB+="OLDPREFIX=${PREFIX}") hardcodes the value in the packing
list. That's not really a problem when dealing with ports but that's
a problem with packages since pkg_add -p option only overrides the
first @cwd occurrence.

This patch allow us to use @cwd without any argument. If no
directory argument is given, it will set current working directory
to the first prefix given by the @cwd command.

PR: bin/77212
Submitted by: flz

Do not bitch about bsdpan modules not having origin.

Based on PR: bin/82269
Submitted by: Steven Hartland <killing at multiplay dot co dot uk>
MFC after: 3 days

remove a few uses of the word `really'

Reviewed by: hrs
Approved by: re (scottl)
MFC after: 1 week

Implement @noinst field which has at the moment the same meaning and
function as @comment has. But will be valid only for files and not
for md5 sums, rcsid's and comments in the future.

Submitted by: flz
Approved by: re@ (scottl)

Use INDEX-6 for versions of FreeBSD > 600000

For variables that are only checked with defined(), don't provide
any fake value.

Add a -q flag to pkg_version. It suppresses the output of <,=,>.
This makes it possible to "portupgrade `pkg_version -q -l '<'`".

Don't put DEBUG to CFLAGS, we have the standard DEBUG_FLAGS for this.

Use the length modifier 'll' instead of 'q' to print long longs.

Downgrade WARNS level to more tolerable value. Attempt to fix
casts as lvalue usage whenever possible.

- pkg_add spawns itself as argv[0] when installing dependent packages, to
enable the use as a port on older systems

- use absolute paths in all calls to external programs, to account for strange
PATH settings

- use INDEX or INDEX-5 depending on FreeBSD version, to enable the use on
FreeBSD 4.x as a port.

- conditionalize all 4.x/5.x differences on __FreeBSD_version, so that the
pkg_install tools can be kept in sync on 4.x and 5.x

- Bump PKG_INSTALL_VERSION

Reviewed by: portmgr (bento run)
MFC after: 4 weeks

- pkg_info: flag -r: (show packages this packages depends on (documentation change))

- pkg_info: new flag -j (show the requirements script)

- pkg_info: fix verbose output when used on packages

- better handling of corrupt entries in /var/db/pkg

- differ between corrupt entires and packages not installed

- various small fixes

PR: 56989, 57016, 57029, 26468

- match package version numbers with relational operators

- use glob patterns when matching packages by origin

- csh-style {...} choices in glob matching

- pkg_info: new flag -E (list matching package names only)

- pkg_version: new flag -T (test if a given name matches a pattern)

- new flag -X (interpret pattern as an extended regular expression)

PR: 56961

- complete rewrite of the version number parsing code, restoring compatibiliy of 5.x with 4.x and portupgrade

- parse version numbers of ports containing an underscore followed by a number correctly

- handle special strings pl, alpha, beta, pre and rc

PR: 56961

When invoking tar, make sure the mode option (-x) is first.

Fix a case of undefined behavior due to overlapping buf objects in
snprintf (buf, size, fmt, buf, etc). This only works by chance with our
libc, but fails (with a truncated string) on e.g. glibc.

Okayed by: sobomax
MFC after: 1 week

Backout 1.14, it seems to be causing problems with libxml build
and I don't have time to investigate this code in much detail right
now.

Use strncmp() and not strcmp() here to properly match
installed ports.

Submitted by: Harold Gutch <logix@foobar.franken.de>

pkg_create incorrectly does not add trailing '\n' when it receives
either COMMENT or DESCR from the command line. When a port is
installed, one gets both +COMMENT and +DESCR files with a trailing
'\n' character. However, +COMMENT does not contain a trailing '\n'
when it is installed from a package due to this behavior of pkg_create.

Therefore, make sure it behaves exactly the same regardless of
where got its information; either command line or files. The modified
functions are used by pkg_create.

PR: 52097
Reviewed by: bento, kris,
portmgr, re,
Michael Nottebrock <michaelnottebrock@gmx.net>,
Martin Horcicka <horcicka@FreeBSD.cz>
Approved by: re (scottl)
MFC after: 1 week

* Add explicit conflict-checking to the package tools. Packages can
register a list of other packages with which they conflict (via the
-C option to pkg_create), and they will refuse to install (unless -f is
specified) if one of the listed packages is already present.

* Update documentation for the new feature as well as fleshing out some
existing documentation.

* Bump PKG_INSTALL_VERSION so this feature can be tested for.

Submitted by: Sergey Matveychuk <sem@ciam.ru>
PR: bin/47145
MFC after: 2 weeks

Revert url.c:1.2: there was a race condition between when I produced
the fix and when I committed it -- in between the two, DES committed
a different fix, rendering this fix unneeded.

Requested by: des

Don't pass NULL in as a flags argument to libfetch's fetchGetURL()
as this can result in a NULL pointer deference when parsing the
flags later. This change fixes "pkg_add -r" on 5.0-CURRENT for
me; not quite clear how the problem was introduced.

Prevent buffer overflow in format_cmd() by properly tracking maximum
buffer size.
Reported by: Lionnel CHAPTAL <Lionnel.Chaptal@IPricot.com>
MFC after: 1 week

Assume that packages passed on stdin are in bzip2 format, not gzip.
(sysinstall depends on this feature for package addition.)

Comment on hard-coded bzip2 usage in the spirit of rev. 1.58.

Clean up a little. Don't need to include fetch.h any more, and I can't
find any consumers of HOSTNAME_MAX.

Break fileGetURL() out into its own file so that pkg_install/lib consumers
pkg_{create,delete} don't need to needlessly link with libfetch.

Whitespace cleanup

Make the threatened fts(3) ABI fix. FTSENT now avoids the use of the struct
hack, thereby allowing future extensions to the structure (e.g., for extended
attributes) without rebreaking the ABI. FTSENT now contains a pointer to the
parent stream, which fts_compar() can then take advantage of, avoiding the
undefined behavior previously warned about. As a consequence of this change,
the prototype of the comparison function passed to fts_open() has changed
to reflect the required amount of constness for its use. All callers in the
tree are updated to use the correct prototype.

Comparison functions can now make use of the new parent pointer to access
the new stream-specific private data pointer, which is intended to assist
creation of reentrant library routines which use fts(3) internally.

Not objected to in spirit by: -arch

Sigh, finally activate pkgwrap.c.

Big pointy hat to: sobomax (who apparently needs some sleep)
Submitted by: walt <wa1ter@hotmail.com>

Introduce notion of the package tools revision and allow to wrap all tools
included into pkg_install according to the content of /var/db/pkg_install.conf
file, which specifies version and alternative location of the tools. Format
of the said file is very simple: one line which specifies revision of the
alternative version of the tools and their location separated by space,
i.e.:

20030102 /usr/local/sbin

This would allow bsd.port.mk to install and use up to date version of tools
on older system from ports.

Also add new `-P' flag to pkg_info, which causes it to report currently
installed version of package tools.

Discussed with: will

Don't treat unknown commands as fatal errors: they are probably just
the signature of out-of-date pkg_tools and newer packages.

Reviewed by: jkh
MFC after: 3 days

Grrr.... mumble.... a linger gzip'ism still existed.

Submitted by: Craig Rodrigues <rodrigc@attbi.com>

Find remaining hardcoded gzip bits and change to bzip2 bits.
Mark with XXX so someone that cares about being able to handle either
bziped or gziped packages knows where the remaining nits are.

Preserve suid/gid permissions when extracting packages remotelly
fetched with -r option

Reviewed by: will
Approved by: will

Ensure that origin field in the packing list structure is initialised. This
fixes problem when there are origin-less packages installed on a system

Reported by: Mark Knight <markk@knigma.org>
MFC after: 3 days

Fix typos; each file has at least one s/seperat/separat/
(I skipped those in contrib/, gnu/ and crypto/)
While I was at it, fixed a lot more found by ispell that I
could identify with certainty to be errors. All of these
were in comments or text, not in actual code.

Suggested by: bde
MFC after: 3 days

Another .tbz2 -> .tbz leftover. Properly look for pkgname.tbz.

MFC after: 1 day

C replacement for the Perl5 code.

Submitted by: reg

Eliminate free'ing already freed chunks.

MFC after: 1 week

- Make use of DEPOROGINs (if there are any) when installing package;
- fix few bogosities here and there;
- move some common routines into the library.

MFC after: 2 weeks

Mark all internal libraries with INTERNALLIB.

replace __FUNCTION__ with standardized __func__.

Requested by: jhb

fix build:
you may not use string concatination with __FUNCTION__, replace all occurances
of:
__FUNCTION__ ": error string"
with:
"%s: error string"

New feature: allow origins of all dependencies be recorded into package list
using new `@comment DEPORIGIN:...' directive. This would allow us to make
many neat things including:

- easier binary upgrades;
- source upgrades without using external tools by simply extending
bsd.port.mk and pkg_install tools;
- mixed-mode upgrades (source + binary);
- depreciate and deorbit silly +REQUIRED_BY files in the near future.

This feature is no-op until appropriate bsd.port.mk patch is committed, and
even when it is already committed packages generated will remain 100%
compatible with old set of pkg_install tools (module all those neat
features, of course).

MFC after: 6 days

Handle PLIST_ORIGIN properly.

Poked by: kris
Pointy hat grabbed by: sobomax
MFC after: 6 days

Add ability to query installed packages based on origin.

Fix SCM ID's.

Make sortdeps() working with dependency lists containing one or no elements.

Submitted by: roam
MFC in: 1 week

Choose more paranoid modes for the temporary directory so the user can't
easily browse its contents.

Noted by: Antoine

Improve recently committed fix for -t bug.

Submitted by: "Andrew L. Neporada" <andr@dgap.mipt.ru>
MFC after: 3 days

Make `-t' flag in pkg_*(1) tools actually working.

PR: 30843
Submitted by: Vladimir B. Grebenschikov <vova@express.ru>
MFC after: 2 weeks

- Introduce a notion of `packing list format version'. This allows making
non-backward compatible changes in the format of packing list and handle
them gracefully;
- fix a longstanding issue with symlinks handling. Instead of recording
checksum for the file symlink points to, record checksum for the value
returned by readlink(2). For backward compatibility increase packing list
format minor version number and provide a fallback to a previous behaviour,
if package in question was created with older version of pkg_* tools;

Submitted by: Alec Wolman <wolman@cs.washington.edu>, sobomax

- don't record MD5 checksum for device nodes, fifo's and other non-regular
files.

Submitted by: nbm
MFC in: 2 weeks

WARNS=2 cleanup.

Tested on: i386, alpha
MFC after: 2 weeks

Various fixes and improvements:
- fix harmless compiler's warnings (unused variables and missed prototype);
- before refusing to delete package because "there are packages installed
that require this package" check that packages in question is actually
installed;
- add new `-r' option to pkg_delete(8), which instructs it to delete not only
packages specified at command line, but all packages that depend on
specified packages as well.

MFC after: 2 weeks

Reduce private "lesser known" function redirection to improve clarity.

Approved by: jkh

fix for pkg_add to symlinked prefix directories that are more than 1 link deep.

PR: bin/28274
Submitted by: John Hein <jhein@timing.com>
MFC after: 1 month

Cosmetics: replace dozen instances of "(tmp = getenv(PKG_DBDIR) ? tmp : DEF_LOG_DIR)"
with macro.

MFC after: 1 month

Cosmetics: kill blank lines at the end of file.

MFC after: 1 month

Usability tweak:

Use '' quotes instead of `' to delimit names of files and packages in
warning and error messages, because it is easier to cut-n-paste name in
question that way (single click) without confusing the shell. And yes,
I know that it is less eye-candy...

MFC after: 1 month

When there is a file that can't be deleted due to checksum mismatch print name
of that file to stdout to simplify debugging. IMO it was a mistake to print
this warning only when `verbose' mode is on.

MFC after: 1 month

The security officer requested this be backed out for discussion.

Remove s_strl*(). I am not sure what was thought they accomplished.

When reading the code I had to stop, say "ok, what does *these*
modifications of strl*() do? Pull out grep. Oh, not in add/, maybe above
in ../lib/? Yep. So what do they do? Comments above them are misleading,
guess I'll have to read the code. Oh, they just test strl* against the
size and return the result of the test. Now I can continue to read the
code I was.

The uses of s_strl*() then test that result and errx()'s.
Lets think about the "optimized" code I am removing:

In general the compiler pushes the three args to strl* onto the stack and calls
s_strl*. s_strl* has to indirectly access 3 args from the stack. Then push
them on the stack a 2nd time for the real strl* call. s_strl* then pops the
return from strl* off the stack; or moves it from the register it was returned
in, to the register where tests can happen. s_strl* then pops the three
arguments to strl*. Perform the test, push the result of the test, or move it
from the result register to the return value register. The caller to s_strl*
now has to either pop the return value of s_strl* or move it from the return
value register to the test register. The caller then pops the three args to
s_strl* off the stack (the same args that s_strl* itself had to pop off after
the real call to strl*). The s_strl* caller then performs a simular test to
what has already been done, and conditionally jumps. By doing things this way, we've given the compiler optimizer less to work with.

Also, please don't forget the that call to s_strl* has possibly jumped to code
not in the cache due to being far away from the calling code, thus causing a
pipeline stall.

So where is the "optimization" from s_strl*?
It isn't code clarity.
It isn't code execution speed. It isn't code size either.

Use reallocf(3) to increase size of malloc'ed buffer instead of
malloc()+memcpy()+free().

Perform a major cleanup of the usr.sbin Makefiles.
These are not perfectly in agreement with each other style-wise, but they
are orders of orders of magnitude more consistent style-wise than before.

Fix the type of the NULL arg to execl()

Idea from: Theo de Raadt <deraadt@openbsd.org>

Use getdtablesize() instead of OPEN_MAX.

Submitted by: bde

Pass on the verbose option to fetchGetURL().
In the child that's set up to run tar(1), close all file descriptors except
stdin, stdout and stderr.

PR: bin/27760

Style policy: reformat multiline comments to conform to style(9).

Clean up some of the evil string handling in this code, replace
warnx()+exit() with errx() and replace a big if..then..else construct
to determine the package download directory with a lookup table.

Reviewed by: jkh
MFC after: 2 weeks

- Add which(1)-like functionality into pkg_info;
- fix a harmless bug in match_installed() function introduced in my last
commit;
- uniformely reorder includes across files.

Submitted by: Garrett Rooney <rooneg@electricjellyfish.net>
Not objected by: jkh, -ports

When creating a package sort dependencies in such a way that if dependency
A depends on dependency B then dependency A will be in all cases listed
before B, so ``pkg_add -r'' will fetch/install packages in the correct order.

Previously dependencies were sorted just by its names, which is why
``pkg_add -r'' never actually worked properly.

To be usefull, hovewer, this fix requires that all packages have been
rebuilt, so it will take some time until users would be able to feel
posititive improvements. For the same reasons it is desirable to propagate
these changes to the 4-stable package building cluster *before* 4.3 ports
freeze, so packages for 4.3-RELEASE would be properly prepared.

Prompted by: kris
Insanely appreciated by: obrien
Silently approved by: jkh, -ports

When matching installed packages against glob keep track of which patterns
actually triggered a match and which did not, and add patterns that didn't
into resulting list, so caller will have a chance to notify user that package
isn't installed. This should fix current, POLA-breaking behaviour when user
doesn't receive a notification if he specifies several packages, some of which
aren't installed.

Properly detect and report malloc(3) failures.

- Merge recently added into pkg_info(1) regex/glob matching functionality into
pkg_delete(1) as well;
- add a new `-a' option for pkg_delete(1) to delete all installed packages;
- add a new `-i' option for pkg_delete(1) to request simple rm(1)-like
interactive confirmation before attempting to delete each package.

Silently approved by: jkh, -ports

- Add ability to handle bzip2-compressed packages;
- fix cosmetics to shut-up compiler in -pedantic mode (axe several unused vars
and provide default clause in several switch() statements).

No response from: -ports

Add `_PATH_DEVZERO'.
Use _PATH_* where where possible.

Insetad of using fileno(ftp) as stdin for tar, use a pipe and have the
parent pass the data received from the server to the child.

Cause fatal error messages to be a little more helpful to the programmer
concerning where they're taking place.

Switch from [r]index() to str[r]chr() functions, which are more ISO
compliant.

Prompted by: Edward Welbourne <eddy@vortigen.demon.co.uk>

Fix symlink-to-a-dir handling in pkg_delete.

Reviewed by: -ports
Tested by: bento

Use libfetch instead of libftpio. This adds support for http and IPv6.

Do not preserve UIDs, permissions nor flags of package system files in
tar files. This fixes clean-up problems during package creation and
does not affect the actual files to be included in the package.

The fix submitted on the attributed PR was identical to the one
obtained from NetBSD.

PR: 17386
Reported by: Adrian Filipi-Martin <adrian@ubergeeks.com>
Obtained from: NetBSD

1. If checking for directory-ness, check "dir" and "dir/" to catch
the case where we have a symlink pointing at a dir.

2. Restore stomped character before returning in make_hierarchy()

Use lstat(2) rather than stat(2) in isdir(), so that a symlink to a
directory is not considered a directory. I have a feeling all the other
stat(2) calls should instead be lstat(2) calls, but I have not suffiently
determined that the current behavior [especially in isfile()] isn't
depended upon by someone.

Ok'ed by: JKH

Deal with package filenames which contain spaces.

Submitted by: Ming-I Hsieh <mihs@wm28.csie.ncu.edu.tw>
PR: 15667

$Id$ -> $FreeBSD$

-Wall fixes. Ambigious if/elses, uninitialized values, missing headers,
missing prototypes, and empty format strings.

Revive the PKG_ADD_BASE stuff a bit.

Make it possible to have separate install scripts as well as have
an on-delete script.

Submitted by: Rajesh Vaidheeswarran <rv@fore.com>

Fix support for uncompressed (".tar") package types. It's not completely
fixed (chained dependency checking for pkg_add is broken, for one thing)
but at least you can now create one package and use it.

Revert 1.33; lstat() is indeed the wrong thing to do here. I'll
revisit this after release.

Use lstat() rather than stat in determining whether a file exists,
otherwise we'll miss the "symlink exists but points nowhere" case.
Submitted by: asami and/or possibly <ru@ucb.crimea.ua>

correct usage of cleanup() everywhere - this became bogus when the warnx/errx
stuff was added and nobody seems to have noticed. :)
Noticed by: Josh MacDonald <jmacd@paris.CS.Berkeley.EDU>

o Set paths internally for pkg_add for tools we might want to invoke.
o Fix bogus suffix handling.
o Tell user when an FTP url is being xferred rather than being silent.
This sort of violates "the unix way" but it stops people from whacking
^C when they think it's hung, too. Sometimes visual indication of
success is important. Doesn't spit out anything if not on a TTY.

Silence potentially bogus warning.

Paranoia: use mkdtemp instead of mktemp
PR: bin/3212
Reviewed by: jkh@FreeBSD.ORG

Clarify "Cannot delete file as directory" errmessage.

PR: bin/1985

Fix dependent-package loading when pkg_adding with an FTP URL. I broke
this when I changed the playpen code to disallow recursion. I know
this fixes a PR someplace but I can't seem to find it with the GNATs
search tool.

Eliminate the idea of nested "playpens" entirely - it just obfuscated
the code and, in at least one case, made it more dangerous to no gain.

MF22: Additional paranoia in cleanup.

- do string-magic on the right variable when trying
to find dependent packages on the local disk (i.e. package given as
filename, not URL)
- document $PKG_PATH
- use -v on pkg_add for dependent packages

Submitted by: hubertf@NetBSD.ORG
Obtained from: NetBSD

Don't return from cleanup() - if it's removed the playpen then very
bad things will happen if we return to unsuspecting code.
Noted to my great dismay by: Tim Vanderhoek <hoek@hwcn.org>

Use err(3). Cosmetics in usage string and man page.

Eliminate an annoying message.

Change the naming conventions for preserve'd files and also document
the option in pkg_create. Now preserved files start with a . and are
named .<filename>.<pkgname>.backup so that their purpose is more clear.

Note that just using the preserve option without proper pkg_deps
is also foolish since packages being deleted in the incorrect order with
preserve on can generate some odd results.

YAMF22 - deletes and preserve option fixes.

Add support for @option preserve - moves existing files out of the way
before replacing them (using pkgname to make this hopefully unique).
Delete also moves them back, if they exist, resulting in a package
which can be "backed out" with reasonable safety.

Revert $FreeBSD$ to $Id$

Make the long-awaited change from $Id$ to $FreeBSD$

This will make a number of things easier in the future, as well as (finally!)
avoiding the Id-smashing problem which has plagued developers for so long.

Boy, I'm glad we're not using sup anymore. This update would have been
insane otherwise.

Use new ftpGetURL interface, as result, report FTP error
via ftpErrString or net error via hstrerror(h_errno)

Fix PR#1557 - chained package dependencies using URLs didn't work.

Put verbosity in the right place.

Add MD5 checksums to packing list for all files during pkg_create and use
later at pkg_delete time to verify that you're deleting what you added.
This, of course, does NOT cover the case where a file you still need
hasn't changed! That's a tougher problem to solve, and this provides
only the minimal amount of safety belt. MD5 checksums are stored in comment
fields, so packages produced with these tools are backwards compatible with
the older ones.

Make pkg_install understand and use libftpio instead of its own
copies of the ftp support routines. Also some cosmetic and minor
bug fixes I've been meaning to incorporate for awhile.

Change min_free to return free space as an off_t instead of a size_t.

Fix a couple of miscellaneous bugs and make pkg_add also support reading
from stdin.

Initialize the argument list in the unpack() routine to zero.
This prevents a core dump due to the stack being trashed if pkg_add
is asked to process something that really isn't a package file.

Fix to catch a stupid oversight on my part which allowed this thing
to remove the current directory under certain circumstances.
Submitted by: Richard J Kuhns <rjk@watson.grauel.com>

Bring my pkg_install improvements forward from 2.1.

Fix the `pkg_add relative to .' bug. Thank you, Ville!
Submitted by: Ville Eerola <ville@vlsi.fi>

Bring forward my changes from 2.1

Don't be needlessly verbose in the pkg_create.

Make the play-pen searching routine even more agile in trying
to create a /usr/tmp dir if it makes sense to do so.

Use size_t properly in all fs size ops. This kills a bug
encountered by someone who had his /tmp on a partition with 3.7GB of
free space!
Spotted by: Francisco Rosich Viana <rosich@power.ci.uv.es>

Some fixes to make this "TMPDIR agile".
Submitted by: jmacd + some of my own fixes.

Make "out of space in ${PKG_TMPDIR}" handling a bit more robust.
Print the right thing, do the right thing. Back to you, Satoshi!

Round this out. Now seems to transfer packages properly over ftp
as it was supposed to. PKG_PATH more sorely needed than ever now..
Hmmmm.

Argh! I hate my life! I can't seem to do anything right here today.
This should restore functionality to the pkg_install suite.

Sync this up with ftp.c.

Remove dependency on sysinstall.

Add ftp.c to makefile - that might make it work better. Sheesh!

Change the environment variable this looks for from FTP_PASSIVE to
FTP_PASSIVE_MODE. It would be really nice if we could standardise on
this name so that all tools (like ncftp) that offer passive/active
ftp selection would work seamlessly with one user environment variable
setting.

Totally eliminate the dependency on libftp (which will be removed from the
FreeBSD source tree) and switch to the internal ftp routines developed
by Poul-Henning and used in sysinstall.

You can run pkg_info everywhere expect /var/db/pkg directory :-)
Running there you got any kind of strange errors from tar caused
by treating directories as tar files!
Fix it by adding new isfile(name) (check for reg. files) to simple fexists(name) calls.

Fix coredump for -v flag
Submitted by: Thomas Gellekum <thomas@ghpc8.ihf.rwth-aachen.de>
Obtained from:

Use lstat() instead of access() for checking file existence. It works
for symlinks too, and according to Rod, access() is evil anyway.

Reviewed by: jkh

Remove trailing whitespace.

Fix a long-standing bug that broke pkg_info utterly and probably made
pkg_add a little wiggy too.

Add code for dealing with URLs.

Check dependencies BEFORE doing final extraction of package.
Fix a bug where direct-extract mode would have still invoked the
packing list reordering commands.

Export the min_free() function.

1. Add an new @option directive. You'll understand it better in a moment.
2. Fix a long-standing bug in pkg_add where the failure of one package in
a multipackage installation (pkg_add *.tgz) would blow you right out of
the water. Ick.

Some patches to make deletions of non-existent files and directories slightly
less fatal (so that the database entry is subsequently removed anyway).
Submitted by: Edward Wang <edward@edcom.com>

Correct a bogusly formatted printf().

Add support for the ${PKG_DBDIR} environment variable. /var/db/pkg is the
default, and probably not something you'd ever change, but now it's
possible to do so for local/custom installations.

Use a more nicely formatted message when we run out of space.

Rename TMPDIR to PKG_TMPDIR, just to make Jeff happy.. :-)

Make pkg_install use /var/tmp instead of /tmp
Suggested by: hsu

More changes from John T. Kohl. Man pages, mostly.
Submitted by: jtkohl

Many of John T. Kohl's patches from NetBSD. Thanks, John!
Submitted by: jkohl

Add a new basename_of() function.

Print a better verbose message when creating packages.

Fix bug in home directory calculation that I introduced with the free space
checking code. Fix the free space checking code, while I'm at it! ;)
Thanks to Garrett for spotting the home dir problem.

Be optionally verbose about free-space checking. I need this to ascertian
whether or not this always works.

Add the ability to verify that a package won't extract in the space
available. Thanks to Michael Elbel for pushing me in the right direction.

Don't let pkg_delete remove the recorded location until everything has
absolutely, positively been successfully removed.
Reviewed by:
Submitted by:
Obtained from:

Add @srcdir option for customizing creation of packages.
Submitted by: jkh

Argh! What a boneheaded bug! Finally fix the blank-line-in-packing list
problem! I am pond scum!

Close /dev/tty that is opened in y_or_n()...this is called many times
so it will run out of ptys if not properly closed.

The environment variable PKG_PREFIX is set to the first @cwd command
in the packing list, or the argument to -p if it is specified, before
the requirements/installation/deinstallation scripts are called. This
enables the scripts to be written to work on the final installation
destination, even if the user uses -p to override the package's default.

Add new find_plist() function.

Make y_or_n() pay attention to AutoAnswer global.

Added code for doing % sequence expansion, added new @unexec command.

Fixed an embarassing bug with @owner, @mode & @group

Fixed bug in copying script files from absolute locations

Made staging area configurable rather than assuming /tmp (-t)

Lots of misc tweaks, support for arbitrary separators in pkg_info, more
intelligent name handling in pkg_create. Most of these files are changed
because of rcsid's being different in my cvs tree and freefall's (foo).

Removed annoying -Wall

This commit was generated by cvs2svn to compensate for changes in r327,
which included commits to RCS files with non-trunk default branches.