| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
fixes PR#23355 finally.
|
|
Fixes build on NetBSD-1.6.2 and NetBSD-1.6ZE. Existing patches have
either been incorporated or are unnecessary.
|
|
Changes since 5.6.4.4.4:
5.7:
- Solaris USB workaround;
- 15 min timeout for FLUSH CACHE in DVD-RW DAO;
- revalidate recording speed;
- load media upon start-up (Linux used to auto-close tray upon open,
but not the others, which is why this functionality is added so
late);
5.8:
- elder Ricoh firmwares seem to report events differently, which
triggered growisofs and dvd+rw-format to end-less loop at startup
[event handling was introduced in 5.6 for debugging purposes];
- int ioctl_fd is transformed to void *ioctl_handle to facilitate
port to FreeBSD;
- FreeBSD support contributed by Matthew Dillon;
- volume descriptors from second session were discarded in
Restricted Overwrite since 5.6;
5.9:
- some [SONY] firmwares make it impossible to tell apart minimally
and fully blanked media, so we need a way to engage DAO manually
[in DVD-RW]... let's treat multiple -dvd-compat options as "cry"
for DAO;
- refuse to finalize even DVD-R media with -M flag (advise to fill
it up with -M /dev/cdrom=/dev/zero too), apparently DVD-units
[or is it just SONY?] also "misplace" legacy lead-out in the same
manner as DVD+units;
- oops! DAO hung at >4MB buffer because of sign overflow;
- couple of human-readable error messages in poor_mans_pwrite64;
- work around Plextor firmware deficiency which [also] manifests as
end-less loop upon startup;
5.10:
- increase timeout for OPC, NEC multi-format derivatives might
require more time to fulfill the OPC procedure;
- extended syntax for -use-the-force-luke option, it's now possible
to engage DVD-R[W] dummy mode by -use-the-force-luke=[tty,]dummy
for example, where "tty" substitutes for the original non-extended
option meaning, see the source for more easter eggs;
- FreeBSD: compile-time option to pass -M /dev/fd/0 to mkisofs to
make life easier for those who mount devfs, but not fdescfs;
- eliminate potential race conditions;
- avoid end-less loop if no media was in upon tray load;
- interpret value of MKISOFS environment variable as absolute path
to mkisofs binary;
- to facilitate for GUI front-ends return different exit codes, most
notably exit value of 128|errno denotes a fatal error upon program
startup [messages worth popping up in a separate modal dialog
perhaps?], errno - fatal error during recording and 1 - warnings
at exit;
- to facilitate for GUI front-ends auto-format blank DVD+RW media;
- Linux: fix for failure to copy volume descriptors when DVD-RW
Restricted Overwrite procedure is applied to patched kernel;
- FreeBSD: growisofs didn't close tray upon startup nor did the rest
of the tools work with open tray;
- bark at -o option and terminate execution, the "problem" was that
users seem to misspell -overburn once in a while, in which case it
was passed down to mkisofs and an iso-image was dumped to current
working directory instead of media;
- generalize -M /dev/cdrom=file.iso option, but if file.iso is not
/dev/zero, insist on sane -C argument to be passed prior -M and
double-verify the track starting address;
5.11:
- authorship statement in -version output;
- make speed_factor floating point and print "Current Write Speed"
factor for informational purposes;
- Pioneer DVR-x06 exhibited degraded performance when recording DVD+;
- Pioneer DVR-x06 failed to complete DVD+ recording gracefully;
- alter set-root-uid behaviour under Linux from "PAM-junky" to more
conservative one;
5.12:
- single Pioneer DVR-x06 user reported that very small fraction of
his recordings get terminted with "LONG WRITE IN PROGRESS," even
though growisofs explicitly reserves for this condition... It
turned out that at those rare occasions unit reported a lot of free
buffer space, which growisofs treated as error condition. It's not
clear if it's firmware deficiency, but growisofs reserves even for
this apparently rare condition now.
- [major] issue with MODE SENSE/SELECT on SANYO derivatives, such as
Optorite, is addressed;
- Linux can't open(2) a socket by /dev/fd/N, replace it with dup(2);
- more relaxed command line option parsing and simultaneously a
zealous check to make sure that no mkisofs options are passed
along with -[ZM] /dev/cdrom=image;
- report I/O error if input stream was less than 64K;
- -M /dev/cdrom=/dev/zero didn't relocate the lead-out in DVD-RW
Restricted Overwrite;
5.13:
- workarounds for Panasonic/MATSUSHITA DVD-RAM LF-D310;
- Solaris: media load upon start-up;
|
|
|
|
platforms) don't have it.
|
|
|
|
machines so we can't use that to detect ac-power. for now, revert to what
we were doing before.
XXX perhaps acpiacad should be used if it exists, and if it doesn't fall
XXX back to using "acpibat.*discharging". then people with broken acpiacad
XXX can simply not include that driver in their kernel, but worry about
XXX that again after pkgsrc freeze.
also, check __NetBSD_Version__ to see when various envsys(4) API changes
occured, and avoid using new methods on old systems. this should fix PR#23355.
|
|
1.)
An integer overflow in ls in the fileutils or coreutils packages may allow
local users to cause a denial of service or execute arbitrary code via a
large -w value, which could be remotely exploited via applications that use
ls, such as wu-ftpd.
2.)
ls in the fileutils or coreutils packages allows local users to consume a
large amount of memory via a large -w value, which can be remotely exploited
via applications that use ls, such as wu-ftpd.
See
http://cve.mitre.org/cgi-bin/cvename.cgi?name=CAN-2003-0853
http://cve.mitre.org/cgi-bin/cvename.cgi?name=CAN-2003-0854
and the original report
http://www.guninski.com/binls.html
for details.
Patches taken from Red Hat's Security Advisory RHSA-2003:309-01.
reported by reed@
bump PKGREVISION
|
|
|
|
|
|
Redo Grant's changes (USE_PERL5=run, s/graph/image).
No need for USE_PKGINSTALL.
Align variables values.
|
|
* Recognize "charge" in addition to "energy".
* Test ENVSYS_FCURVALID.
|
|
that we're not on AC power via the acpiacad's 'disconnected' entry, not
one of the battery's "discharging" entries, as the latter may be true
even while on AC power.
|
|
Bump PKGREVISION.
|
|
1) remove superfluous long in stage2/fsys_reiserfs.c (forward port from
current version in grub CVS)
2) provide an explicit memcpy stub so the memcpy inserted by gcc3 as
consequence of a pass-by-value can be linked against (patch from
Joachim Kainz in grub bug report 3343)
|
|
committed.
|
|
No, really.
|
|
|
|
* Clean up temp file
* Remove runtime check for graphviz package - that's really handled
via dependencies
* Note requirement for perl
Addresses PR 23207 by Robert Elz / munnari.OZ.AU BANG kre /
|
|
|
|
|
|
move DEPENDS to appease pkglint.
|
|
|
|
|
|
Read dmesg(8) output (e.g. from /var/run/dmesg.boot), and output the
device tree as GIF image to stdout.
Usage: cat /var/run/dmesg.boot | /usr/pkg/bin/dmesg2gif | xv -
|
|
doesn't work, try to use ACPI to get the battery charge value, and AC
charging switch.
Known to work on my Sony VAIO (PCG-GRX315MP) with ACPI, and with thanks to:
Lubomir Sedlacik for testing on an IBM T22 with ACPI
Manuel Bouyer for testing on an IBM Thinkpad 600e with APM, and
Brian A. Seklecki for testing on a Sony PCG-R505
|
|
|
|
XXX: both program and man page are missing a license.
|
|
|
|
Remove (currently) unnecessary dependency on Text::Substitute, and
catch some more rdist output variations.
|
|
Correctly handle various rdist messages to fix post-rdist summary.
|
|
0.11.35
* Brought the copyright in About into 2003. :)
* Copying now also copies owner and group information, except for
symbolic links. If you really need it for links, let me know.
Also note that copying owner info requires root access, and that
group copying is very limited for non-root users. See chgrp(2).
* Hopefully made Copy signal failure when target file doesn't have
full size, which in turn should help Move behave better.
* Added an option to always save configuration on exit, which does
away with the dialog. Kind of silly, but easy to implement.
* Changed logic in internal dir-copy code, so it doesn't fool itself
into thinking it fails quite as easily. Note: when copying a dir,
gentoo will first *delete* the destination, to ensure that the
copy is indeed 1:1 and not a merge. BEWARE!
* The internal routine that rescans a pane no longer activates it,
which means less flickering focus indication when the other pane
rescans due to FAM detecting a modification in its contents.
* FAM rescans are now rate-limited, and will not occur more often
than once every 300 milliseconds. This limit is not settable at
runtime, you need to edit source (gfam.c:96) to change it.
* Added call to gtk_set_locale() during startup, suggested by
tarot.
* Silently assume compiler is GCC, and disregard CFLAGS set by
autoconf. In return, --enable-debug now disables optimizing.
|
|
Refix build on -current now that procfs and kernfs have churned again.
|
|
|
|
|
|
Indicate which hosts have failed rdist stage
|
|
* TAB completion against history. (Contributed by Eduard Roccatello)
|
|
* Fixed bug #972: The total value in the tooltip is now displayed with
fractional portion
* Fixed bug #962: Any errors about wrong configuration are now only
displayed in a dialog after changing configuration and not after
startup. This should help users whose network device is created after
dialing-in.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
New in v0.12.5 (2003/09/27)
---------------------------
Fixed bug in --test-server when using --restrict security options
(test would fail improperly). Thanks to Maik Schreiber for report.
--list-changed-since and --list-at-time now work remotely.
Thanks to Morten Werner Olsen for bug report.
Fixed logic bug that could make restoring extremely slow and waste
memory. Thanks for Jacques Botha for report.
Fixed bug restoring some directories when mirror_metadata file was
missing (as when made by 0.10.x version).
Regressing and restoring as non-root user now works on directories
that contain unreadable files and directories as long as they are
owned by that user. Bug report by Arkadiusz Miskiewicz. Hopefully
this is the last of the unreadable file bugs...
|
|
curses.buildlink2.mk. This was wrong because we _really_ do want to
express that we want _n_curses when we include the buildlink2.mk file.
We should have a better way to say that the NetBSD curses doesn't
quite work well enough. In fact, it's far better to depend on ncurses
by default, and exceptionally note when it's okay to use NetBSD curses
for specific packages. We will look into this again in the future.
|
|
|
