Re: Typos in plus49.html

Thanks for the big review! Janne do you care about it or do I merge
it?

On Thu, Mar 24, 2011 at 11:39:55AM +1300, [email protected] wrote:
> Hi, guys.
> 
> I'm looking forward to 4.9 release; in the meantime here are a few typos.  A few
> are very trivial, some are changes to make for (in my opinion) easier reading,
> but some definitely need fixing.
> 
> The list goes from top-to-bottom on plus49.html; each paragraph is current text
> followed by what I think it should be.
> 
> Thanks.
> 
> Make makewhatis(8) more silent unless launched with -v option.
> Can't be "more silent" - so either "quieter" or "less noisy" or "less verbose"
> Make makewhatis(8) quieter unless launched with -v option.
> 
> Disabled sbt(4) and ubt(4) on all kernel.
> Disabled sbt(4) and ubt(4) on all kernels.
> 
> Make bgpd(8) print nicely 4-byte AS.
> Make bgpd(8) nicely print 4-byte AS.
> 
> Allow relayd(8) debug logging before daemonization to help identification of
> configuration issues. Fixes PR6314.
> Allow relayd(8) debug logging before daemonization to help identify
> configuration issues. Fixes PR6314.
> 
> Clear bpf(4) filter memory before using it. Prevents from leaking previous
> kernel stack contents through a malicious bpf(4) filter.
> Clear bpf(4) filter memory before using it. Prevents leaking previous kernel
> stack contents through a malicious bpf(4) filter.
> 
> Fixed iked(8) traffic selector configuration that it is always "from $localnet
> to $peernet" and not depending on the initiator/responder mode.
> Fixed iked(8) traffic selector configuration so that it is always "from
> $localnet to $peernet" and not depending on the initiator/responder mode.
> 
> Make mandoc(1) skipping .TH with excessive arguments instead of throwing a fatal
> error.
> Make mandoc(1) skip .TH with excessive arguments instead of throwing a fatal error.
> 
> Make mandoc(1) skip .br .sp .nf .fi .na with arguments, using empty strings like
> groff instead of throwing a fatal errors for
> For what?  Guessing it can just stop there, or say "for the same"?
> Make mandoc(1) skip .br .sp .nf .fi .na with arguments, using empty strings like
> groff instead of throwing a fatal errors.
> 
> make random(4) be more careful with nanotime() calls in early entropy storage,
> since at least sparc may not have the clock mapped.
> Make random(4) be more careful with nanotime() calls in early entropy storage,
> since at least sparc may not have the clock mapped.
> 
> Make i386 and amd64 only use memory mapped extended PCIe config space access for
> the busses advertised in the MCFG table, and fall back on the traditional method
> for other busses.
> Make i386 and amd64 only use memory mapped extended PCIe config space access for
> the buses advertised in the MCFG table, and fall back on the traditional method
> for other buses.
> 
> When ssh(1) exit because of a ServerAliveTimeout, mention the hostname that
> caused it.
> When ssh(1) exits because of a ServerAliveTimeout, mention the hostname that
> caused it.
> 
> Make acpi(4) MCFG ignores tables with an empty bus range.
> Make acpi(4) MCFG ignore tables with an empty bus range.
> 
> Fixed properly PR6009.
> Properly fixed PR6009.
> 
> Make rc.subr(8) check for a login class that match the rc script name and use it
> if it is find, otherwise use daemon.
> Make rc.subr(8) check for a login class that matches the rc script name and use
> it if it is found, otherwise use daemon.
> 
> Prevent some NULL pointer dereference in osiop(4).
> Prevent some NULL pointer dereferences in osiop(4).
> 
> Prevent issuing a m88k cache maintenance operation until the last one is not
> completed.
> Should the "not" be there?
> Prevent issuing a m88k cache maintenance operation until the last one is completed.
> 
> Added a param to uvm_map_hint to not skip over the heap, and use it as a last
> resort if mmap fails otherwise to enable more complete address space utilization.
> Not sure ... might be ...
> Added a param to uvm_map_hint to not skip over the heap, and use it as a last
> resort if mmap otherwise fails to enable more complete address space utilization.
> 
> Prevent pms(4) for crashing when doing an ioctl(2) (eg wsconsctl -a) before
> first activation (wsmoused or X).
> Prevent pms(4) from crashing when doing an ioctl(2) (eg wsconsctl -a) before
> first activation (wsmoused or X).
> 
> Fixed a tmux(1) crash due to sides effects server_kill_window with RB_FOREACH.
> Fixed a tmux(1) crash due to side effects of server_kill_window with RB_FOREACH.
> 
> Prevent m88k systems from sharing their exception page with the one used by the
> firmware at address 0. Make it possible to caught NULL pointer dereference.
> Prevent m88k systems from sharing their exception page with the one used by the
> firmware at address 0. Make it possible to catch NULL pointer dereferences.
> 
> Make tmux(1) store it's sessions in a red-black tree by name rather than a list.
> Make tmux(1) store its sessions in a red-black tree by name rather than a list.
> Make ucycom(4), utrh(4) and utwitch(4) the last drivers using there own macros
> instead of usb_lookup().
> Make ucycom(4), utrh(4) and utwitch(4) the last drivers using their own macros
> instead of usb_lookup().
> 
> SECURITY FIX: Insufficent initialization of the pf rule structure in the ioctl
> handler may allow userland to modify kernel memory. By default root privileges
> are needed to add or modify pf rules.
> SECURITY FIX: Insufficient initialization of the pf rule structure in the ioctl
> handler may allow userland to modify kernel memory. By default root privileges
> are needed to add or modify pf rules.
> 
> Make math(3) HUGE_VAL, HUGE_VALF, HUGE_VALL, INFINITY, NAN expand to the
> constant expressions with the help of GCC version posterior to 3.3.
> Make math(3) HUGE_VAL, HUGE_VALF, HUGE_VALL, INFINITY, NAN expand to the
> constant expressions with the help of GCC version prior to 3.3.
> 
> Make tmux(1) read ${X} environment variables in strings and $HOME from the
> global environment rather than getenv, this allows them to be updated during the
> configuration file.
> What does "during the configuration file" mean?
> 
> Added rsu(4), a driver for Realtex RTL8188SU, RTL8191SU and RTL8192SU 802.11n
> USB devices. Enabled it in i386 and amd64 GENERIC kernels.
> Added rsu(4), a driver for Realtek RTL8188SU, RTL8191SU and RTL8192SU 802.11n
> USB devices. Enabled it in i386 and amd64 GENERIC kernels.
> 
> Fixes panic on bridge(4) destroy.
> Fixed panic on bridge(4) destroy.
> 
> Fixed lynx(1) SSL error when going to sites using wildcards certificates.
> Fixed lynx(1) SSL error when going to sites using wildcard certificates.
> 
> Introduced ifa_update_broadaddr in ifmedia(4) to update an ifaddrs broadcast
> address,
> Introduced ifa_update_broadaddr in ifmedia(4) to update an ifaddr's broadcast
> address
> 
> Prevent revision 0xc4 and earlier of the Acer Labs M5229 UDMA IDE from doing DMA
> for LBA48 commands. It makes large disks slower but working properly.
> Prevent revision 0xc4 and earlier of the Acer Labs M5229 UDMA IDE from doing DMA
> for LBA48 commands. It makes large disks slower but work properly.
> 
> Added security fix for the TrueType FX font variation loader of FreeType in
> Xenacara (CVE-2010-3855).
> Added security fix for the TrueType FX font variation loader of FreeType in
> Xenocara (CVE-2010-3855).
> 
> Added urtwn(4), a driver for Realtex RTL8188CU/RTL8192CU 802.11n USB devices.
> Added urtwn(4), a driver for Realtek RTL8188CU/RTL8192CU 802.11n USB devices.
> 
> Makt systat(1) print relative number of mbuf(9) livelock that happened between
> two systat updates instead of showing absolute figures.
> Make systat(1) print relative number of mbuf(9) livelocks that happened between
> two systat updates instead of showing absolute figures.
> 
> Make ldapd(8) truncate additional incomplete requests when the input buffer
> already get two complete requests. Fixes the number of bytes consumed from the
> input buffer.
> Make ldapd(8) truncate additional incomplete requests when the input buffer
> already received two complete requests. Fixes the number of bytes consumed from
> the input buffer.
> 
> Fixed a few logic errors in misp64 comparison instruction emulation: make sure
> the less than relation is correctly computed, and check for both operands being
> signalizing NaNs, instead of only the first NaN found, to decide whether to
> raise an invalid exception or not.
> Should be mips64?  And signaled?
> Fixed a few logic errors in mips64 comparison instruction emulation: make sure
> the less than relation is correctly computed, and check for both operands being
> signaled NaNs, instead of only the first NaN found, to decide whether to raise
> an invalid exception or not.
> 
> Improved isakmpd(8) interoperability with Fortigates devices.
> Improved isakmpd(8) interoperability with Fortigate's devices.
> 
> In ikectl(8): added a -q (quiet) command line option that will be use by ike CA
> to set openssl batch mode, allow to specify the initial CA password on the
> command line, allow to create certificate for clientAuth or serverAuth only.
> In ikectl(8): added a -q (quiet) command line option that will be used by ike CA
> to set openssl batch mode, allow specifying the initial CA password on the
> command line, allow creating certificate for clientAuth or serverAuth only.
> 
> Make ospfd(8) build the sum of the ls_checksums when counting the lsa. It
> permits to quickly verify if two LSDBs are in sync.
> Make ospfd(8) build the sum of the ls_checksums when counting the lsa. It allows
> quick verification that two LSDBs are in sync.
> 
> Deactivate all vscsi(4) childs when it's closed so they don't retry the commands
> still on the queue when killed.
> Deactivate all vscsi(4) children when its closed so they don't retry the
> commands still on the queue when killed.
> 
> Introduced a DKF_NOLABELREAD flag in disk(9), which is used by a device to prevent
> prevent what?  I think the point ABOVE this line is what it is meant to prevent
> "automatic reading of disklabel on attach. Applied it in fd(4) and cd(4). "
> 
> Make inteldrm(4) always unconditionally set the C3_LP_WRITE_ENABLE bit in
> ARB_STATE on gen3 hardware. Prevent some 945 chipset from hanging.
> Make inteldrm(4) always unconditionally set the C3_LP_WRITE_ENABLE bit in
> ARB_STATE on gen3 hardware. Prevents some 945 chipsets from hanging.
> 
> Avoid USBD_IOERROR in uhub(4) by determining the size of the buffer for USB hub
> interrupt pipe with bNbrPorts in hup descriptor.
> Is that last hup meant to be hub?
> Avoid USBD_IOERROR in uhub(4) by determining the size of the buffer for USB hub
> interrupt pipe with bNbrPorts in hub descriptor.
> 
> Added watchdog suspend/resume logic to glxpcib(4/LOONGSON) and make
> glxpcib(4/LOONGSON) save and restore it's configuration on suspend/resume.
> Added watchdog suspend/resume logic to glxpcib(4/LOONGSON) and make
> glxpcib(4/LOONGSON) save and restore its configuration on suspend/resume.
> 
> Make mandoc(1) format multiple authors (.%A) in the same way as groff: two
> authors with "and", but without a comma; three or more with commata, and an
> "and" before the last one.
> commata?  Should be "commas"?
> 
> Make snmpd(8) return error on GET requests without an instance identifier, both
> for tables and scalar values. Fixes PR6468 by not calling table get functions
> with an unexpectedly short OID. Also fixes PR6071. Scalar variables without an
> instance specified now returns a noSuchInstance error. GetNext requests
> correctly returns the .0 instance but it must be specified explitely: $ snmpget
> -v2c -c public localhost SNMPv2-MIB::sysDescr.0
> Make snmpd(8) return error on GET requests without an instance identifier, both
> for tables and scalar values. Fixes PR6468 by not calling table get functions
> with an unexpectedly short OID. Also fixes PR6071. Scalar variables without an
> instance specified now returns a noSuchInstance error. GetNext requests
> correctly returns the .0 instance but it must be specified explicitly: $ snmpget
> -v2c -c public localhost SNMPv2-MIB::sysDescr.0
> 
> Added firmware some missing firmware to install media: ral(4) RT2860 for i386,
> amd64 and sparc64; run(4) RT2870 and RT3071 for landisk, macppc, socppc and sparc64.
> Added some missing firmware to install media: ral(4) RT2860 for i386, amd64 and
> sparc64; run(4) RT2870 and RT3071 for landisk, macppc, socppc and sparc64.
> 
> Fix lazy binding issue on mips64 run-time link-editor. Fixes an nm(1) crash that
> happen when passed multiple object files on the command line.
> Fix lazy binding issue on mips64 run-time link-editor. Fixes an nm(1) crash that
> happens when passed multiple object files on the command line.
> 
> Prevent azalia(4) from unmuting the speaker muted by plugged headphones when the
> master volume control in unmuted.
> (is the last "in" meant to be "is"?)
> Prevent azalia(4) from unmuting the speaker muted by plugged headphones when the
> master volume control is unmuted.
> 
> Make sure local arrays of chars are word-aligned on strict alignment
> architectures with gcc(1), just like what is done for global ones. Though not
> explicitly allowed by the C standard it have been historically handle correctly
> by most C compilers.
> Make sure local arrays of chars are word-aligned on strict alignment
> architectures with gcc(1), as is done for global ones. Though not explicitly
> allowed by the C standard it has historically been handled correctly by most C
> compilers.
> 
> Fixed a file descriptor leaks in ccdconfig(8).
> Fixed file descriptor leaks in ccdconfig(8).  OR
> Fixed a file descriptor leak in ccdconfig(8).
> 
> Worked around a rare race condition that can happen if daily(8) zaps old files
> from /var/tmp while pkg_add(1) is installing or updating a package and doesn't
> yet have moved nor even read some of the files in /var/tmp/pkginfo.* (typically
> +DESCR).
> Worked around a rare race condition that can happen if daily(8) zaps old files
> from /var/tmp while pkg_add(1) is installing or updating a package and haesn't
> yet moved or even read some of the files in /var/tmp/pkginfo.* (typically +DESCR).
> 
> Prevent wsconsctl(8) from ignoring errors silently if the device could not be open.
> Prevent wsconsctl(8) from ignoring errors silently if the device could not be
> opened.
> 
> Fixed a NULL pointer dereferencement in midicat(1) option parsing when `-' is
> used as filename.
> Fixed a NULL pointer dereference in midicat(1) option parsing when `-' is used
> as filename.
> 
> Make sure mandoc(1) correctly print .Lk arguments in -Tascii.
> Make sure mandoc(1) correctly prints .Lk arguments in -Tascii.
> 
> Make sure mandoc(1) throw away non-ASCII characters.
> Make sure mandoc(1) throws away non-ASCII characters.
> 
> Removed support for loading 64bits kernels from cdboot(8/HPPA).
> Removed support for loading 64 bit kernels from cdboot(8/HPPA).
> 
> Fixed a off by one bug in NTFS.
> Fixed an off-by-one bug in NTFS.
> 
> Forbade duplicate tmux(1) session names to be created.
> Forbade duplicate tmux(1) session names from being created.