src/oping.h: Don't use the ncurses internal _nc_unicode_locale() function.
The "-u" and "-U" options are now handled in a single variable so that the
last option "wins".
The "-u" and "-U" options are now handled in a single variable so that the
last option "wins".
add commandline flag for forcing or disabling unicode
note that "forcing" may not work: ncurses still expects your locale to
be correct, so maybe the force flag should also explicitely set a utf8
locale?
note that "forcing" may not work: ncurses still expects your locale to
be correct, so maybe the force flag should also explicitely set a utf8
locale?
move unicode detection to a separate function
this will ease transition to a commandline option override
this will ease transition to a commandline option override
autodetect unicode, and fallback to ACS scancodes on failure
this is fairly ugly - we use a hidden ncurses function to detect
unicode (but at least we don't reimplement it), then we mess around
with the character arrays to display the rights symbols
i am not sure i like this, but i prefer to braindump than to stall, so
let's move forward.
this is fairly ugly - we use a hidden ncurses function to detect
unicode (but at least we don't reimplement it), then we mess around
with the character arrays to display the rights symbols
i am not sure i like this, but i prefer to braindump than to stall, so
let's move forward.
Refactor the graph printing into its own function.
This makes it much easier and cleaner to add a second, non-Unicode
implementation and possibly even a non-color version.
This makes it much easier and cleaner to add a second, non-Unicode
implementation and possibly even a non-color version.
switch to ncursesw in the build chain
better wrapping: add a cursor and respect inner borders
rework scaling algorithm so it covers the whole range
i did this after some stress testing, now it covers the whole range on input
i did this after some stress testing, now it covers the whole range on input
cleanup safety checks
make sure that if we go beyond the max ratio, we just mark it as one
make sure that if we go beyond the max ratio, we just mark it as one
wrap around the window once the histogram grows too big
remove a little quirk i introduced by mistake
this was breaking the border
this was breaking the border
make histogram colors with a nicer background
restore usual ping display, now that histograms are in a different window
put histograms into the independent window, to eventually restore ping times display
implement some prettyping-inspired display instead of textual
this is a first stab at porting "prettyping.sh" into noping. the idea
is to show a histogram of ping times instead of numerical values. with
some work, we could actually show both, but this focuses on
implementing the hard part (the histogram) properly.
it is very rudimentary for now:
* the math may be wrong for the size of the bar, I was mostly in a
rush to make unicode work and have something pretty quickly.
* all hosts are on the same line: each should have its own line
* the histogram should be displayed in a separate window
* it should use background colors as well
* it should fallback when the terminal is not unicode-capable
* the scaling logic is fully automatic, which necessarily gives weird
results at first
this requires switching to the ncursesw library and other unicode
ncurses hackery described in this post:
http://newsgroups.derkeiler.com/Archive/Rec/rec.games.roguelike.development/2010-09/msg00050.html
pretty ping is a awk/bash script, MIT-licensed, available here:
https://bitbucket.org/denilsonsa/small_scripts/src/tip/prettyping.sh
this is a first stab at porting "prettyping.sh" into noping. the idea
is to show a histogram of ping times instead of numerical values. with
some work, we could actually show both, but this focuses on
implementing the hard part (the histogram) properly.
it is very rudimentary for now:
* the math may be wrong for the size of the bar, I was mostly in a
rush to make unicode work and have something pretty quickly.
* all hosts are on the same line: each should have its own line
* the histogram should be displayed in a separate window
* it should use background colors as well
* it should fallback when the terminal is not unicode-capable
* the scaling logic is fully automatic, which necessarily gives weird
results at first
this requires switching to the ncursesw library and other unicode
ncurses hackery described in this post:
http://newsgroups.derkeiler.com/Archive/Rec/rec.games.roguelike.development/2010-09/msg00050.html
pretty ping is a awk/bash script, MIT-licensed, available here:
https://bitbucket.org/denilsonsa/small_scripts/src/tip/prettyping.sh
Bump ABI version; update ChangeLog.
Mac OX S 10.7 compile fixes.
I've attached a patch that's required to be able to compile liboping on
MacOS 10.7 (Lion). The issue is that the IPv6 API's are changing from
RFC 2292 to 3542, and Apple's headers require you to state which API you
want to use by defining a constant. I've changed it to use either based
on the presence of constants, but to prefer 3542 on MacOS.
One problem I see with this is that if you have both multicast and
unicast TTL's set, then this won't report both of them, but I'm not sure
if that's even valid and it would require more significant rework.
Also, I noticed you have fixed the IP_RECVTOS issue in head. That patch
is also required to be able to compile on Solaris and MacOS.
Signed-off-by: Florian Forster <octo@verplant.org>
I've attached a patch that's required to be able to compile liboping on
MacOS 10.7 (Lion). The issue is that the IPv6 API's are changing from
RFC 2292 to 3542, and Apple's headers require you to state which API you
want to use by defining a constant. I've changed it to use either based
on the presence of constants, but to prefer 3542 on MacOS.
One problem I see with this is that if you have both multicast and
unicast TTL's set, then this won't report both of them, but I'm not sure
if that's even valid and it would require more significant rework.
Also, I noticed you have fixed the IP_RECVTOS issue in head. That patch
is also required to be able to compile on Solaris and MacOS.
Signed-off-by: Florian Forster <octo@verplant.org>
guard use of IP_RECVTOS which is unavailable on kfreebsd
configure.ac: autoconf 2.65 works just fine, too.
README: Add note about capabilities.
Manual pages: Update the copyright information.
Bump version to 1.6.2; Update ChangeLog.
configure.ac: autoconf 2.67 works just fine.
Merge pull request #1 from barak/master
upstream update + tweaks
upstream update + tweaks
git ignore more build debris
Update configure.ac and Makefile.am per autotools
Tweak configure.ac and Makefile.am per autoreconf instructions.
libtoolize: Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
libtoolize: rerunning libtoolize, to keep the correct libtool macros in-tree.
libtoolize: Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
libtoolize: `AC_PROG_RANLIB' is rendered obsolete by `LT_INIT'
Accept autoupdate modifications.
Manually integrate portions of autoscan configure.scan.
Tweak configure.ac and Makefile.am per autoreconf instructions.
libtoolize: Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
libtoolize: rerunning libtoolize, to keep the correct libtool macros in-tree.
libtoolize: Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
libtoolize: `AC_PROG_RANLIB' is rendered obsolete by `LT_INIT'
Accept autoupdate modifications.
Manually integrate portions of autoscan configure.scan.
src/Makefile.am: Make setcap and chmod fault-tolerant.
This is required because this weird Debian fakeroot utility gives the
impression as if the install was run by root, when in fact it isn't.
What good this should do is beyond me.
This is required because this weird Debian fakeroot utility gives the
impression as if the install was run by root, when in fact it isn't.
What good this should do is beyond me.
src/liboping.c: Fix a compiler warning
… regarding a non-static format string. Thanks to Brian Edwards for
reporting this problem.
… regarding a non-static format string. Thanks to Brian Edwards for
reporting this problem.
Bump version to 1.6.1; Update ChangeLog.
Update and unify copyright headers.
Build system: Automatically set capabilities flag / set-UID bit when installing as root.
noping: Fix compatibility with ncurses 5.8.
Hello,
My name is Gaetan and I maintain liboping in Arch Linux; our distro is
in the process of switching to ncurses-5.8, and I wanted to report that
I had to patch liboping to make noping work with the newer ncurses.
Specifically, if I update ncurses and run the old noping binary, then
nothing is displayed in the ncurses window (although the summary output
displayed on stdout at the end is correct). Same if I recompile noping
against the new ncurses library.
The reason seems to be that calling newwin() with ncols=0 does not
produce a full width window anymore. Anyway, the simple patch attached
fixed this issue for me.
Cheers.
Signed-off-by: Florian Forster <octo@verplant.org>
Hello,
My name is Gaetan and I maintain liboping in Arch Linux; our distro is
in the process of switching to ncurses-5.8, and I wanted to report that
I had to patch liboping to make noping work with the newer ncurses.
Specifically, if I update ncurses and run the old noping binary, then
nothing is displayed in the ncurses window (although the summary output
displayed on stdout at the end is correct). Same if I recompile noping
against the new ncurses library.
The reason seems to be that calling newwin() with ncols=0 does not
produce a full width window anymore. Anyway, the simple patch attached
fixed this issue for me.
Cheers.
Signed-off-by: Florian Forster <octo@verplant.org>
src/oping.c: Fix a compiler warning about uninitialized variables.
Some verions of GGC don't notice that usage_qos_exit() won't return and
complain about the variables "prec" and "class" being used uninitalized.
This commit adds the "noreturn" attribute to that function and
initializes the variables. Thanks to James Bromberger for reporting this
problem.
Some verions of GGC don't notice that usage_qos_exit() won't return and
complain about the variables "prec" and "class" being used uninitalized.
This commit adds the "noreturn" attribute to that function and
initializes the variables. Thanks to James Bromberger for reporting this
problem.
Bump version to 1.6.0; Update ChangeLog.
README: Remove whitespace.
src/liboping.c: Change coding style a bit.
liboping: use SO_TIMESTAMP when available
Current implementation of liboping relies on getting enough CPU shares
and getting those in a timely manner in order to determine correct ping
response times.
Avoid this limitation when kernel is able to provide packet reception
times using SO_TIMESTAMP. (as is done by iputils's ping utility)
Signed-off-by: Florian Forster <octo@verplant.org>
Current implementation of liboping relies on getting enough CPU shares
and getting those in a timely manner in order to determine correct ping
response times.
Avoid this limitation when kernel is able to provide packet reception
times using SO_TIMESTAMP. (as is done by iputils's ping utility)
Signed-off-by: Florian Forster <octo@verplant.org>
README: Document Linux capabilities and UNIX set-UID root solutions.
oping(8): Document the new SetUID behavior in connection with the "-f" option.
Bump version to 1.5.1; Update ChangeLog.
src/oping.c: Mark unused argument to avoid compiler warnings.
src/oping.c: Remove __attribute__ flags when using non-GNU compiler.
src/liboping.c: Always set the QoS member. Default to 0.
This should fix the following warning / error:
liboping.c:599: warning: comparison is always true due to limited
range of data type
Thanks to James Bromberger for reporting this!
This should fix the following warning / error:
liboping.c:599: warning: comparison is always true due to limited
range of data type
Thanks to James Bromberger for reporting this!
oping: Add alias for the "Voice Admit" DSCP.
Bump version to 1.5.0; Update ChangeLog.
ping_iterator_get_info(3): Document the "PING_INFO_RECV_QOS" flag.
src/oping.h: Change the default payload message.
Merge branch 'vm/qos'
oping(8): Document the displayed "qos=" field.
src/oping.c: Only print QoS information if it's not 0/0.
AUTHORS: Add Vladimir Melnikov.
src/oping.c: Add function for formatted printing of the QoS byte.
oping(8): Document the "be" and "cs<n>" DSCP abbreviations.
Rename "ToS" to "QoS" everywhere, to account for DSCP being used in IPv6 …
… and recommended for IPv4.
… and recommended for IPv4.
src/oping.c: Remove useless "strlen".
src/oping.c: Add the "be" and "cs[0-7]" DSCPs.
oping(8): Document the "-Q" option.
src/oping.c: Implement named arguments for DiffServ flags.
oping: Improve parsing of the "-z" argument.
Fix printing the ToS value.
Apparently "%#04x" prints 0 as "0000" instead of "0x00". *sigh*
Apparently "%#04x" prints 0 as "0000" instead of "0x00". *sigh*
ping_set_tos: Improve error handling.
ping_receive_one: Fix a typo.
src/liboping.c: Fix building with debugging enabled.
Rename "PING_INFO_TOS" → "PING_INFO_RECV_TOS".
ping_setopt(3): Document the new "PING_OPT_TOS" option.
src/liboping.c: Set the "IP_RECVTOS" and "IPV6_RECVTCLASS" socket options.
src/liboping.c: Use an uint8_t for storing the ToS.
Implement support for QoS / ToS fields.
Hello Florian,
I've made some functionality extension to debug QoS in the network -
ability to set tos (present in most standard "pings") and to verify
received tos field (didn't see any ping command with this feature).
May be it will need to someone else.
Best regards,
Vladimir
Hello Florian,
I've made some functionality extension to debug QoS in the network -
ability to set tos (present in most standard "pings") and to verify
received tos field (didn't see any ping command with this feature).
May be it will need to someone else.
Best regards,
Vladimir
src/Makefile.am: Use "LDADD" for applications, "LIBADD" for the library.
Use "LDFLAGS" only for "-version-info".
Use "LDFLAGS" only for "-version-info".
Relicense: Use the LGPL v2.1 (or later) for liboping.
The "oping" and "noping" utilities are still GPL v2.
Sebastian Harl, the only other copyright holder on the library, agrees
with this license change.
The "oping" and "noping" utilities are still GPL v2.
Sebastian Harl, the only other copyright holder on the library, agrees
with this license change.
oping(8): Document the colors used by noping.
Bumbed version to 1.4.0, updated ChangeLog.
oping(8): Document "noping", too.
README: Document "oping" and "noping".
README: Convert to Unicode.
src/oping.c: Define NCURSES_OPAQUE.
src/oping.c: Resizing the scrolling region seems to work with this parameters.
I don't understand what I'm doing different to previously, but whatever
it is it seems to do the trick.
I don't understand what I'm doing different to previously, but whatever
it is it seems to do the trick.
noping: Further cleanup of the curses stuff.
noping: Use colors to highlight "unusual" response times.
noping: Print host name in bold.
noping: Handle resizing of X windows gracefully.
src/oping.c: Move "context_get_packet_loss" into a separate function.
src/oping.c: Move calculation of average and stddev to separate functions.
src/oping.c: Add ncurses based front-end.
src/oping.c: Move printing header and footer into separate functions.
contrib/liboping.spec: Added "provided as is" header.
contrib/liboping.spec: Added example spec file.
I'm using collectd and the ping plugin for a project I'm working on. I noticed
that collectd is distributing a starter spec file for building RPMs but that
there wa s no such spec file for liboping. I created one which I have attached
to this email. I'm not too familiar with creating RPMs but I think it's a good
start and I was able to get a package built and installed using it. I'm giving
it to you so if you wish to put in on your website for others to start with
when creating packages t hen you may. I hope others may find this useful.
I'm using collectd and the ping plugin for a project I'm working on. I noticed
that collectd is distributing a starter spec file for building RPMs but that
there wa s no such spec file for liboping. I created one which I have attached
to this email. I'm not too familiar with creating RPMs but I think it's a good
start and I was able to get a package built and installed using it. I'm giving
it to you so if you wish to put in on your website for others to start with
when creating packages t hen you may. I hope others may find this useful.
minor doc fixes for liboping
Hello,
I was reading the documentation of your liboping library as I'd like to
use it for a small project of mine and couldn't help noticing a few typos
in the docs, please find attached a tiny patch fixing them.
Thanks for writing this library!
VZ
Signed-off-by: Florian Forster <octo@verplant.org>
Hello,
I was reading the documentation of your liboping library as I'd like to
use it for a small project of mine and couldn't help noticing a few typos
in the docs, please find attached a tiny patch fixing them.
Thanks for writing this library!
VZ
Signed-off-by: Florian Forster <octo@verplant.org>
ChangeLog: Oops, there were actually a couple more changes in that release.
Bumped version to 1.3.4 and updated ChangeLog.
src/liboping.c: ping_receive_all: Clear the timer when a FD signals an error.
The loop likely never ends if the error isn't handled, so the least we
can do is ignore the file descriptor in the future.
The loop likely never ends if the error isn't handled, so the least we
can do is ignore the file descriptor in the future.
bindings/perl: Bumped version to 1.21 and updated Changes.
bindings/perl/Oping.xs: Fix an off-by-one error when allocating a Perl string.
In the previous code the null byte was included in the Perl string,
making the string handling awkward.
In the previous code the null byte was included in the Perl string,
making the string handling awkward.
oping: Temporarily drop privileges if supported by the system.
This tries to solve the read-from-any-file-issue more elegantly by
dropping privileges first thing after start-up and regaining privileges
just to open the sockets.
This has the advantage that the "-f" option is available for all users
again and files are opened according to *their* permissions.
Systems not supporting the _POSIX_SAVED_IDS feature still behave as
before, i. e. if real and effective user IDs don't match, only "-" may
be specified as input file.
This tries to solve the read-from-any-file-issue more elegantly by
dropping privileges first thing after start-up and regaining privileges
just to open the sockets.
This has the advantage that the "-f" option is available for all users
again and files are opened according to *their* permissions.
Systems not supporting the _POSIX_SAVED_IDS feature still behave as
before, i. e. if real and effective user IDs don't match, only "-" may
be specified as input file.
ChangeLog: Updated the changelog to reflect the last change.
oping: Allow “-f -” even if uid and euid differ.
Thanks to Sebastian for the idea :)
Thanks to Sebastian for the idea :)
Bumped version to 1.3.3; Updated ChangeLog.
oping: Disable the “-f” option when real and effective user IDs don't match.
Thanks to Steve Kemp who reported this issue as Debian bug #548684.
<http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=548684>
Thanks to Steve Kemp who reported this issue as Debian bug #548684.
<http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=548684>
Bumped version to 1.3.2; Updated ChangeLog.
src/oping.h: Do not check HAVE_*_H for system headers.
Imho, this is not a good thing to do for an installed header since it requires
every user to define those macros as well.
Imho, this is not a good thing to do for an installed header since it requires
every user to define those macros as well.