Whew, it was nip-and-tuck there for awhile but finally
we solved the SIGWINCH overload problem one finds with
most X window managers. Now if a window manager should
try to inundate ol' top with repeated SIGWINCH signals
they won't even be received so can't impact us at all.
And we achieve this miracle having never even issued a
sigprocmask, so all the top code executes with signals
totally unblocked. Intuition suggests it probably rubs
even more salt in the wound, but au contraire mon ami!
The key to our success was simply trading the 'select'
call for its cousin 'pselect'. Not only does that call
provide nanosecond granularity (vs. the former's usec)
but it takes a sigset_t parm which can then atomically
block the troublesome SIGWINCH guy until user input or
optional timeout. Net result? No more signal overload!
Now, if only we could just coax all terminal emulators
into one identical standard buffering scheme plus find
some way to emulate the most recent SIGWINCH, it would
be perfect. We would then obviate the user requirement
of typing yet 1 more key before seeing proper results.
(everything is perfectly justified plus right margins)
(are completely filled, but of course it must be luck)
Reference(s):
http://www.freelists.org/post/procps/unwanted-topinspect-window-enclosure-with-the-terminal-size-changehttp://www.freelists.org/post/procps/Sourceforge-project,7
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit primarily involves renaming functions plus
reorganizing logic in preparation for the next changes
which will hopefully yield the 'final solution' to the
excessive SIGWINCH signals under most window managers.
In this specific patch, the most significant change is
the introduction of a new 'ioa' function (io avail) to
focus all logic dealing with unsolicited user keyboard
input and exposed to signals and/or optional timeouts.
That new function is where our signal overload will be
ultimately defeated, if it is at all humanly possible.
(everything is perfectly justified plus right margins)
(are completely filled, but of course it must be luck)
Reference(s):
http://www.freelists.org/post/procps/unwanted-topinspect-window-enclosure-with-the-terminal-size-change
Signed-off-by: Jim Warner <james.warner@comcast.net>
Prior to this commit, top has always taken the easiest
(safest?) approach when dealing with those troublesome
SIGWINCH interrupts. Whenever the user was on a screen
other than the main display, any signal received would
force an immediate exit, returning to the main screen.
With these changes, top will retain the user's current
position regardless of what screen he/she was viewing.
In support the following additional changes were made:
* the initial help screen requires an explicit end key
` not 'any other key' formerly used to request an exit
* the colors mapping screen instructions were improved
* ^Z response was made immediate, eliminating the flag
* the sigaction's SA_RESTART flag had to be eliminated
* sigprocmask logic was normailize to the bare minimum
* the POSIX.1-2004 async-signal safe functions used in
` the signal handlers were acknowledged and documented
(everything is perfectly justified plus right margins)
(are completely filled, but of course it must be luck)
Reference(s):
http://www.freelists.org/post/procps/unwanted-topinspect-window-enclosure-with-the-terminal-size-change
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit just addresses the following minor issues:
. eliminate the leading tab character upon error exits
. standardized single key input as 'keyin', not 'chin'
. symbolic keys changed to guarantee no negative value
. placed most 'case' statement labels on a unique line
. standardized lvalue/rvalue convention in while loops
. fixed prototype declaration in the 'debug_END' macro
(everything is perfectly justified plus right margins)
(are completely filled, but of course it must be luck)
Signed-off-by: Jim Warner <james.warner@comcast.net>
Some testing under a new distro revealed what appeared
to be a broken top Inspect request. When the selection
was made, the resulting output was scrambled/scrunched
at the bottom of the screen (as if ^J's were missing).
This anomaly surfaced under Fedora-18 which happens to
use the above ncurses version. The bug was not present
in version 5.9.20120714 (available with openSUSE 12.2)
or the more widely available version of: 5.9.20110404.
It has now been confirmed that this problem originated
in version 5.9.20120825. It was then that buffering of
output was changed from stdio to some internal ncurses
scheme so as to avoid problems with its SIGTSTP logic.
Thanks to a very prompt response from Thomas E. Dickey
we also learned that contrary to the documentation the
putp logic does not call putchar internally. Thus, the
single putchar that Inspect was employing was actually
mixing 2 different buffering schemes: ncurses & stdio.
Thus, from now on we'll use putp() exclusively and try
to achieve single char output as efficiently as we can
while meeting that putp() string argument requirement.
(everything is perfectly justified plus right margins)
(are completely filled, but of course it must be luck)
Reference(s):
https://bugzilla.redhat.com/892674
Signed-off-by: Jim Warner <james.warner@comcast.net>
There existed a small chance that the display could be
corrupted when a search string was found within a row.
For that to happen, conditions like these were needed:
. a very short Locate string was active in some window
. the string matched part of a terminfo <esc> sequence
. that sequence was used in highlighting running tasks
. the 'x' toggle was active (sort column highlighting)
One solution to this potential problem was to manually
turn off sort column highlighting before using Locate.
But rather than rely on a user remedy, we'll automate.
Since other top provisions were already being enforced
when Locate was in use (off 'i' and/or 'u'/'U'), we'll
now also force column highlighting off when the search
string in a given window is not empty. However, unlike
idle tasks and user filtering, when that search string
*is* emptied, we restore highlighting for that window.
(everything is perfectly justified plus right margins)
(are completely filled, but of course it must be luck)
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit just addresses the following minor issues:
. restored both lost end-of-job reporting capabilities
. added missing initializers to the DEF_RCFILE #define
. added 'nls_maybe' eye-catcher to the 'Scaled_sfxtab'
. removed a now superfluous 'READMINSZ' assertion test
. man document references to 'top' are more consistent
(everything is perfectly justified plus right margins)
(are completely filled, but of course it must be luck)
Signed-off-by: Jim Warner <james.warner@comcast.net>
This potential problem is caused by frequently spawned
and short lived tasks which happen to sort above a row
containing a match from an active Locate request. It's
most likely to be visible when under Forest View mode.
This commit will eliminate a potential duplicated row.
Signed-off-by: Jim Warner <james.warner@comcast.net>
Both of the 'file' and 'pipe' Inspect read buffers are
always 2048 bytes bigger than is actually needed which
provided for some slightly simplified row paint logic.
However, with no real rcfile inspect entries, and thus
operating in 'demo' mode, the allocated buffer is only
2048 bytes total. This can produce a valgrind warning.
This commit simply puts the inspect 'demo' buffer on a
par with other allocated real buffers (an extra 2048).
(everything is perfectly justified plus right margins)
(are completely filled, but of course it must be luck)
Signed-off-by: Jim Warner <james.warner@comcast.net>
In a effort to anticipate a potential future wishlist,
the recent '-o' sort override command line switch will
now support an override of the sort direction as well.
By prepending a '+' or '-' to any valid field name the
user will be able to guarantee a specific desired sort
direction. The '+' forces a high-to-low (normal) order
while a '-' reverses that to yield a low-to-high sort.
Without this addition users would be left to the mercy
of whatever was last specified for Curwin as reflected
in the rcfile or top's default of a high-to-low order.
(everything is perfectly justified plus right margins)
(are completely filled, but of course it must be luck)
Signed-off-by: Jim Warner <james.warner@comcast.net>
The recent change to task area memory scaling was just
a little short of optimum in its consistency and upper
limits. In fact, top could only scale memory fields up
to a maximum of 99.9999 TiB (with VIRT a little more).
While that seems like more than enough it was actually
artificially low, due to an unnecessary decimal place.
So, this commit standardizes both precision and widths
to achieve a minimum amount of scaling beyond the user
requested target plus reclaim some horizontal spacing.
. VIRT & DATA are now 7 bytes wide (not eight and six)
. other memory fields are 6 wide (were formerly seven)
. as before, KiB shows whole numbers only (no decimal)
. MiB, for its precision, shows a single decimal place
. all other memory ranges display three decimal places
The net result is a more homogeneous display with less
forced scaling and the recovery of three lost columns.
(now that we know a '.' + 2 spaces is squeezed to one)
(everything's perfectly justified, but it's just luck)
Reference(s);
http://www.freelists.org/post/procps/top-enhancements-i-hope,1
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit will allow user filtering as long as there
is a valid number representing a potential user ID. It
will serve, for example, chroot environments where the
specific user may be unknown to a host, or vice versa.
(now that we know a '.' + 2 spaces is squeezed to one)
(everything's perfectly justified, but it's just luck)
Reference(s):
http://www.freelists.org/post/procps/top-enhancements-i-hope,1
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit was adapted from the openSUSE patch, whose
original comments appear between the --- marker below.
It should be noted, however, that the original changes
were ineffective (wasted) because 'Batch' mode had not
yet been set when signals were being checked and their
handlers established. Thus, SIGHUP was never bypassed.
(of course, only our comments are perfectly justified)
------------------------------------------------------
Do not setup SIGHUP signal handler if we are in the batch mode
Top enables a signal handler for the SIGHUP signal (loss of terminal). While
this makes sense for top's default interactive mode, it doesn't make any sense
for batch mode. If you run top in nohup just to collect data over time and
disconnect top finishes which is not what one would expect.
------------------------------------------------------
Reference(s):
http://www.freelists.org/post/procps/top-enhancements-i-hope,1
Signed-off-by: Jim Warner <james.warner@comcast.net>
Everyone has either accepted this potential distortion
or patched top to eliminate it. Now, the time has come
to regain some consistency when calculating that %CPU.
We'll now limit such values to: 100.0 * total threads.
And, it took way too long to address this little flaw.
(now that we know a '.' + 2 spaces is squeezed to one)
(everything's perfectly justified, but it's just luck)
Reference(s):
http://www.freelists.org/post/procps/top-regression-reportshttp://www.freelists.org/post/procps/top-regression-reports,1
Signed-off-by: Jim Warner <james.warner@comcast.net>
The delay interval is kept in the rcfile in a floating
point format and is thus susceptible to changes in the
locale between invocations. So values written as #,###
could not be read if a new locale uses decimal points.
This commit takes control of our own decimal point and
will henceforth make top immune to locale switcharoos.
(now that we know a '.' + 2 spaces is squeezed to one)
(everything's perfectly justified, but it's just luck)
Reference(s):
http://www.freelists.org/post/procps/top-has-a-localedpendent-config-file-toprchttp://lists.opensuse.org/opensuse-bugs/2012-12/msg01466.html
Signed-off-by: Jim Warner <james.warner@comcast.net>
With the increased width and precision of memory and %
columns, the proliferation of 0's when there's nothing
to report seems like a distraction versus useful data.
This commit introduces the '0' toggle which can either
display or suppress those zeros. And, like the scaling
states this new state is also preserved in the rcfile.
(now that we know a '.' + 2 spaces is squeezed to one)
(everything's perfectly justified, but it's just luck)
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit is an unrequested outgrowth of the earlier
change dealing with summary area memory field scaling.
That user selectable scaling provision is now extended
to include 6 (at present) task oriented memory fields.
The new companion 'e' (lower case) interactive command
has been added and, like the 'E' command, it can cycle
each of the currently displayed memory columns between
KiB through TiB. There are, however, some differences.
Where '+' indicates summary area truncation at a given
radix, task memory fields are automatically scaled for
their column. Thus, not all rows use the same scaling.
And, while summary area field widths were not changed,
the task memory columns were widened in order to offer
more meaningful data when the radix was increased. The
precision is automatically increased in step with each
radix: MiB displays 2 decimal places, GiB 3 and TiB 4.
To compliment that additional precision, both the %CPU
and %MEM fields were widened by 1 column and now offer
precision up to 3 decimal places. But, unique to %CPU,
widening could already have occurred due to the number
of processors in some massively parallel boxes. At any
rate, total extra width for both memory and percentage
fields could amount to twenty (precious) columns more.
So for both the memory and % fields the original width
(along with loss of precision) can be restored via new
compiler conditionals which this commit also provides.
p.s. and it will be rcfile preserved for any restarts!
(now that we know a '.' + 2 spaces is squeezed to one)
(everything's perfectly justified, but it's just luck)
Reference(s):
http://www.freelists.org/post/procps/top-regression-reports
Signed-off-by: Jim Warner <james.warner@comcast.net>
Earlier this year, the switch from KiB to Mib as shown
in top's summary area was postponed to those occasions
when KiB exceeded 8 digits. In hindsight that may have
moved top in the wrong direction, given the difficulty
of digesting such large numbers of digits at a glance.
This commit adds a new 'E' interactive command used to
cycle the displayed memory amounts ranging from KiB to
TiB. Thus, users can choose the radix they wish shown.
p.s. and it will be rcfile preserved for any restarts!
(now that we know a '.' + 2 spaces is squeezed to one)
(everything's perfectly justified, but it's just luck)
Reference(s):
http://www.freelists.org/post/procps/top-regression-reports
commit 95f2201730
Author: James Cloos <cloos@jhcloos.com>
Date: Mon Feb 6 00:00:00 2012 -0500
Signed-off-by: Jim Warner <james.warner@comcast.net>
This is doubtless not quite the solution envisioned by
the submitter, but 'V' should probably never have been
used as an alias for 'v' now that we have forest-view.
(now that we know a '.' + 2 spaces is squeezed to one)
(everything's perfectly justified, but it's just luck)
Reference(s):
Bug-Redhat: https://bugzilla.redhat.com/848290
Signed-off-by: Jim Warner <james.warner@comcast.net>
When 'type' was added to the alternate status line for
the '=' key, the need for the demonstration 'type' was
overlooked. This commit avoids the '(null)' potential.
(now that we know a '.' + 2 spaces is squeezed to one)
(everything's perfectly justified, but it's just luck)
Reference(s);
commit 3b2b9a95e6
Signed-off-by: Jim Warner <james.warner@comcast.net>
With the recent inspect search highlight provisions in
place, the lack of highlighting in task based searches
has grown from being only irritating to a real defect.
Thus, this commit introduces parallel functionality to
those searches initiated within a visible task window.
And just as separate inspect searches are possible for
each selection, per window task searches are provided.
However, it should be noted that there are differences
between task based searches and inspect type searches:
* There is no concept of out-of-view data when dealing
. with task rows -- if the data can't bee seen, it has
. not, in fact, been constructed from a proc_t struct.
* While inspect data is output at the character level,
. up to now all task display data was only potentially
. output and it was always based on a complete string.
* With task search highlighting, rows now containing a
. match must be output in pieces and, therefore, can't
. be optimized away like other rows which haven't been
. been altered. This is because top cannot predict the
. the contents of a search string or, how many matches
. might occur in a given row. Short search strings and
. many matches would raise buffer needs geometrically.
(now that we know a '.' + 2 spaces is squeezed to one)
(everything's perfectly justified, but it's just luck)
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit extends Inspect provisions for 'find/next'
to each individual selection. Thus a user can maintain
multiple active searches without having to reissue the
locate command whenever the current selection changes.
To emphasize this feature the View screen now displays
the current active locate string or 'N/A' if inactive.
Such a reminder is important when no found matches are
present on the 1st display page, given that they would
otherwise be apparent via the additional highlighting.
(now that we know a '.' + 2 spaces is squeezed to one)
(everything's perfectly justified, but it's just luck)
Signed-off-by: Jim Warner <james.warner@comcast.net>
We have modeled the Inspect search provisions on those
provided by the 'less' pager. With this commit we take
the next step and provide for highlighting any strings
matched (and in view). Of course, top will continue to
adjust the beginning column so as to bring out-of-view
matches into view, while highlighting visible matches.
However, top won't emulate every 'less' behavior since
the following are seen as flaws in the user interface.
* when viewing true binary data, less makes no attempt
. to smooth the right margin by truncating unprintable
. symbols, thus creatng ragged unappealing right edges
* when viewing true binary data, less will always fail
. search requests regardless of surrounding characters
* less refuses to bring out-of-view found matches into
. view by adjusting the left-most column, if necessary
(now that we know a '.' + 2 spaces is squeezed to one)
(everything's perfectly justified, but it's just luck)
Signed-off-by: Jim Warner <james.warner@comcast.net>
Previously top would warn users if an older version of
an rcfile was about to be overwritten. That's assuming
that RCFILE_NOERR was not defined. This left, however,
other potential rcfile issues or questions unattended.
For example, if a faulty 'inspect' redirected echo had
overwritten all window entries or if the inspect entry
was not 'pipe' or 'file' (actually, just a 'p' or 'f')
then top would silently accept it but look no further.
With this commit top will try to process every inspect
entry, while preserving unrecognized entries. Plus all
other non-fatal rcfile errors will now alert a user to
the potential overwrite when the 'W' command is given.
(now that we know a '.' + 2 spaces is squeezed to one)
(everything's perfectly justified, but it's just luck)
Signed-off-by: Jim Warner <james.warner@comcast.net>
The Inspect find algorithm has always been challenging
given the possibility that 'rows' might contain binary
data. Be that as it may, two small changes have proven
to dramatically improve the performance of such scans.
The first involves the case wherein if no match on the
'substring' portion of a row was found, then a pointer
representing the substring was increased by the length
of the search string, not the better/longer substring.
Thus, portions of the substring were always rescanned!
The second performance boost was achieved in this way:
pre-scanning each raw row for just the first character
in the search string now determines if a full match is
even possible. Therefore, repeated unproductive strstr
calls on individual substrings within that row will be
avoided. In a nutshell, 1 'if' with '}' did the trick!
(now that we know a '.' + 2 spaces is squeezed to one)
(everything's perfectly justified, but it's just luck)
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit improves display performance when the user
has scrolled horizontally past the end of a top 'row'.
We can avoid the need to memset our buffer with spaces
and putp those spaces individually by exploiting logic
that already exists. If one '\n' character is inserted
into the buffer instead, the next terminfo string sent
will be Cap_clr_eol achieving exactly the same effect!
(now that we know a '.' + 2 spaces is squeezed to one)
(everything's perfectly justified, but it's just luck)
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit introduces an extremely powerful, flexible
brand new capability. Now, users can pause the normal
iterative display and inspect the contents of any file
or output from any script, command, or even pipelines.
It's invoked via the 'Y' interactive command which, in
turn, is supported with simple user supplied additions
as new entries in the top personal configuration file.
A separate new 'Inspect' window supports scrolling and
searching, similar to the main top display. Except it
extends existing 'L'/'&' (locate/locate-next) commands
so that an out-of-view match automatically adjusts the
horizontal position bringing such data into view. And
it provides for multiple successive same line matches.
Also, the basic 'more/less' navigation keys are active
in this new 'Inspect' window, to ease user transition.
There are no program changes required when entries are
added to or deleted from the rcfile. And there are no
known limits to the complexity of a script, command or
pipeline, other than the unidirectional nature imposed
by the 'popen' function call which top cannot violate.
Since it's impossible to predict exactly what contents
will be generated, top treats all output as raw binary
data. Any control characters display in '^C' notation
while all other unprintable characters show as '<AB>'.
The biggest problem encountered was with the find/next
capability since that strstr guy was really diminished
given the possibility that numerous 'strings' could be
encountered *within* many of top's raw, binary 'rows'.
Oh, and another problem was in maintaining the perfect
left & right text justification of this commit message
along with all of the commit summaries. Some of those
summaries (like this very one) are of course, slightly
shorter, to make room for the 'man document' addition.
Enjoy!
Signed-off-by: Jim Warner <james.warner@comcast.net>
As an aid to the above 2 commands, and as a prelude to
an upcoming 'inspect other output' capability, the act
of selecting a process for either has been simplified.
Positioning a task as the first one displayed, via the
up/down arrow keys, will now establish it as a default
selection for the appropriate command. Thus, that pid
will then be incorporated in a subsequent input prompt.
Signed-off-by: Jim Warner <james.warner@comcast.net>
An earlier commit improved the scroll coordinates
message performance by offloading most of the work
to those occasions when column headers were rebuilt.
The only remaining per-frame costs were the addition
of some terminfo escapes and the Frame_maxtask count.
This commit further reduces those per-frame costs to
the absolute minimum.
Reference:
commit fbfaa868ba
Signed-off-by: Jim Warner <james.warner@comcast.net>
The 'refactor and enhance column width management'
recent redesign produced many subsequent benefits,
the latest of which is automatically sized fixed-width
non-scalable columns.
As expected, there was a cost associated with these
many enhancements. That cost has now been identified
as a 1-4% performance degradation, depending on which
fields are being displayed.
This increased cost arises principally from current
drawing related function calls, whereas top-3.3.3 did
most of its drawing via macros effectively inlining
those duties.
This commit inlines the equivalent drawing functions,
thus eliminating the function call penalty, and places
this top on a par with top-3.3.3. The trade off is a
modest additional 4k in executable size.
Signed-off-by: Jim Warner <james.warner@comcast.net>
The recent introduction of a column widths override
(the 'X' command) provided for a user input amount
to be added to default field size which ranged from
5 to 10 bytes.
While that approach could prevent truncated data, the
different default sizes would almost certainly mean
some precious screen real estate was waisted.
This commit introduces the concept of dynamic widths
where top will add only enough to a field default to
prevent truncation for that specific field.
Now users have a choice between their explicit width
override or a width chosen by top to exactly match
display needs. The former is immediate but likely
wastes some horizontal space while the latter is
iterative but will be sized precisely.
Original 'X' Command:
commit 384afa494a
commit 47e1d063ac
Extensions to 'X' Command:
commit bbf8e44fb4
commit 7557f3f754
Signed-off-by: Jim Warner <james.warner@comcast.net>
There were some gaps in the alternate navigation keys
top provided. Additionally, some inconsistencies
existed in the supporting key table.
This commit adds the following new key equivalents,
mirroring the standard vim navigation keys:
. ctrl+alt+ k = pgup, ctrl+alt+ j = pgdown
. ctrl+alt+ h = home, ctrl+alt+ l = end
Also, the supporting table entries now consistently
follow these "directions":
. up/pgup, down/pgdown, left/home, right/end
Signed-off-by: Jim Warner <james.warner@comcast.net>
To support the cursor navigation keys, after saving
the termios structure top issues 'smkx/keypad_xmit'
during startup. However, some terminals appear to
treat that directive as persistent which leaves a
corrupted tty state after top exit.
This commit reverses the above terminal directive
via 'rmkx/keypad_local' just prior to restoring the
saved termios structure at program end.
For discovering this bug, and providing the 'rmkx'
clue to its solution, thanks to:
Kirill A. Shutemov <kirill@shutemov.name>
Signed-off-by: Jim Warner <james.warner@comcast.net>
This 'Sleeping in function' field was made variable
width because the length of current kernel symbols
usually exceeded the former top's 9 character limit.
As a variable width field it would steal valuable
horizontal display positions from other, more likely,
displayed fields such as COMMAND or CGROUPS.
With the advent of the new 'X' toggle, no fixed-width
non-scalable field need suffer permanent truncation.
Thus, WCHAN is being made fixed width with a default
size of 10 characters.
Signed-off-by: Jim Warner <james.warner@comcast.net>
top/top.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
I have no idea what the maximum length of a terminal
name might be. However, the library provides for up
to 128 characters (ouch).
So just to be safe, this commit extends the ability
to widen columns to embrace this field.
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit accommodates those fields which may have
suffered truncation due to these default limits:
. 5 digits for uid/gid type fields
. 8 characters for user/group type fields
With a new interactive command, users can increase the
width of all such fields, or return to the defaults.
Note:
There are no restrictions on the amount added to
the defaults. The user is free to vastly exceed
screen limits which simply means such fields can
never be displayed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit affords user control over justification
for both column headings and the subordinate data.
Separate toggles are provided for control of numeric
data and string data.
Signed-off-by: Jim Warner <james.warner@comcast.net>
Now that column headings are independent of column
data format and require no carefully managed padding
bytes they are candidates for nls translation.
This commit migrates all column headings to the .pot
file with additional translator guidance in the form
of maximum sizes to avoid truncation.
It also places these new additions adjacent to their
associated descriptions, which were already present.
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit accomplishes the following objectives:
* remove extra task_show parm added with 'Locate'
* avoid column overflow with subsequent misalignment
* eliminate spaces for column heading padding
* decouple column headings from column data formats
* eliminate all hardcoded column format specifiers
* generalize the inter-column spacing management
* remove Fieldstab.desc in favor of direct nls access
* set the stage for nls support of column headings
* set the stage for dynamic changes to justification
Signed-off-by: Jim Warner <james.warner@comcast.net>
This flaw was revealed under 'man2htm' and dates back
to the first Gitorious revised top submission.
Reference:
commit fd62123562
Date: Thu Mar 31 22:15:12 2011 +1100
Signed-off-by: Jim Warner <james.warner@comcast.net>
The TREE_RESCANS #define (formerly TREE_ONEPASS) has
been eliminated and the approach to forest view mode
redesigned. The chance of dangling children has been
eliminated and overhead reduced.
We now order processes on start_time (non-display)
and are therefore immune to any pid, ppid or tgid
anomalies when pid values wrap.
The new algorithm also accommodates any distortions
caused by the 3.3 kernel 'hidepid' provisions --
something guaranteed to produce dangling children
under the former approach.
Related References:
commit a2086dfdf6
commit cd608f462e
commit 41ed28aa5d
Signed-off-by: Jim Warner <james.warner@comcast.net>
The recent introduction of scrollable variable width
columns makes a process 'environment' a potentially
useful addition to top's displayable fields.
This commit exploits the following new library flag:
PROC_EDITENVRCVT
Signed-off-by: Jim Warner <james.warner@comcast.net>
In an effort to avoid dangling children when in forest
view mode, top defaulted to a complete rescan of every
proc_t for each child encountered.
That expense was never really cost justified and now
with the 3.3 kernel 'hidepid' provisions it no longer
can offer such protection.
With this commit, the TREE_ONEPASS define is changed
to TREE_RESCANS so as to reverse the default scan
behavior.
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit represents mostly spelling corrections
in comments. It also includes a few very minor logic
changes/relocations.
Signed-off-by: Jim Warner <james.warner@comcast.net>
This section purported to list fields in alphabetical
order, but this was not always true.
With this commit, strict ascii collating sequence is
now observed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
With the introduction of intra-column scrolling, the
scroll coordinates message was enhanced to give some
hint of positioning within a scrolled column.
Rather than rebuild this somewhat costly string from
scratch with each frame, we'll now do the bulk of the
work only when column headers are constructed.
The only remaining per frame costs will then be the
addition of a few terminfo escapes and the current
Frame_maxtask count.
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit introduces horizontal scrolling within any
variable width column. Thus, an entire command line,
complete list of control groups, etc. can now be
viewed -- not just a screen width's portion.
It is activated when any variable width column:
. is (via field selection) or
. has become (via the right arrow key)
the only displayed field.
Then, the right and left arrow keys can be used in the
normal way to continue scrolling within that column.
The amount scrolled with each key press is currently
set as the normal tab stop increment of 8 characters.
Signed-off-by: Jim Warner <james.warner@comcast.net>
The library does not weed out potential duplicate PID
values when sampling with the PROC_PID flag. This was
treated as merely an inefficiency by top and safely
ignored prior to the advent of forest view mode.
Now, however, if the -p switch duplicates certain PIDs,
*and* those processes have no PPID, *and* top's forest
view mode is active or activated, then a SEGV will be
generated (and caught).
This rather obscure buglet is thus limited to pid #1
(/sbin/init) and pid #2 (kthreadd). With any other
duplicate PIDs the worse case scenario was a '?' in
place of the usual forest view artwork.
This commit silently ignores any duplicate process ids
and thereby avoids the potential segmemtation fault.
Signed-off-by: Jim Warner <james.warner@comcast.net>
All warnings where about unnecessary quoting. The scriptlet
below will tell what was wrong.
for I in ./top/top.1 ./ps/ps.1 ./*.[0-9]; do
echo "== $I warnings =="
man --warnings=all $I > /dev/null
done
This should probably be turned to 'make check' script.
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
Since its inception top has provided for monitoring
only specific process ids via the -p command line
switch. This provision has also embraced the top
process itself, even though its pid wasn't yet known.
This commit simply documents the special zero value,
which would otherwise be an invalid process id.
Signed-off-by: Jim Warner <james.warner@comcast.net>
In attempting to keep at least one task visible when
scrolling vertically, a negative task index would be
produced when pid monitoring was in effect and no
matching pid was found.
Since there were already other conditions where no
task might displayed, the faulty source line has been
removed.
Bug-Debian: http://bugs.debian.org/668335
Signed-off-by: Jim Warner <james.warner@comcast.net>
If stream status is not checked at the end of execution below problem
would not report error, or non-zero exit code. The uptime is just an
example same was true with all commands of the project.
$ uptime >&- ; echo $?
uptime: write error: Bad file descriptor
1
$ uptime >/dev/full ; echo $?
uptime: write error: No space left on device
1
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
When top introduced true line input editing, the
ability to paste keystrokes was lost. This remains
a necessary evil so that top has an opportunity to
translate cursor motion keystrokes into terminfo
escapes during line input. Motion keys themselves,
of course, can never be pasted.
If pasting ever became more important than input
editing, then native termios support should have been
available via a define called TERMIOS_ONLY. But a
recent commit, eliminating what was thought to be
obsolete logic, rendered the alternate linein()
function virtually useless.
Similar to top-3.2.8, when native termios input is
functional, these abberations can be experienced:
. cursor motion keys will appear as escapes
. excessive input can cause line wraps
. ^Z during i/p is not be honored until <Enter>
. SIGWINCH during i/p corrupts screen temporarily
In hindsight, it now seems that the ability to paste
keystrokes may indeed outweigh any shortcomings of
native termios support. This is especially true if
one is preparing to search ('L') for some lengthy
process command line contined in the clipboard.
Thus, this patch fixes the alternate linein() function
and changes TERMIOS_ONLY to TERMIO_PROXY so that top
now defaults to using native termios input. In turn,
that will restore the paste keystrokes capability.
Reference(s):
commit: 045538e01b
Reported by: sergio <mailbox@sergio.spb.ru>
Bug-Debian: http://bugs.debian.org/663334
Signed-off-by: Jim Warner <james.warner@comcast.net>
Admittedly, top referred to memory quantities in
a variety of non-standard ways. This commit brings
the program and supporting documentation into strict
compliance with IEC standard binary names.
According to wikipedia, as of 2012 this IEC standard
was still not in widespread use. However, I'm making
this change now for the anal-retentive among us.
Reference:
http://en.wikipedia.org/wiki/Megabyte
Reported by: Roman Mamedov <rm@romanrm.ru>
Bug-Debian: http://bugs.debian.org/662786
Signed-off-by: Jim Warner <james.warner@comcast.net>
These per instance warnings have been eliminated:
warning: range expressions in switch statements are non-standard
warning: padding struct to align 'winflags'
warning: dereferencing type-punned pointer will break strict-aliasing rules
These per instance warnings have not been addressed
since they simply trade one warning for another:
From:
warning: ISO C does not permit named variadic macros
warning: ISO C does not support the '%Lu' gnu_scanf format
To:
warning: anonymous variadic macros were introduced in C99
warning: ISO C90 does not support the 'll' gnu_scanf length modifier
Lastly, since all C compilers have supported use of
C++ style comments for the past 20 years, the top
program will never trade them for the often more
cumbersome C style comments simply to avoid this
once per source file warning:
warning: C++ style comments are not allowed in ISO C90
Signed-off-by: Jim Warner <james.warner@comcast.net>
Under openSUSE, old top uses additional fields for
out-of-memory reporting. As a result, under the
original approach to rcfile conversion, new top would
issue a fatal corrupt window entry message asking that
the rcfile be deleted.
This patch extends the conversion range to include
the extra openSUSE field characters. It's effective
when ./configure specifies the --enable-oomem option
which in turn defines OOMEM_ENABLE.
This commit also makes the conversion logic slightly
more forgiving. While enforcing an upper limit on the
expected number of old style field characters, amounts
less than that will be handled seemlessly.
Reference:
commit 4b98733132
Signed-off-by: Jim Warner <james.warner@comcast.net>
At one time, new top silently defaulted when an rcfile
was found to be incompatible. This is exactly what
the old top did. However, after some discussion it
was decided top should alert the user and thereby
save the system administrator some headaches.
Now, some are upset over the fatal error, proving you
can't please everybody. But in all fairness, given
the difficulty of customizing old top, any reluctance
to delete an old saved rcfile is understandable.
To ease transition to this new top, old style rcfiles
will now be honored and converted to the new format.
And if not disabled at ./configure time via CFLAGS,
a user will be warned when an old style rcfile is
about to be overwritten using the 'W' command.
Lastly, the config validation logic was enhanced to
help ensure both types of rcfile haven't been edited
manually and possibly made unuseable.
Reported-By: sergio <mailbox@sergio.spb.ru>
Bug-Debian: http://bugs.debian.org/651213
Reported-By: martin f krafft <madduck@debian.org>
Bug-Debian: http://bugs.debian.org/651863
Signed-off-by: Jim Warner <james.warner@comcast.net>
When sort column header emphasis was extended to a
monochrome screen, the ability to emphasize selections
on the Fields Management screen was lost when colors
were not being displayed.
This patch corrects that bug by using the capclr_hdr
terminfo string instead of capclr_msg.
Reference:
commit 0c6aa6af41
Signed-off-by: Jim Warner <james.warner@comcast.net>
Until the 'locate/search' provisions were added,
top avoided the need for any function prototypes
through careful source file organization. But
the addition of the find_string function required
a prototpe for task_show, lest a massive file
reorganization be undertaken.
This commit moves the actual protype out of top.h
and places it adjacent to the caller in order to
avoid a warning when top_nls.c is compiled.
References:
commit 270e8e7eeb
commit d6e6a9aa38
Signed-off-by: Jim Warner <james.warner@comcast.net>
All top.h defines were lumped together as:
'Development/Debugging defines'
This commit establishes this new category:
'Defines represented in configure.ac'
And that new category now contains OOMEM_ENABLE,
which enables the SuSE out-of-memory additions
Signed-off-by: Jim Warner <james.warner@comcast.net>
At one time the 'open_psdb_message' library call was
supported with the concept of a postponed message
which would display after top startup completed.
In turn, that required logic to strip the '\n' which
was embedded (inappropriately) in any such message.
Nowdays top treats such a returned error as fatal so
there is no need for the 'strim' function which is
being removed with this commit.
Signed-off-by: Jim Warner <james.warner@comcast.net>
We originally approached the potential problem of
% CPU distortions as unique to Nehalem type cpus.
The latest information suggests that it may have
been due to a kernel anomaly that has since been
corrected.
Yet even without such a cpu, wide disparities in
tics allocation among all available cpus have
sometimes been observed -- spikes as it were in
the normal pattern. This has happened under both
version 2.26.38-13 and 3.0.0-15 kernels.
The small amount of additional code addressing the
original problem carries very little overhead. It
is being retained to afford protection against any
future tic accounting aberrations.
In this commit, supporting programmer comments have
been divorced from any particular cpu type. Also,
another variable and manifest constant will now be
eliminated when CPU_ZEROTICS is defined.
References:
commit 02508b3d76http://www.freelists.org/post/procps/CStates-handling-new-switch,50
When the calibrate_fields function was broken up for
mainainability, an obscure regression was introduced.
For the resulting bug to affect the display, all of
the following conditions would have to be met:
. USE_X_COLHDR was not defined
. column highlighting had been turned on
. many, perhaps all, fields were displayable
. the user then typed the <End> key
. and the current sort column just happened to
be immediately to the left of the left-most
visible field
This patch corrects for that remote possibility.
Reference:
commit d0e16acf15
Signed-off-by: Jim Warner <james.warner@comcast.net>
This patch adds the -p option to the uptime(1) command, which changes
the uptime displayed from something like:
10:35:52 up 2:33, 1 user, load average: 1.69, 1.65, 1.63
to:
up 2 hours, 33 minutes
I originally implemented this as the up(1) program about 14 years ago.
In 2008 or 2009, I created a patch for procps to add this functionality
to uptime and submitted it to the project. Never heard from the
project and no new releases of procps had been made. Then I found out
about this project and decided to port my patch to it. So here it is.
This is really just for fun. There is no real technical reason to
have this functionality. But even now, 14 years later, I still get
emails asking where the source code for up is. So I thought it would
be nice for the uptime command on Linux to sport the up functionality
by default.
The calibrate_fields function had grown too large and
was adversly impacting maintainence. So half of the
logic was split out into a new function.
Now, maintainence of column headers and the required
library flags is organized as follows:
. adj_geometry (calibrate_fields helper)
provides low-level support for sigwinch, memory
. build_headers (calibrate_fields helper)
constructs the headers and library flags
. calibrate_fields
establishes which fields will be displayed
( note the alpha order mentioned in a prior commit )
Signed-off-by: Jim Warner <james.warner@comcast.net>
The top program is carefully organized into sections
and those sections are carefully placed so as to avoid
the need for prototypes. *
Additionally, names of functions are carefully chosen
to maintain alphabetical order within each section.
The names of most 'helper' functions, which are always
placed immediately above the calling functions, often
only met the spirit of the alphabetical law, not the
actual letter of that law.
This commit alters the names of such helper functions
so as to mainatin strict ascii alphabetical order
within each section.
* the single exception to prototypes is find_string,
which calls the task_show function, and would have
prompted a massive reorganization.
Signed-off-by: Jim Warner <james.warner@comcast.net>
The recently added logic dealing with "missing" tics
is mutually exclusive with logic associated with a
define called CPU_ZEROTICS.
This commit expands the use of that define to exclude
such Nehalem logic as appropriate.
It also extends programmer notes in top.h to include
an attribution for initiating the topic of potential
Nehalem type % CPU distortions, acknowledging:
Jaromir Capik, <jcapik@redhat.com>
References
commit ce1410a51a
commit 9e7dd43ab7
commit a9041a5526
The mem and swap lines have enough room to show eight significant
digits, so switch to showing MB when >=100MB, not >=10MB.
The extra detail is valuable; it should not be elided.
Signed-off-by: James Cloos <cloos@jhcloos.com>
Signed-off-by: Jim Warner <james.warner@comcast.net>
Due to a poorly constructed temporary fprintf
used during development, an earlier commit went
a little too far in its computations. The net
result was the code looked nice but actually
accomplished nothing.
It is the /proc/stat line 1 (summary line)
whose tics must be used in establishing the
threshold boundary. And that calculation
need be performed just once per frame.
This commit ensures one threshold calculation
per delay interval no matter how many cpus
are ultimately displayed.
It also corrects scalability by factoring in
the total number of online processors.
Reference:
commit 9e7dd43ab7
Along the way to width override support (-w switch),
this top began clearing the screen far more often
than his predecessor. In fact, it happend with each
user keystroke.
This commit dramatically reduces those occurances.
The screen will now be cleared only when an actual
SIGWINCH is received.
Thanks for identifying this flaw belongs to:
James Cloos, cloos@jhcloos.com
References:
http://www.freelists.org/post/procps/Merge-request
Blame: c2dcbef482
Author: Jim Warner <james.warner@comcast.net>
Date: Thu May 26 11:33:32 2011 +0200
subject: added output width/height override support to top, + misc
The original approach to potential % CPU distortion due
to Nehalem type cores being turned off completely when
idle worked ok until the user typed something.
At that point, elapsed tics would no longer equal the
calculated value producing an undesirable 100% idle
condition until the next update or <Enter/Space> key.
This commit employs actual elapsed tics in determining
whether a cpu should be considered idle and thus makes
top's individual cpu display immune to user keystrokes.
This patch provides for cpu cores which can be turned
off completely when idle (Nehalem, etc.) thus registering
very few or no tics since the last update cycle.
When CPU_ZEROTICS is not defined (the default), any
displayed cpu with less than a certain amount of total
tics will show as 100% idle. That amount is tempered
by the delay interval and total number of cpus.
This commit also satisfies the Debian 'top_nohz' patch
(11/24/09) in a slightly more efficient manner. That
patch concerned kernels built with CONFIG_NO_HZ.
Reference:
http://www.freelists.org/post/procps/CStates-handling-new-switch,4
Prior to this patch, top was able to handle any hotplugged
cpus *added* to the system in two distinct ways.
1) Newly added cpus would be detected by sysinfo_refresh
calling the library's cpuinfo function, which occurs
at most every 5 minutes.
2) The user could force a refresh using either the
<Enter> or <Space> keys.
Unfortunately, the *loss* of a cpu would produce an early
exit due to a /proc/stat read failure. Such a failure
can be produced in the following way:
sudo echo 0 > /sys/devices/system/cpu/cpu??/online
This commit allows top to tolerate the loss of cpus.
It also provides for more efficient CPU_t management,
especially for massively parallel cpu environments.
Note: Changes to the cpu compliment can produce a single
cycle distortion of cpu percentages. Such distortion is
most visible when each cpu is being displayed. It can
be eliminated with a forced refresh via <Enter>/<Space>.
With the introduction of the 'locate string' provisions,
the precedent for calling tertiary helper functions from
secondary do_key helper functions was established.
This commit simply migrates some additional keys out of
the do_key function itself and into the more generalized
key table.
Normally, when the chosen sort column is displayed via the
'x' command toggle the entire column is highlighted. And
while this version of top substantially reduced the cost
of such highlighting, a small pathlength increase remained.
The USE_X_COLHDR define was an experimental alternative which
eliminated all recurring runtime costs for such emphasis by
highlighting the column header, not the entire column.
The previous implementation required colors to be turned on
(the 'z' toggle) for such highlighting to be visible. This
commit extends column header emphasis to include monochrome
displays as well.
Reference:
http://www.freelists.org/post/procps/post-nls-merge,6
Since its inception, this top has improperly handled an
empty HOME environment variable. Under those conditions
a path to the root directory would have been constructed.
That caused no real harm upon startup since the display
defaults would have been employed. However, except for
root, it would have been impossible to save the rc file.
This commit keeps the promise made in the documentation.
This commit addresses a long standing buglet (debian #441166) which
surfaces when the display mode is switched between task and threads.
An extra procps refresh is now forced upon such a transition which
parallels the approach used at startup for the exact same reason.
Reference: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=441166
Some of the latest changes to Makefile.am files are missing.
This patch restores the LOCALEDIR variable, among others,
and dispenses with the include directives in the ps/ and top/
subdirectories since they're no longer needed.
This commit corrects some outdated programmer comments.
Additionally, certain nls justifications might become
increasingly obscure with the passage of time so some
previous nls commit text has been added as comments.
This patch represents the final resting place for miscellaneous
changes not otherwise encountered or for which the resolution
was incomplete or incorrect.
The gettext documentation leads one to believe that
a printf type call is necessary for gettext string
extraction. That turns out to be misleading and
all one really needs is the runtime gettext string
address resolution.
Thus, we can avoid our original snprintf/strdup
overhead and establish an address for the original
or translated string just by issuing the _() macro.
We create these nls string tables so that:
1) top avoids the overhead of repeated
runtime function calls
2) we can control the order of top's
strings in the .pot file
3) translator comments don't obscure
and clutter the main program
Until this patch, top had used some strings with
special escape sequences to produce colors, normal
text, bold text, etc. They took the following form,
explained by an excerpt from program comments:
...
Our special formatting consists of:
"some text <_delimiter_> some more text <_delimiter_>...\n"
Where <_delimiter_> is a single byte in the range of:
\001 through \010 (in decimalizee, 1 - 8)
and is used to select an 'attribute' from a capabilities table
which is then applied to the *preceding* substring.
...
Unfortunately, these nonprinting values revealed
insurmountable inconsistencies in both the front-end
and back-end translation tools.
The xgettext (extraction) program would take those
special escapes, convert them and then output raw
binary values. Thus the .pot file would contain
lots of unprintable stuff making it unreadable.
If the following was added to po/Makevars, most of
those special escapes would be preserved in their
escape notation:
XGETTEXT_OPTIONS = ... --escape
But two escapes were converted from octal notation
and there was no way to prevent it:
\007 --> \a
\010 --> \b
After a pass through the msginit program, most of
the escapes were reconverted to raw binary values
making translation impossible. There was no
"--escape" option for the back-end programs like
there was for xgettext.
But the real killer was the escape \004, also used
in some of top's special strings. This value would
be silently accepted by xgettext, only to produce
the following fatal error in back-end programs like
msginit, msgfmt and msgen:
.pot:2647: context separator <EOT> within string
To quote from one of the references below:
"Would you create a suite of tools that silently
allow what is destined to become a fatal error
to pass unnoticed?"
So the bottom line was: top's special strings, in
use for the past nine years, had to be redesigned.
References:
http://www.freelists.org/post/procps/procpsng-nls-support,11http://www.freelists.org/post/procps/procpsng-nls-support,14
