When we transitioned from 'char' based fields to 'int'
proper initial cursor positioning (highlight) was lost
in the 'fields_utility' function. The highlight should
appear on the current sort field. Instead, however, it
was often positioned on the first field listed or some
completely different field other than that sort field.
So this patch will restore the master branch behavior.
[ and on the assumption that a sort field is visible ]
[ we will search first for a field with the 'on' bit ]
Reference(s):
. Mar, 2022 - trade 'char' for 'int'
commit 46aa96e438
Signed-off-by: Jim Warner <james.warner@comcast.net>
At this point I don't know if the Rc.id ever carried a
value of 'b', 'c', 'd' or 'e' in a public release. But
I do know that those values ought not to be recognized
in any configuration file. Now they won't be for sure.
[ the net result is that now a user with such a file ]
[ will see the 'incompatible rcfile' message instead ]
[ of that misleading 'window entry corrupt' message. ]
Signed-off-by: Jim Warner <james.warner@comcast.net>
When the configuration file is first written, there is
just the normal single newline which ends that rcfile.
However, when any existing rcfile was rewritten, there
was one extra newline '\n' character added at the end.
This will happen just once. The file does NOT continue
to grow with the extra blank lines always being added.
Anyway, this patch will put an end to such a practice.
[ along the way i fixed an unrelated whitespace oops ]
Signed-off-by: Jim Warner <james.warner@comcast.net>
This new integer based version of 'fieldscur' deviated
a bit too much from that character based logic when it
comes to ancient 3.2.8 rcfile support. This results in
an inappropriate extra field shown as the last column.
If we're truly dealing with 20 year old 3.2.8 rcfiles,
it's important the 3.3.9 extension (RCF_PLUS_H) not be
concatenated to what was read. That's because a search
for the special SUSE characters will always find a '\'
in the 26th position (after normal 3.2.8 'fieldscur').
[ for symmetry, we'll also avoid RCF_PLUS_J (3.3.17) ]
All other 'old' rcfiles are transformed without error.
Reference(s):
. Mar, 2022 - 3.2.8 support restores
Signed-off-by: Jim Warner <james.warner@comcast.net>
This patch will restore the previously abandoned older
rcfile support. To be honest, I wasn't sure this would
be possible given the vast difference between formats.
But after some very intense hacking our newlib top can
again fully support seamless migration from some older
'char' based file to this new 'int' based file format.
[ top even supports a 20+ year old 3.2.8 config file ]
Signed-off-by: Jim Warner <james.warner@comcast.net>
With each of those references shown below, the maximum
number of fields was increased. However, with a 'char'
based implementation we're nearing the upper limits of
total displayable fields. We currently use 76 of a max
of 86 fields. With extra effort, 94 might be possible.
But 94 is the absolute upper limit of possible fields!
Moreover, the current implementation yields characters
that were unprintable in the rcfile. This could become
an issue with that 'inspect' feature when/if an rcfile
is edited to add entries (as opposed to using 'echo').
So, with this commit the internals of field management
has been completely reimagined. It is now based on the
integer type, not a character. And whereas that former
design used the high order bit to show the 'on' state,
thus yielding an unprintable character, the new design
uses the low order bit for the state. As such, numbers
will be kept small and an even number will be an 'off'
field whereas an odd number will become an 'on' state.
The bottom line is that this new design will afford an
unlimited number of new fields while keeping an rcfile
completely free of that potential unprintable garbage.
And it is embarrassingly easy to extend the maximum of
supportable fields from the currently implemented 100.
Who knows, maybe a future patch will prove this point.
[ unless a subsequent commit proves otherwise, given ]
[ the dramatic differences in rcfile contents, i had ]
[ to abandon the practice of supporting old rcfiles. ]
Reference(s):
. Nov, 2013 - RCF_PLUS_H introduced
commit af4e6533ba
. Jul, 2016 - RCF_PLUS_J introduced
commit d5c5051fb3
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit just compliments a change referenced below
by providing the value which includes reaped children.
Reference(s):
. Feb, 2022 - added %CUU field
commit 7647e96b0a
Signed-off-by: Jim Warner <james.warner@comcast.net>
When scale_tics was refactored and then Ctrl-E support
added to top, the complete range of scaling values was
not visible. Namely, a single 'd' (days) & 'w' (weeks)
was never seen with ^E. With this commit they will be.
Reference(s):
. Mar, 2022 - introduced ^E tics scaling
commit 402bf1903b
. Mar, 2022 - refactored scale_tics
commit 71eb90c1b2
Signed-off-by: Jim Warner <james.warner@comcast.net>
That normalization of the 'scale_tics' function in the
prior commit convinced me that I won't please everyone
with my arbitrary choices for the scaling transitions.
So, this patch will provide the users with a means for
setting their own scaling transition points with a new
toggle. Ctrl-E was chosen since the 'e/E' toggles were
already present as a means of scaling (albeit memory).
[ this toggle will also serve an educational purpose ]
[ by allowing one to see all the scaling conventions ]
The scaling a user establishes is saved in the rcfile.
Signed-off-by: Jim Warner <james.warner@comcast.net>
This patch refactors the 'scale_tics' function to more
closely parallel uptime shown on the first line of the
summary area. The old logic has been preserved through
the header file's new #define SCALE_FORMER provision.
However, the former logic was actually a big disaster.
These are some potential problems with that old logic:
1. With respect to our time fields top no longer deals
solely with cpu time. So, the old limits of '68 weeks'
could possibly be insufficient to reflect those times.
2. Given the widths of top's new time fields, the code
never got beyond scaling to hours. For example, with a
ridiculously large span of 19 years, the scaled result
would then be shown as '167832h'. We never reached the
days ('6993d') or even the weeks ('999w') equivalents.
3. Similarly, with that 'TIME+' field and a large tics
value, results would then appear as 'MMMMMM:SS' rather
than the more meaningful 'HH:MM:SS' or days and hours.
So henceforth we will adopt these scaling conventions:
MMM:SS.hh ... minutes:seconds.hundredths
MMM:SS ...... minutes:seconds
HH,MM ....... hours,minutes
D+H ......... days+hours (with 'd' & 'h' suffixes)
D ........... days (with 'd' suffix)
W+D ......... weeks+days (with 'w' & 'd' suffixes)
W ........... weeks (with 'w' suffix)
Note that, unlike our former scaling logic, that 'MMM'
portion won't be allowed to grow unconditionally. It's
limited (arbitrarily?) to 360 total minutes (6 hours).
Additionally, the 'HH' guy will be limited to 96 hours
(4 days) while that 'D' limit was set at 14 (2 weeks).
Whenever a limit is hit, scaling will advance a level.
Reference(s):
. Feb, 2022 - added 'ELAPSED'
commit 9348d3b008
. Feb, 2022 - added 'STARTED'
commit 7647e96b0a
Signed-off-by: Jim Warner <james.warner@comcast.net>
When the 'STARTED' field was added, in the message for
the commit referenced below, I explained why 'ELAPSED'
shouldn't be implemented though it might be preferred.
Well, after climbing out of my box to do a little more
thinking, I came up with the way to add that 'ELAPSED'
field while avoiding the possible performance penalty.
Just do not show what would change with every refresh!
If we do not show the seconds portion of a scaled tics
amount then the problem goes away. And this comes with
an additional benefit. The HH,MM (hours,minutes) style
then is readily compared with that system uptime shown
as HH:MM. The only difference is just the comma/colon.
[ assuming the top uptime/load average toggle was on ]
Reference(s):
. introduced 'start time' field
commit 7647e96b0a
Signed-off-by: Jim Warner <james.warner@comcast.net>
When the 'STARTED' field was added, the width was made
the same as the 'TIME+' field. Thus, a full time could
be shown (which then included hundredths of a second).
That kind of granularity is totally unnecessary. After
all, this column is potentially confusing enough since
it is so counterintuitive. So, this commit will reduce
the width of the field with some help from scale_tics.
Henceforth it will not include those ol' centiseconds.
[ along the way let's expand the man document with a ]
[ a remainder about content representation & scaling ]
Reference(s):
. introduced 'start time' field
commit 7647e96b0a
Signed-off-by: Jim Warner <james.warner@comcast.net>
This patch will exploit some new library capabilities.
[ one will raise eyebrows, the other likely will not ]
A new 'STARTED' field was added which shows the time a
process started after system boot. As such the largest
interval represents the most recently started process.
This is the field that will likely be questioned since
it's somewhat counterintuitive. But were we to instead
use TIME_ELAPSED, the value will change with every top
refresh. This will defeat any PUFF macro optimization.
The new '%CUU' field will probably be better received.
It represents the cpu usage over the life of the task.
When a process was showing high %CPU usage, this field
can be used to determine if it's an anomaly or normal.
[ and as with %CPU, %CUU shows a '?' when running in ]
[ a namespace when /proc was mounted with subset=pid ]
Signed-off-by: Jim Warner <james.warner@comcast.net>
The assignment was originally created to reduce length
of one line of code (i.e. for readability). But we can
achieve this objective without adding executable code.
Signed-off-by: Jim Warner <james.warner@comcast.net>
There was a potential problem with these macros should
that '#define SCROLLVAR_NO' be active and they were to
appear in an 'if' statement (like is necessary for the
master branch version of top under that EU_CMD label).
[ now they're always usable without requiring an ';' ]
Signed-off-by: Jim Warner <james.warner@comcast.net>
This patch eliminates those warnings referenced below.
They are sometimes associated with an obscure #define.
We'll also corrrect one header file function prototype
so it will aggree with the actual function definition.
Reference(s):
top.c: In function 'adj_geometry':
top.c:1874:20: warning: comparison of integer expressions of different signedness: 'int' and 'long unsigned int' [-Wsign-compare]
1874 | if (Screen_cols < DOUBLE_limit) Curwin->rc.double_up = 0;
| ^
top.c: In function 'zap_fieldstab':
top.c:2359:26: warning: comparison of integer expressions of different signedness: 'int' and 'unsigned int' [-Wsign-compare]
2359 | if (wtab[EU_CPN].wmin < digits) {
| ^
top.c:2365:26: warning: comparison of integer expressions of different signedness: 'int' and 'unsigned int' [-Wsign-compare]
2365 | if (wtab[EU_NMA].wmin < digits) {
| ^
top.c: In function 'keys_summary':
top.c:5128:45: warning: comparison of integer expressions of different signedness: 'int' and 'long unsigned int' [-Wsign-compare]
5128 | if (w->rc.double_up && Screen_cols < DOUBLE_limit) {
| ^
top.c: In function 'sum_tics':
top.c:5605:22: warning: unused variable 'num_syst' [-Wunused-variable]
5605 | int ix, num_user, num_syst;
| ^~~~~~~~
top.c:5605:12: warning: unused variable 'num_user' [-Wunused-variable]
5605 | int ix, num_user, num_syst;
| ^~~~~~~~
Signed-off-by: Jim Warner <james.warner@comcast.net>
The 'iokey' function's parameter 'action' was utilized
with literal numbers in the calling functions. So this
change will replace those literal numbers with #define
constants which, one hopes, will clarify their impact.
Signed-off-by: Jim Warner <james.warner@comcast.net>
When in forest view mode, that focus toggle ('F') is a
useful tool occasionally. But, if a focused parent has
enough cloned siblings to exceed screen rows, it could
be hard to remember that such a toggle remains active.
So, this patch will provide a subtle visual clue added
to the leftmost position in the COMMAND column. Now if
the focus toggle was active, regardless of total tasks
affected, the users will always know when it's active.
Reference(s):
. -7/24/21, introduced new focus toggle
commit 3e922e671d
. 09/23/21, ensure focused tasks stay focused
commit 69978e3650
Signed-off-by: Jim Warner <james.warner@comcast.net>
[ you wouldn't believe how many back-and-forths were ]
[ involved in Craig convincing me there were several ]
[ inconsistencies. i am so dense sometimes (often?). ]
Signed-off-by: Jim Warner <james.warner@comcast.net>
Over the years the summary_show function has increased
from around 77 lines of code & comments to its current
size of 243 lines. This is well beyond an ideal length
of available screen rows. So this patch will split it.
We'll take the cpu and memory duties and make separate
functions out of them. Of course, this will incur some
additional call overhead but, given current cpu/memory
logic, any such increase really becomes insignificant.
Now summary_show's a svelte 57 lines of code/comments.
[ this is like what was done to that do_key function ]
[ a decade ago except overhead of new function calls ]
[ plus table lookup was even less of a concern since ]
[ a human was involved, not normal iterative output. ]
Signed-off-by: Jim Warner <james.warner@comcast.net>
In the patch referenced below the focus task logic was
improved to ensure that newly cloned tasks 'above' the
focused task did not force an effect like the up arrow
key. That commit also acknowledged that when some task
'above' ended, it *would* act like the down arrow key.
Well, with this commit a task ending 'above' a focused
task no longer distorts the focus. That's assuming the
new '#define FOCUS_HARD_Y' is specified plus the total
focused tasks does not exceed the current screen rows.
Thus, the manual scrolling with up and down arrow keys
is allowed when the total focused exceeds screen rows.
[ but keep in mind that when a focused task has been ]
[ hardened some otherwise useful toggles will not be ]
[ available. keystrokes like 'v' and even 'F' itself ]
[ can not be applied to another task with no scroll. ]
Reference(s):
. Sep, 2021 - 'focus' logic improved
commit 69978e3650
Signed-off-by: Jim Warner <james.warner@comcast.net>
This just updates the copyright dates in the documents
where I was already represented. Others are unchanged.
Signed-off-by: Jim Warner <james.warner@comcast.net>
As the issue cited below illustrates, a pids namespace
with proc mounted as subset=pid denies our library any
access to non-task data. In top's case, the result was
a fatal error message which involved "cpu statistics".
With this patch top will now assume an error involving
global cpu (stat) or memory (meminfo) data means we're
running under a restricted pids namespace. As such, an
attempt will be made to still display task level data.
[ if our assumption is incorrect, it's of no matter. ]
[ instead of a fatal error, we'll still try to offer ]
[ a user some minimally useful bit of functionality. ]
Reference(s):
https://gitlab.com/procps-ng/procps/-/issues/227https://www.freelists.org/post/procps/three-for-newlib,1
. 1st cut at subset=pid
commit bcb837b8c7
Signed-off-by: Jim Warner <james.warner@comcast.net>
Not only does that library tweak help to simplify some
top code, but now that ps snprintf fmtstr will finally
be accurate. That is two birdies with a single pebble!
Signed-off-by: Jim Warner <james.warner@comcast.net>
The separate threads for background updates were added
to top in the commit shown below. At that time cleanup
logic was added to end-of-job processing to cancel any
active threads and destroy any semaphores then in use.
That seemed like simple good stewardship with an added
benefit of avoiding potential valgrind 'possibly lost'
warnings for 320 byte blocks. Those blocks represented
an initial stack allocation for each of three threads.
All of that worked perfectly if such code was executed
under the main thread. In other words, if the keyboard
or a signal directed to any thread was used to trigger
program end. However, it might not always be the case.
Each of those 'refresh' routines supporting a separate
thread interacts with a newlib interface. As a result,
each is required to check the library's return result.
Should some error be detected, 'error_exit' is called.
Now we've got big problems with that eoj cleanup code.
One can't 'pthread_cancel' and 'pthread_join' a thread
from withing that same thread. Thus, when an error was
returned by the library with threads active, top would
hang with no possibility of removal short of a reboot.
So, this commit only executes that cancel/join cleanup
code when we are running under the main thread. Should
program end be triggered by a library error under some
sibling thread, all such cleanup will now be bypassed.
In the latter case, we will rely on documentation that
says any thread calling exit(3) will end every thread.
[ now, the only time we'll see any valgrind warnings ]
[ is with a library error, which is the least likely ]
[ scenario for running valgrind & top to begin with. ]
[ besides, if the valgrind warnings became a problem ]
[ one could easily add a 'warning-suppression' file. ]
Reference(s):
. Sep 2021, top introduced threads
commit 29f0a674a8
Signed-off-by: Jim Warner <james.warner@comcast.net>
When multi-threading was introduced in the patch shown
below, the former calls to sigprocmask were traded for
a pthread_sigmask call. This was done unconditionally.
As a result, even when those threads weren't enabled a
need to link with libpthread was created. In hindsight
the need should only arise when top is multi-threaded.
This commit will make pthread_sigmask use conditional.
Reference(s):
. 09/2021, separate threads introduced
commit 29f0a674a8
Signed-off-by: Jim Warner <james.warner@comcast.net>
When a visual separator was added to 2 abreast summary
items in a recent commit, this bug was introduced. So,
from that earlier patch we'll revert one line of code.
The bug surfaced under an 80/even column terminal only
when that '4' toggle was off. With an an 81/odd column
screen, it existed in both single and 2 abreast modes.
[ this commit also goes the extra distance to ensure ]
[ two abreast mode maximizes available screen width. ]
Signed-off-by: Jim Warner <james.warner@comcast.net>
In the patch referenced below, some logic was moved in
that frame_make function in an effort to overlap those
background threads with additional main thread duties.
Unfortunately, the auto-size column feature was broken
in the process. So this patch will revert a portion of
the earlier patch so as to restore the broken feature.
Reference(s):
. Sep 2021, top overlap additional processing
commit f11f43b4f7
Signed-off-by: Jim Warner <james.warner@comcast.net>
When that 'F' focus command has been applied to a task
in forest view it should remain as the topmost process
in a particular window. But without this patch that is
not guaranteed. Newly forked/cloned tasks 'above' such
a process result in task(s) appearing which shouldn't.
The effect was as if that up arrow key scrolled beyond
the topmost parent task, which would never be allowed.
[ since scrolling is permitted within a focus range, ]
[ when any task 'above' our focus/topmost task ends, ]
[ we respond as if scrolled with the down arrow key. ]
[ that result is completely appropriate. if the user ]
[ wishes to return to a focused parent, the up arrow ]
[ or home key can be used to accomplish such a goal. ]
Signed-off-by: Jim Warner <james.warner@comcast.net>
In an effort to improve the performance & decrease the
cost of our separate background update threads we will
relocate some overhead so that it might be overlapped.
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit just ensures that at program end all those
semaphores are removed in the same order for symmetry.
From thread's view of importance, we do least to most.
Signed-off-by: Jim Warner <james.warner@comcast.net>
Rather than having those 3 separate background threads
enabled by default, we'll turn them off until somebody
chooses to activate them. That seems more appropriate.
Signed-off-by: Jim Warner <james.warner@comcast.net>
After the stage had been set in the previous patch, in
this patch we will actually implement those background
updates via 3 separate threads. The design was simple:
. the do-while loops have now been made truly infinite
. 2 semaphores per thread allow needed synchronization
. 1 semaphore will provide for each thread to sem_wait
. 1 semaphore will provide for display o/p to sem_wait
. and all 3 thread's program name was made descriptive
A complication was the potential for a signal directed
to one of our new threads. Rather than having a thread
try to deal with such signals, we pass a mask with all
signals blocked at pthread_create time. Thereafter any
subsequent signals are forwarded to the parent thread.
[ also sigprocmask was exchanged for pthread_sigmask ]
[ since warned about use "in multithreaded process". ]
[ plus we also modified each of those POSIX comments ]
[ about 2004 to agree with current signal-safety(7). ]
Sadly, after all this effort there were no performance
benefits to having separate threads. In fact there was
a measurable performance degradation when running with
ever smaller delay intervals. But even with a delay of
1/10 second the 'real' cost increase is only about 1%.
There is one way whereby any additional costs might be
eliminated (at least seemingly). One could introduce 2
separate sets of contexts for each of those 3 threads.
Then retrieval & display could be overlapped. However,
the resulting display wouldn't represent the real-time
results. Rather it would be stale by 1 delay interval.
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit sets the stage for turning three functions
into code that can support a separate thread to update
their respective data while working in the background.
It involved relocating 1 function, renaming 2 routines
and adding parameters plus return addresses to each of
three functions. Those latter changes will be required
when issuing 'pthread_create' calls in the next patch.
The final step was organizing this code into what will
become the infinite do-while loop supporting a thread.
Signed-off-by: Jim Warner <james.warner@comcast.net>
The master branch top is pretty well insulated against
signal-unsafe function calls during end-of-job action.
However, this newlib branch top had exposed himself to
a bunch of such unsafe activity in the form of 'unref'
function calls to the new library. In those guys there
will be several 'free' invocations which are not safe.
So, this commit will simply correct such an oversight.
Signed-off-by: Jim Warner <james.warner@comcast.net>
This darn patch began as 1 correction to an extraneous
comma in the man document. Then, it grew to include an
adjustment for a couple of additional man page things.
After that, I figured I might as well tighten up logic
dealing with those awful gaps in the getopt_long code.
[ the error_exit mentioned in the associated comment ]
[ will only be taken when '=' ends the argv vectors! ]
Signed-off-by: Jim Warner <james.warner@comcast.net>
For quite some time now, top has stood out like a sore
thumb regarding the approach to cmdline options & help
text. Only short options were used and that same help
text was displayed for '-h' (help) plus 'v' (version).
[ also, top 'rolled his own' when it came to parsing ]
[ options while avoiding that getopt implementation. ]
Well, with this commit all of that has changed and top
now has added a long form of his options. Additionally
he employs getopt_long() for the bulk of that parsing.
[ however, top will still avoid separate fputs calls ]
[ characteristic of other procps-ng programs when it ]
[ comes to help. rather all such text is one string. ]
Along the way, the following major getopt deficiencies
were addressed, assuming the absence of a new #define:
* an equals sign ('=') is allowed on both option forms
* whitespace is allowed before & after the equals sign
* optional arguments needn't abut their related option
for short form nor is an '=' required with either form
Signed-off-by: Jim Warner <james.warner@comcast.net>
The commit referenced below forced a return to row one
whenever there was a thread mode transition. Now, with
our new focus ('F') feature, we should adopt a similar
philosophy so as to avoid potential display anomalies.
Reference(s)
. Jun 2018, force row 1 w/ thread mode transition
commit 19ec80bd41
Signed-off-by: Jim Warner <james.warner@comcast.net>
This patch makes it possible to honor our '-H' threads
switch even when operating under that '-p' switch used
to select/filter only certain pids. Of course, what we
have done is just exploit the new library enhancement.
[ and we're pretending we don't know the enumerators ]
[ PIDS_FETCH_THREADS_TOO + PIDS_SELECT_PID are equal ]
[ to that PIDS_SELECT_PID_THREADS enumerator itself. ]
[ thankfully, those dependencies can be assured with ]
[ a 'make check-lib' which exploits ITEMTABLE_DEBUG! ]
Signed-off-by: Jim Warner <james.warner@comcast.net>
