Some configuration systems might link a htop configuration file and we
don't really want to replace the symlink but rather its source. This
will also allow us to fail in case the source is read only.
Signed-off-by: Sefa Eyeoglu <contact@scrumplex.net>
Rarely (if ever) do we see references in the documentation to `pcp htop` (it's mainly just `pcp-htop`, with a hyphen instead of a space). I fixed the documentation to reflect this.
Feel free to ignore this if the space is intentional.
Over in the PCP project, Coverity has spotted a couple of minor
issues in the htop code.
This commit addresses them:
Settings.c
need to take control of the umask before calling mkstemp()
pcp/Platform.c
Although Platform_getMaxPid() returns a pid_t (which
strictly speaking is signed), the value is used to call
Row_setPidColumnWidth() and a negative value here is not good
... using INT_MAX instead of UNIT_MAX on the (unlikely) error
path makes no practical difference but may keep Coverity quiet
Kernel threads are commonly not based on an executable and their cmdline
therefore just a human readable string.
On Linux this string might contain slashes, e.g. kworker/7:5H-ttm, which
cause Process_writeCommand() to print only the trailing parts if the
option *Show Program Path* is disabled.
Reported-and-Suggested-By: mmrmaximuzz
Fixes: #1391
Compare the return value of sscanf(3) and fscanf(3) explicitly against
the expected number of parsed items and avoid implicit boolean
conversion. Such an implicit conversion would treat EOF (-1) the same
as at least one item parsed successfully.
Reported by CodeQL.
Make it more clear the variable stores the previous tty number (similar
to lasttimes), while the current one gets set in the intermediate
LinuxProcessTable_readStatFile().
In case the string to be printed contains invalid wide characters,
continue to translate after the invalid character instead of aborting.
Closes: #1373
The last argument of RichString_appendnWide() should be the number of
bytes of the source multi-byte string, not the number of characters in
the resulting wide-character string.
While abstracting processes from row display, code relating to
row parent sorting was moved to row-level that in fact must be
done at process-level (using the Row_compareByParent_Base call
circumvented this unintentially).
Resolves#1378
Every now and then users complain about settings to being persisted.
Most of the time this turns out to be caused by user: Exiting with
signal causes htop not to save settings to htoprc.
Make htop print a warning when exiting on signal.
If the user removes all meters from the header on exit htop will write
no values for the settings column_meters_X and column_meter_modes_X.
The parser skips all configuration settings with no value, so on the
next start no header meter related key is processed and htop will add a
set of default meters to the header.
Write instead an invalid value of `!`, so the keys are parsed and htop
will not add the default meters back.
Closes: #1248
Christian Göttsche
Daniel Lange
fraggerfox
Sefa Eyeoglu
Nathan Scott
kew24
Ken McDonell
Sergey Fedorov
Benny Baumann
Christian Hesse
Dominik Hassler
Samuel Venable
Explorer09