Releases: dbohdan/memsparkline
Releases · dbohdan/memsparkline
v0.9.4
- Fixed: print the summary even when the command we run fails.
- Fixed: only print a newline before an error message if the error message is about receiving a signal.
The newline changes
████████ 1.9^CError: received signal: interrupt
into
████████ 1.9^C
Error: received signal: interrupt
However, we need to check whether the error is a signal so we don't print an extraneous newline before other error messages.
v0.9.3
v0.9.2
- Fixed memsparkline not counting the memory usage of the first process it tracks.
- Fixed a test intermittently failing with a fairly high frequency.
- Stopped ignoring the error when getting the memory usage of a descendant process fails.
- Added Bash and fish completion files.
v0.9.1
v0.9.0
v0.8.1
v0.8.0
- Minor breaking change in the CLI: memsparkline will process the options
-h
/--help
and-v
/--version
like other options. This means they must be given before the first positional argument. These options were special and processed in any position. It was supposed to make help easier to access, but in practice it caused problems more than it helped. What this change means is that the option-h
in the commandmemsparkline ls -h
used to make memsparkline print help and exit; now-h
is passed tols
.
v0.7.1
v0.7.0
- memsparkline has been ported from Python to Go.
- Static binaries are now available.
- Changed
-d
/--dump
to append to the output file for consistency with-o
/--output
. You can accidentally lose old data if you expect--dump
to behave like--output
. Since we prefix every usage figure with a timestamp, it should be reasonably easy to separate the output from different runs. - NetBSD is not supported until gopsutil supports NetBSD.