halsampler — sample data from HAL in realtime




sampler and halsampler are used together to sample HAL data in real time and store it in a file.

sampler is a realtime HAL component that exports HAL pins and creates a FIFO in shared memory. It then begins sampling data from the HAL and storing it to the FIFO.
halsampler is a user space program that copies data from the FIFO to stdout, where it can be redirected to a file or piped to some other program.


instructs halsampler to read from FIFO CHAN .
FIFOs are numbered from zero, and the default value is zero, so this option is not needed unless multiple FIFOs have been created.

instructs halsampler to read COUNT samples from the FIFO, then exit.
If -n is not specified, halsampler will read continuously until it is killed.

sets an explicit HAL component name.

instructs halsampler to tag each line by printing the sample number in the first column.

instructs halsampler to write to FILENAME instead of to stdout.


A FIFO must first be created by loading sampler with
halcmd loadrt
or a loadrt command in a .hal file.

Then halsampler can be invoked to begin printing data from the FIFO to stdout.

Data is printed one line per sample. If -t was specified, the sample number is printed first.
The data follows, in the order that the pins were defined in the config string.

For example, if the sampler config string was "ffbs" then a typical line of output (without -t ) would look like:
123.55 33.4 0 -12

halsampler prints data as fast as possible until the FIFO is empty, then it retries at regular intervals, until it is either killed or has printed COUNT samples as requested by -n.

Usually, but not always, data printed by halsampler will be redirected to a file or piped to some other program.

The FIFO size should be chosen to absorb samples captured during any momentary disruptions in the flow of data, such as disk seeks, terminal scrolling, or the processing limitations of subsequent program in a pipeline.

If the FIFO gets full and sampler is forced to overwrite old data, halsampler will print 'overrun' on a line by itself to mark each gap in the sampled data.
If -t was specified, gaps in the sequential sample numbers in the first column can be used to determine exactly how many samples were lost.

The data format for halsampler output is the same as for halstreamer input, so 'waveforms' captured with halsampler can be replayed using halstreamer.
The -t option should not be used in this case.


If a problem is encountered during initialization, halsampler prints a message to stderr and returns failure.

Upon printing COUNT samples (if -n was specified) it will shut down and return success.
If it is terminated before printing the specified number of samples, it returns failure.
This means that when -n is not specified, it will always return failure when terminated.




Original version by John Kasunich, as part of the LinuxCNC project. Improvements by several other members of the LinuxCNC development team.


Report bugs to the Machinekit forum!forum/machinekit

Copyright (c) 2006 John Kasunich. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.