Age | Commit message (Collapse) | Author |
|
We have a pipe index (0 to n_read), not a process index (0 to n_proc).
These are two separate lists, with no correspondence between them.
See 7dbbfa4 for another fix related to this issue.
|
|
|
|
|
|
|
|
|
|
|
|
Crash introduced by 775bbc2d. How silly of me.
Thanks to Ken Beyerlein for fixing this.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
They're x/y offsets, not fs/ss
|
|
|
|
They're x/y coordinates, not fs/ss. Fixed the labelling as well.
|
|
|
|
Because the status message is only printed after possibly skipping over
a non-indexed pattern, if n_chunks is used it might skip entire
messages. Better to use n_images, which is always incremented just before
possibly printing the message.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This makes the Greek letters come up at SLAC.... again.
Reverts 779d87eb08fa2b185c43a82814aecc0e1a00eef9.
|
|
Fixes corrupted output when heavily multi-processing.
|
|
|
|
|