Henrik Lemke
98d8e4a0c5
- docs/did_you_know/: one short tip per feature added since May 2026 (<date>-<slug>.md, YAML front matter: added/tags/details), listed newest first; index.md describes the format for adding more. - New pages for features that had none: launching, command_history, tab_completion, live_gui, tweak, scanning, status_server, channel_access, shared_files, schematic_panels. Wired into index.rst together with the existing queues page. - Fix archiver_stripchart/listening_monitor: the method is archiver.strip_plot (strip_chart does not exist on DataHub); document the per-device .strip_plot(). - NAMESPACE_MONITOR_FORMAT.md: scan monitoring is wired into Daq's scan callbacks since 2026-09-09. - scripts/eco-docs (+ -install-user-service): build the docs from an own clone and serve them over HTTP as user systemd units, rebuilt by a timer; a failed build never replaces the served site.
1.7 KiB
Tab completion on lazy devices
Most components in the Bernina namespace are lazy: they are only built (with real EPICS connections) the first time you use them. Tab completion has to look inside a device to list its children, so without care, pressing Tab would quietly build devices you never meant to touch.
Two Tabs to initialize
If the device you are completing into has not been initialized yet, the first Tab does nothing except tell you so:
In [1]: prepump.<TAB>
eco: 'prepump' is not initialized yet -- press Tab again to initialize it and complete.
In [1]: prepump.<TAB>
check gp isolate_all line1_usd line2_lic pump_down roots_pump show vent ...
The second Tab initializes it and completes as usual. This works at any depth
(bernina.something.<TAB>) and still counts if you kept typing between the
two Tabs.
Only the one component named in the expression is ever initialized, never its siblings that happen to share a prefix. Completing into a device that is already initialized works on the first Tab, as before.
What makes completion work on eco objects
You normally don't need to know this, but if completion breaks after an
IPython upgrade, these are the three settings involved. eco sets them in every
session it starts (terminal, desktop console, start_console()):
Completer.use_jedi = False. jedi resolves every candidate to inspect it, which on a lazy namespace means initializing everything with a matching prefix. It also returns no matches at all on eco's dynamic objects.- eco's lazy-proxy class is allow-listed in IPython's
guarded_evalpolicy. Without it, completing two levels deep (prepump.line1_usd.<TAB>) silently returns nothing. - The two-Tab gate described above.