RCP: 0.7.9, CQP corpus re-load "not ready"

Currently, reloading a X CQP corpus in the current session (search engine restart) can lead to a “X corpus not ready” status whereas the X corpus could be correct (on disk).

Solution

• after X import or load, check the X corpus status: if the status is not ready ->
  1. drop each CQP corpus (to free as much of C heap memory as possible)
  2. unload cqplib (to free remaining memory pointers - typically in C stack memory - but assuming possible memory leaks)
  3. load cqplib
  4. load each CQP corpus
     • check the X corpus status: if the status is not ready -> invite to quit TXM and relaunch it (see #2351)

(from redmine: issue id 2352, created on 2018/03/06 by Serge Heiden)

• Relations:
  • relates #2351 (closed)

added import type::bug + 1 deleted label

changed milestone to %TXM 0.8.0

closed

assigned to @mdecorde