TXM: 0.7.6, corpus import must not accept invalid corpus names

Currently, CQP search engine cannot manage corpora with a name beginning with a digit: 0..9.

More generally, corpus names must follow a ‘CQP naming policy’ to be manageable by CQP.

Currently, the User Interface of corpus import or corpus load accept corpus names
that don’t respect CQP naming policy.

The result is that when the import or the load is finished with success,
the corpus is available in the corpus view but is not useable by CQP.

Solution

• corpus import:
  • in the import parameters window, the corpus name input field must indicate a corpus naming policy violation if the currently given corpus name doesn’t respect it (for example with a red cross) OK the corpus name check is done when the user starts the import process.
  • the import process must NOT start until the corpus name doesn’t respect the corpus naming policy. OK

see ticket #1033 for load process changes

Validation test

• try the “42”, “àéuazeDF” corpus names
• and attempt to start the import the “Nom du corpus* …” label background must go red

MD: OK Linux64

(from redmine: issue id 975, created on 2014/08/08 by Serge Heiden)

added import priority::urgent type::bug + 1 deleted label

changed milestone to %TXM 0.7.7

MD: in the import editor :

• When user select the source directory : the generated corpus ID manages corpus with a name beginning with a digit: 0..9.
• When user start the import : the corpus ID is validated with the [A-Z][-A-Z0-9]* regular expression

(from redmine: written on 2014/09/12 by Matthieu Decorde)

Checked, corpus is not imported but some directories are created, see: #1290

(from redmine: written on 2015/04/02 by Sebastien Jacquot)

closed

assigned to @mdecorde