- Issue created by @fgm
- π«π·France fgm Paris, France
More info: when generating the migrations with migrate:upgrade, the
migrate_plus.migration.upgrade_d7_taxonomy_vocabulary
migration entity forprocess.vid.length
has value 30, which explains why the vocabulary names are truncated to 30 characters instead of the 32 allowed by machine names.The generated terms migration e.g.
upgrade_d7_taxonomy_term_vocabulaire_de_l_horreur_et_du_malheur
correctly references the new vid using themigration_lookup
plugin, for which the migrate_map entry shows the name has been correctly truncated.However, when checking the migrate_map entry for the field_instance migration, the
destid2
is the unchanged source name of. the vocabulary instead of the updated vocabulary. This appears to be due to theprocess.bundle
usingsource = ['entity_type', '@bundle_mapped']
, wherebundle_mapped
is a constant map only remapping thecomment_forum
bundle tocomment_node_forum
and ignoring the other possible renamings. - Status changed to Needs work
over 1 year ago 4:58pm 27 September 2023 - π«π·France fgm Paris, France
As a quick hack, modifying the generated
migrate_plus.migration.upgrade_d7_field_instance.yml
to add the mapping manually allows the upgrade_d7_field_instance not to fail.process: # ... bundle_mapped: - plugin: static_map source: bundle bypass: true map: comment_node_forum: comment_forum # Add a line like the one below the_name_that_is_really_too_long: the_name_that_is_really_too # ...
Now, the question is how to automate this.