Add attachment annotations, converted_from, as well as a couple other nice things #7108
Conversation
...tions/versions/20250925_1631_e4f49c197158_add_annotations_and_converted_from_id_collation.py
Outdated
Show resolved
Hide resolved
...tions/versions/20250925_1631_e4f49c197158_add_annotations_and_converted_from_id_collation.py
Outdated
Show resolved
Hide resolved
...tions/versions/20250925_1631_e4f49c197158_add_annotations_and_converted_from_id_collation.py
Outdated
Show resolved
Hide resolved
pferreir
force-pushed
the
add-attachment-annotations
branch
2 times, most recently
from
2af5c08 to
38c648c
Compare
pferreir
force-pushed
the
add-attachment-annotations
branch
2 times, most recently
from
1042d1b to
da77adf
Compare
ThiefMaster
left a comment
ThiefMaster
left a comment
There was a problem hiding this comment.
LGTM, but I think it's best to (test and) merge it together w/ the plugin changes
| # remove PG* variables from environment, as they may cause the following commands to fail | ||
| for env_var in ('PGUSER', 'PGPASSWORD', 'PGHOST', 'PGPORT', 'PGDATABASE'): | ||
| os.environ.pop(env_var, None) |
There was a problem hiding this comment.
Would it make sense to discard any PG* env vars?
There was a problem hiding this comment.
I didn't want to be so dramatic, in case there is some legit use for some variable starting with PG. But I'm not sure.
Any particular reason why you would not merge this one first? Just afraid that it will become stale. |
|
Primarily in case anything is missing... but OTOH it looks pretty complete. I guess you tested the collation stuff with weird characters (emoji, chinese and whatever) so it doesn't fail with an error in case of such content? |
I tried it with some special characters. Not necessarily Asian scripts. |
Also add case insensitive collation to the DB
Co-authored-by: Adrian <[email protected]>
ThiefMaster
force-pushed
the
add-attachment-annotations
branch
from
a473fb1 to
611fad6
Compare
I don't like to just cascade there, even though this should only be used on test databases with no valuable data
3 participants
This adds:
converted_from(and backrefconverted_to) to everyAttachment, which can of course beNone, and will help us keep track of attachments generated from other attachmentsannotationsfield/column (JSONB), useful to add metadata such as origin of the attachment (CloudConvert? CodiMD archiver? others?)case_insensitive, to the database, which does case-insensitive "predictable" sorting of names which takes into account also punctuation characters (contrary to Postgres' default). This is added only because I noticed that the sorting in attachments names was unintuitive