Add ability delete received private messages by smorks · Pull Request #6378 · LemmyNet/lemmy

smorks · 2026-03-04T03:25:38Z

This should fix #4120.

It also required PR's in lemmy-ui and lemmy-js-client, that I will open shortly too.

There are some issues with this but I figure I'd get this into a PR before I forget about it again, haha.

The only remaining issue that I know about is that this also causes the message to stop showing for the sender too. Pretty sure it just needs the view modified so it only filters the message for when the recipient of the message matches the current user, but I haven't looked to hard into doing that yet.

Edit: Oops, I just noticed that I should update the timestamp of the migration file too.

crates/api/api_crud/src/private_message/delete_for_recipient.rs

Nutomic · 2026-03-04T14:42:04Z

crates/db_views/notification_sql/src/lib.rs

      .and(not(private_message::deleted))
-      .and(not(private_message::removed)),
+      .and(not(private_message::removed))
+      .and(not(private_message::deleted_by_recipient)),


Needs to be something like this, so that these messages are only excluded if current user is not the creator (best write a unit test to ensure it works correctly).

Suggested change

.and(not(private_message::deleted_by_recipient)),

.and(private_message::creator_id.eq(person_id).or(not(private_message::deleted_by_recipient))),

I tried that code and it appears to crash the server process with a stack overflow error. I can get the exact error message (& stack trace), but I remember running into this before too. Will poke around to see why that's happening, but any insight would be appreciated!

this is the error I receive, tried with RUST_BACKTRACE=1 which didn't seem to give any useful information either:

thread 'actix-server worker 3' (19510) has overflowed its stack fatal runtime error: stack overflow, aborting

Here isn't the right place to try to extract the creator. We have two options:

Do it like you've done here, and don't show messages you've sent that the recipient deleted

Add the deleted_by_recipient logic somewhere inside this function. That does have access to the creator and recipient.

I spose it might be important for message creators to still view their past messages, so the 2nd option is maybe better.

This also needs a unit test in crates/api/api_utils/src/notify.rs

Ok, I managed to get it working, but I had to in-memory filtering, which is probably not ideal?

I couldn't add on to the existing filter you mentioned because as far as I can tell, I have no way of getting to the deleted_by_recipient field of the private message, from the notification?

I also added a unit test, hopefully it's sufficient.

dessalines

Almost there, nice job so far. I'd def add a unit test somewhere in crates/api/api_utils/src/notify.rs . There are other tests for private messages in there already.

dessalines · 2026-03-05T14:38:53Z

crates/db_views/notification_sql/src/lib.rs

      .and(not(private_message::deleted))
-      .and(not(private_message::removed)),
+      .and(not(private_message::removed))
+      .and(not(private_message::deleted_by_recipient)),


Here isn't the right place to try to extract the creator. We have two options:

Do it like you've done here, and don't show messages you've sent that the recipient deleted

Add the deleted_by_recipient logic somewhere inside this function. That does have access to the creator and recipient.

I spose it might be important for message creators to still view their past messages, so the 2nd option is maybe better.

This also needs a unit test in crates/api/api_utils/src/notify.rs

also add tests.

dessalines · 2026-03-05T20:17:11Z

crates/db_views/notification/src/impls.rs

+      .filter(|r| {
+        r.private_message
+          .as_ref()
+          .is_none_or(|pm| !(pm.deleted_by_recipient && pm.recipient_id == my_person.id))
+      })


This works but we should really be doing this on the query side somewhere. I'll take a look and see if I can figure out where.

Agreed. I poked around a bit to try and do it in the query part you referenced before, but couldn't figure it out. I'll take another look later today too.

It can probably be done with query = query.filter(...) just before the let paginated_query. You check if private message is null (none) by checking if one column is null. Probably notification::private_message_id.is_null() if such a column exists.

I figured it out I think: smorks#1

Unfortunately I had to remove one join (post image details) because with notifications, we're hitting the limits of diesel's joins. That's for us to deal with tho.

Ahhhh ok, so that explains the stack overflow I was getting when doing it that way too.

Thanks for looking into that!

No probs, ya sry about that, its a diesel related bug because we have too many joins there.

Nutomic · 2026-03-06T09:23:08Z

crates/db_views/notification_sql/src/lib.rs

-    .left_join(instance_join)
+  // This causes a stack overflow currently, due to diesel join limitations.
+  // image_details aren't critical for for post notifications anyway.
+  // .left_join(image_details_join())


There are only 20 joins here. I tried to uncomment these lines and it works just fine with cargo 1.93

It failed on my local machine for 1.92, but I'll try it for that one.

I'm still getting the same overflowed stack error, even on 1.94

Nevermind, had it added at the wrong place. I'll try again.

I also got overflows on both 1.93 and 1.94, with that line uncommented.

I got it, I had to change the position of the join to make it earlier.

Normally diesel isn't this finicky, its just with the large amount of joins there.

Once I get CI passing, I'll push directly to this branch. Won't be any logical changes, just clippy fixes.

…iv-msgs-dess

dessalines · 2026-03-06T19:31:01Z

I'll handle the clippy errors and such, and push directly to your branch to handle them.

dessalines · 2026-03-07T12:38:37Z

Added back in the image_details_join, needed to add Box::pin to appease clippy, and upgraded to rust 1.94 to fix the diesel join. Once this passes it can be merged.

Nutomic · 2026-03-09T10:49:46Z

crates/db_views/notification/src/impls.rs

@@ -173,12 +173,17 @@ impl NotificationQuery {
    }


Change the definition of this function to:

pub fn list(...) -> impl Future<Output = LemmyResult<PagedResponse<NotificationView>>> { Box::pin(async move { ... })

Then its not necessary to use Box::pin for every single function call.

I'll try it out.

Nutomic · 2026-03-09T14:31:22Z

Thanks @smorks!

smorks · 2026-03-09T16:28:49Z

Thanks @smorks!

No problem! I'll look at getting those UI changes done soon as well so this can be completely done. thanks to both of you for helping me out with this one!

smorks requested review from Nutomic, SleeplessOne1917, dessalines, dullbananas and phiresky as code owners March 4, 2026 03:25

This was referenced Mar 4, 2026

add delete priv msg by recipient LemmyNet/lemmy-js-client#875

Merged

add priv msg delete by recipient button LemmyNet/lemmy-ui#3942

Merged

smorks force-pushed the delete-priv-msgs branch 4 times, most recently from 48848bf to ba3af7c Compare March 4, 2026 04:10

Nutomic reviewed Mar 4, 2026

View reviewed changes

crates/api/api_crud/src/private_message/delete_for_recipient.rs Outdated Show resolved Hide resolved

Nutomic reviewed Mar 4, 2026

View reviewed changes

smorks force-pushed the delete-priv-msgs branch 4 times, most recently from d3d60b0 to 08fdba7 Compare March 4, 2026 20:34

dessalines reviewed Mar 5, 2026

View reviewed changes

smorks added 2 commits March 5, 2026 12:37

add private message delete by recip route and logic

237a591

only filter pm on deleted by recipient if we are the recipient

d923e31

also add tests.

smorks force-pushed the delete-priv-msgs branch from 08fdba7 to d923e31 Compare March 5, 2026 18:38

dessalines reviewed Mar 5, 2026

View reviewed changes

dessalines added 2 commits March 5, 2026 16:14

Filter out private_message.deleted_by_recipient in join

01766a2

Filter out private_message.deleted_by_recipient in join

ff3f62a

Nutomic reviewed Mar 6, 2026

View reviewed changes

dessalines added 2 commits March 6, 2026 14:16

Merge remote-tracking branch 'smorks/delete-priv-msgs' into delete-pr…

3da95bc

…iv-msgs-dess

Try adding Box::pin for clippy.

3c6de89

dessalines added 2 commits March 6, 2026 14:55

Re-add image_details join

34f546d

Upgrading rust version.

cd2fe94

dessalines and others added 7 commits March 6, 2026 16:16

Clippy 1

f3f30f1

clippy 2

609d5a6

clippy 3

abbc20c

Adding a few more box pin

b1eb9a9

clippy 4

397d582

Fixing cargo api common

e165489

Merge branch 'main' into delete-priv-msgs

ba4c5d0

dessalines approved these changes Mar 7, 2026

View reviewed changes

Nutomic reviewed Mar 9, 2026

View reviewed changes

Move Box::pin up a level

a69d722

Nutomic approved these changes Mar 9, 2026

View reviewed changes

Nutomic enabled auto-merge (squash) March 9, 2026 14:31

dessalines mentioned this pull request Mar 9, 2026

Add ability to delete private messages LemmyNet/lemmy-ui#3955

Closed

3 tasks

Nutomic disabled auto-merge March 9, 2026 15:17

pin

d0aa71f

Nutomic enabled auto-merge (squash) March 9, 2026 15:22

Nutomic merged commit 2a696c5 into LemmyNet:main Mar 9, 2026
1 check passed

	.and(not(private_message::deleted_by_recipient)),
	.and(private_message::creator_id.eq(person_id).or(not(private_message::deleted_by_recipient))),

Uh oh!

Conversation

smorks commented Mar 4, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dessalines left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Nutomic Mar 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dessalines Mar 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dessalines commented Mar 6, 2026

Uh oh!

dessalines commented Mar 7, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Nutomic commented Mar 9, 2026

Uh oh!

Uh oh!

smorks commented Mar 9, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

smorks commented Mar 4, 2026 •

edited

Loading

dessalines left a comment •

edited

Loading

Nutomic Mar 6, 2026 •

edited

Loading

dessalines Mar 6, 2026 •

edited

Loading