oss/miniflux-v2
oss/miniflux-v2
1
0
Fork 0
mirror of https://github.com/miniflux/v2.git synced 2025-08-26 18:21:01 +00:00
Code Wiki Activity

feat(storage): add limit parameter to ClearRemovedEntriesContent

Browse source 
Without the limit, this query is going to hangs forever on large
databases with millions of entries.
This commit is contained in:
Frédéric Guillot 2025-08-17 17:30:33 -07:00
parent 905d652511
commit 5403ca09f6
2 changed files with 13 additions and 6 deletions
Download patch file Download diff file Expand all files Collapse all files

2
internal/cli/cleanup_tasks.go
View file

@ -54,7 +54,7 @@ func runCleanupTasks(store *storage.Storage) {
slog.Int64("removed_entries_enclosures_deleted", enclosuresAffected)) slog.Int64("removed_entries_enclosures_deleted", enclosuresAffected))
} }
if contentAffected, err := store.ClearRemovedEntriesContent(); err != nil { if contentAffected, err := store.ClearRemovedEntriesContent(config.Opts.CleanupArchiveBatchSize()); err != nil {
slog.Error("Unable to clear content from removed entries", slog.Any("error", err)) slog.Error("Unable to clear content from removed entries", slog.Any("error", err))
} else { } else {
slog.Info("Clearing content from removed entries completed", slog.Info("Clearing content from removed entries completed",

17
internal/storage/entry.go
View file

@ -297,7 +297,7 @@ func (s *Storage) DeleteRemovedEntriesEnclosures() (int64, error) {
} }
// ClearRemovedEntriesContent clears the content fields of entries marked as "removed", keeping only their metadata. // ClearRemovedEntriesContent clears the content fields of entries marked as "removed", keeping only their metadata.
func (s *Storage) ClearRemovedEntriesContent() (int64, error) { func (s *Storage) ClearRemovedEntriesContent(limit int) (int64, error) {
query := ` query := `
UPDATE UPDATE
entries entries
@ -305,12 +305,19 @@ func (s *Storage) ClearRemovedEntriesContent() (int64, error) {
title='', title='',
content=NULL, content=NULL,
url='', url='',
author=NULL author=NULL,
WHERE comments_url=NULL,
status=$1 AND content IS NOT NULL document_vectors=NULL
WHERE id IN (
SELECT id
FROM entries
WHERE status = $1 AND content IS NOT NULL
ORDER BY id ASC
LIMIT $2
)
` `
result, err := s.db.Exec(query, model.EntryStatusRemoved) result, err := s.db.Exec(query, model.EntryStatusRemoved, limit)
if err != nil { if err != nil {
return 0, fmt.Errorf(`store: unable to clear content from removed entries: %v`, err) return 0, fmt.Errorf(`store: unable to clear content from removed entries: %v`, err)
} }