You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
[BUG] Large chats with media don't import at all #90
I was trying to import a chat (zip file along with media) but nothing happens when I do that. The size of the file I'm importing is about 3.5GB. It seems like the app is doing nothing when I import that file. When I load a ZIP file that's 1.5GB, the app loads for a while but it eventually works.
The chat has about 2500 actual media and thousands and thousands of stickers (which also count as images I guess). And there are about 85000 messages.
Let me know if I can help somehow to help debug the issue.
Triage update (2026-10-13): this remains a valid bug. The current import path still reads the entire archive into an ArrayBuffer and then passes it to JSZip.loadAsync(), so a 3.5 GB ZIP is resident before lazy media handling begins. That load stage also has no progress callback, which explains the apparent stall. I found no streaming implementation, large-file regression benchmark, or candidate PR yet.
The durable fix needs incremental ZIP processing in a worker plus a representative large-file benchmark. Please do not upload the private archive; browser/OS, available RAM, and whether the desktop or web build was used would still help characterize limits.
Activity
[-]Large chats with media don't import at all[/-][+][BUG] Large chats with media don't import at all[/+]rodrigogs commentedon Oct 13, 2026
Triage update (2026-10-13): this remains a valid bug. The current import path still reads the entire archive into an
ArrayBufferand then passes it toJSZip.loadAsync(), so a 3.5 GB ZIP is resident before lazy media handling begins. That load stage also has no progress callback, which explains the apparent stall. I found no streaming implementation, large-file regression benchmark, or candidate PR yet.The durable fix needs incremental ZIP processing in a worker plus a representative large-file benchmark. Please do not upload the private archive; browser/OS, available RAM, and whether the desktop or web build was used would still help characterize limits.