force client side reload on update

[stdweird]

What happened?

It seems that when updating to 0.7.8-rc1 (from 0.7.7-rc1), the user needs to reload the page (eg pressing f5 in firefox) manually; otherwise they are stuck in previous version (or worse, some mixture of old and new).

it's easy enough to spot because the version is displayed at the bottom, but is there no way to notify the user this is needed; or trigger this from the server?

Version Information

using 0.7.8-rc1 from docker repo (but using podman via nomad)

Steps to Reproduce

• deploy new release
• visit url
• notice old version at bottom

What browsers are you seeing the problem on?

Firefox

Relevant log output

no real errors (deleting an old chat triggered an error, but that was probably mixing old and new code somehow).
the version at the bottom is the main indicator, but (imho) you can't expect a normal user to know what the correct version should be.

as an admin, i first thought the old container was being deployed instead of the new one; but i verified that the correct new container was being deployed

Screenshots

No response

Code of Conduct

• I agree to follow this project's Code of Conduct

[danny-avila]

This is a browser caching issue. There are ways to prevent it altogether which is discussed in the documentation: https://www.librechat.ai/docs/configuration/dotenv#static-file-handling

some way to notify the user would be interesting and possible, I will have to think about it

[stdweird]

@danny-avila what "index.html" is this referring to? when looking at the cached html documents when doing a normal reload, i get 0 cached html docs. or did something change in 0.7.8 wrt 0.7.7-rc1?

[danny-avila]

index.html is what gets serviced from the backend to render all the client elements. It points to different static assets every time the compiled code gets changed (from commit to commit).

[stdweird]

sure, but if there are no cached html document on a reload, this is either fixed or something else. when final 0.7.8 is released, i will pay specific attention to this after i deploy it