Skip to content

document is null error, prevents notebook saving #1602

@parente

Description

@parente

There appears to be a weird timing bug with showing the logo for the current kernel spec and with the DOM being ready. I typically see this when accessing a remote notebook server where the connection is a bit slow.

screen shot 2016-07-07 at 9 46 47 am

When it happens, the problem is undetectable (short of watching the JS console) until I go to save the notebook. Then I get the dreaded:

at which point I have to run Jupyter.notebook._fully_loaded = true to save my work (if I know this cheat code) or cuss because I can't save whatever I've typed into the notebook since loading it (if I don't know the cheat code).

The problem seems to be coming from here:

I don't mind taking a crack at a fix since it's really annoying when it bites. But I need a bit of guidance. Is there or should there be a DOM-ready guard somewhere earlier in the call stack before this method gets invoked, or does this method need to guard against DOM-readiness itself before calling img.show or hide? What about other places in the code?

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions