New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
event manager doesn't "connect"(?) unless uzbl-browser is being verbose #393
Comments
Output of |
This may be a race condition. Could you inject |
Oh, and upload the |
Ok, here are the hopefully relevant parts of the log. First, opening the socket:
Then, something I noticed whilst scrolling over the log:
Writes to the socket (fd
|
To me this looks a bit like some missed error handling when writing (possibly long) event messages over the socket. That handling of |
I found the issue and am working on a patch. |
The socket is opened in non-blocking mode:
Thus when writing a lot of data (the long path names) the socket write call returns these I fixed the affected function in #400 |
#401 fixes a source for issues like this in current master (or v0.9.1). (The earlier pull request is a patch for v0.9.0) |
When I just run
uzbl-browser
, then I get neither[Cmd]
nor[Ins]
as mode indicator but[]
. The FAQ says that this could be due to a nonrunningconnected event manager. I'm running on NixOS, where there's often the issue that runtime dependencies are not found due to funny paths (/nix/store/....foobar/bin/baz
), so I tried runninguzbl-browser -v
in the hope to see which path it's trying to find the event manager. I didn't get any useful output, but noticed that now my mode indicator showed[Cmd]
just fine.So, after some further investigation as for how reproducible this behavior is, I come to the conclusion: For some reason my event manager doesn't seem to
start (or"connect"?)unless I either print verbose information (-v
), event information (-p
) or both. If I'd need to guess I'd say that this might be some kind of race condition, where "slowing down" the start ofuzbl-browser
allows the event manager to catch up? Though throttling throughcpulimit
doesn't work ...I guess that isn't too helpful, so I digged into my distros package / build system to find that it's supposedly version
v0.9.0
, downloaded fromhttps://github.com/uzbl/uzbl/archive/v0.9.0.tar.gz
. I'm a bit confused, though, because the checksum doesn't match what I was expecting. Mine:Expected as far as package manager is concerned:
No, happens on every page.
The text was updated successfully, but these errors were encountered: