ucla-cs130 / team22 Goto Github PK

lcov only works for integration test again

async_read_until fails sometimes with reverse_proxy_handler.

telnet - no problem
curl - no problem
chrome - happens with the ucla.edu redirection
nc - doesn't work at all

Known causes of async_read_until fail:

• early termination by client
• EOF received from client
• request header is too long for the buffer
• some other exception thrown inside the read_until
• timeout?

We could just make it a warning?

it's too complicated to fit in a static file handler

sometimes the status page logs 600, what is this?

only crashes on Docker

{, }, " don't work because they don't get through the config parser.

Requests to relative paths with different base urls are added together.
ex.
/proxy1/somefile.png -> somefile.png (google.com/somefile.png)
/proxy2/somefile.png -> somefile.png (wikipedia.org/somefile.png)

Status page
somefile.png 2

mutex::lock() causes a segfault

It is shared state without locks or anything.

It is possible that a connection gets deleted while the status page is querying the connection state.

Possible fixes: (anything that removes the dependency on connection)

• Shared pointer on entire connection
• Shared pointer on connection state
• Connection tells state to status, instead odd stat us asking

spam refresh on /sleep handler or /proxy3 and the server will probably crash.
go to /sleep and close the tab and the server will probably crash.

Missing the example files example_config, config_with_nested, large_config

Request parsing has a bug -- the last header's value has an additional \r postpended.

It shouldn't matter much, but it's a bug nonetheless.

^^^

Hi Guys,

We (Caviar-Hacker-Grid) are your assigned team for the next week.
What would the best way be to contact you if we have any questions?