Git log fetches too much data at once

[wcho]

Big projects have very long git log data. Currently webida git plugin reads all the git log history at once and show it in the git log dialog. It may reduce the number of exec calls but for big projects the data size is too large and server could be crashed by buffer limit. At least it should be streamed and eventually git plugin should load portions of data that actually needed.

[gotchazipc]

Streaming result of git command requires heavy re-writing of implementation of 'exec' in server and client. Server should handle stdout / err to pipe command results to some logical channels (maybe each channel would be bound to a command) on web socket. Client should listen web-socket events on the channels. If there are some proxy or sharding mechanisms, complexity will grow further. And, more serious problem is, (at least) the server code need enermous refactoring and documentation to implement some shiny features.

Until then, we can use --skip & --max-count.