On Fri, 29 Jan 2016 05:19:23 -0800 (PST) Sandor Akarki <s3n...@gmail.com> wrote:
> I have a little thing to do what looked pretty simple up until > today... > > There is a website on which there is a button. Push the button -> > clone a repo -> copy a local file into the cloned repo -> commit -> > push -> delete the repo. It shall be a pretty simple flow. [...] Read the git(1) manual page (run `git help git`) and look for the "ENVIRONMENT VARIABLES" section -- you can set a bunch of variables whose names start with the "GIT_COMMITTER_" prefix. I'd say the whole approach is broken. I would rather have a simple REST-y server running on localhost under some "normal" credentials and accepting "update this file" POST requests. Your website's backend would then perform that request. Variations exist: like the webserver putting files in a dedicated directory with the Git-facing service using inotify/fsnotify events on that directory to detect new files and doing its work. A cron job would be even more low tech. Or you could connect the website's backend and the service through a Redis DB instance or a special work queue software etc. -- You received this message because you are subscribed to the Google Groups "Git for human beings" group. To unsubscribe from this group and stop receiving emails from it, send an email to git-users+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
