9 comments

[ 3.3 ms ] story [ 36.6 ms ] thread
Sigh. Another web request piped into an interpreter.

Not even https.

I love the implication that it only needs the output of git log so it's totally safe to run this unknown script on your box.
The wget version still needs curl... at least it looks like the comitted data is escaped and a big chunk of /dev/zero does internal server error or a gateway timeout only. Why does "www.gitminer.com" fordward to "http://gitminer/" ?
As if that weren't bad enough, the entire tool runs server-side, requiring you to upload your data, for precisely what reason? It's not like these aren't rather boring, bog-standard visualizations.

It would be rather more interesting to have a tool that processed and dumped git history data, locally, into Postgres, sqlite or whatever with a standard DB schema. Nicely structured, that data would then be readily available for investigation with whatever viz tools you had to hand (d3 and other charting libs, R, Tableau, etc.)

Great way to harvest email addresses etc of all your committers. For identity theft, spamming and spoofing fun.
> bog-standard.

Are you British? As an American, this is one of my favorite esoteric expressions, along with the even more marginal "bog-stock".