.ignore/.gitignore for tgrep serve

[Dan Elksnitis (dpelksnitis)]

Hi, and thanks for this tool!

I've been using it extensively in my dev workflow lately in conjunction with a workspace-root .ignore file, but noticed that "tgrep serve" does not respect the ignore rules (it just prevents them from being added to the index) - this leads to a couple of issues:

• other tools can potentially conflict on file handles with the index update (the specific error case I encountered was tgrep holding file handles to .tmp files which perforce creates for transfers during a synch)
• unnecessary output spam from the tgrep serve process - it lists all changed files, not only ones which pass the ignore filter - I've been writing the serve output to a log in the hopes of verifying it's indexing what I expect, in order to refine my ignore exclusions, which doesn't work at the moment at head rev

I've locally applied changes to make tgrep serve also respect the ignore files, which appears to be working as I expect, is this something you'd be interested in taking as a PR? If so I can spend the time to clean up and submit. It's only a handful of lines of code changed.

[Dan Elksnitis (dpelksnitis)]

Friendly bump, I'm still just holding these changes locally. I'll just put a pr together soon and hope that gets a reply I suppose.

[Dan Elksnitis (dpelksnitis)]

I've created a PR to extend the newly-added gitignore functionality to also respect non-git .ignore files, and made a couple of related improvements (piggy backing the startup file walk with the ignore file walk for performance, gate gitignore usage based on whether the folder is actually a git repo). I also submitted another PR with a fix for events silently being lost while flushes are in progress.

[Shengyu Fu (shengyfu)]

.ignore should be respected now, sorry for the delay to address this issue and thanks for your contributions.

[Dan Elksnitis (dpelksnitis)]

Thanks for taking care of it. I'm re-testing my workflows with the 1.0.2 release now, will report back if I encounter any new issues.

[Dan Elksnitis (dpelksnitis)]

Confirmed everything looks good from some preliminary testing on 1.0.2 - I'll report back if I find anything else!