.gitignore patterns beginning with '/' not treated correctly.

[ggreer]

.gitignore patterns beginning with '/' not treated correctly.

[codyps]

I should also note that as a temporary work around I've just adjusted the code to skip patterns starting with '/' so I don't miss important things.

[mutewinter]

Would love to see this get added. Right now patterns like /log/* are not being used for Ag's ignore. I'm running HEAD.

[losingkeys]

Same here (though I'm running version 0.14). The pattern that isn't getting ignored for me is /log/*.log (for reference).

[raimo]

Just ran into this problem also. Using patterns that begin with '/' is the best practice in large projects if the exact directory structure is known, so this really is a crucial feature.

My specific problem is having this in .gitignore:

/tmp/performance

Running ag blabla still returns matches from file tmp/performance/PERF-REPORT-234234234.html

[MSch]

Just ran into this problem too. At first I thought that ag doesn't deal with gitignores in subfolders. Having support for this is IMO crucial, because look at for example the default gitignore of Rails projects:

# Ignore bundler config.
/.bundle

# Ignore the default SQLite database.
/db/*.sqlite3
/db/*.sqlite3-journal

# Ignore all logfiles and tempfiles.
/log/*.log
/tmp

[smathy]

Umm, 4148d6a seems to have reversed the proper handling of the leading / in .gitignore. Reverting that commit makes ag do the right thing for me.

[ggreer]

Actually, reverting that commit will make ag treat slashes incorrectly. Instead of just ignoring /tmp, it will ignore all files and directories named tmp.

It's safer to show extra matches than to silently ignore a file that the user expects to be searched.

[smathy]

Ahh, I see. So we really need something that prepends a . for leading /s

[bitboxer]

I would love to see this fixed 😸

[rik]

This seems to be closed with #355

[ggreer]

That is correct.

[bitboxer]

👍 Thanks for fixing this! ❤️

[smathy]

Thanks 👌