Eric Pruitt

I'm attaching a patch that's a updated version of this PR for anyone that builds jq themselves. Thanks, @haguenau, for the original code. [jq-src-JQ_COLORS-field-color.patch.txt](https://github.com/stedolan/jq/files/4111565/jq-src-JQ_COLORS-field-color.patch.txt) EDIT: I noticed that in cf615152e192729541db6c06f0c009e2898b1c4e,...

Depends on what you mean by abandoned. I'm still using this patch, and the last time I synchronized my copy of jq with the upstream repo (December), the patch still...

I'm going to give implementing this a shot. Per [my comment on Kim Vandry's repository]( https://github.com/vandry/mairix/issues/1#issuecomment-353749362), I'm going to create a create a new table that will include all tokens...

My estimate was off; the database with the re-indexed messages takes up ~220 MB (vs ~95 MB expected) with my changes. Aside from that, the preliminary code seems to work,...

Normalizing the header terms to cut down on the size of the database is going to be a fairly involved task, so I'm not going to try to do that...

> Then the o: flag could be used without specifying a particular header, which might often be helpful if the data being searched for could occur in several headers. The...

> Seeing the size increase and wondering - wouldn't "naive full text indexing" of the headers result in smaller database size? Potentially, but I have no intention of spending any...

I'll open a PR later today with my latest version of the patch.

I've opened https://github.com/jqlang/jq/pull/2638.

This is strange. I'm guessing something in GAWK changed since the error appears with older versions of this code. Until I can figure this out, deleting those lines won't impact...