-k is mostly useless

[refi64]

By default, --known-types is on. And there's no way to turn it off! In particular, this means that there's no reason for -k, since the feature is already on. Idea: make -k be an "off" switch for it.

[gvansickle]

Hi @kirbyfan64 . I added -k for command-line compatibility with ack. It defaults to on because that's currently the only way ucg works: a file has to match it's list of known file types, or it gets ignored under the assumption that it's either binary or of no interest to the user.
I do intend at some point to add support for scanning all files which look like text by default (which is what ack and ag do now), and then that option will have some meaning.

[refi64]

@gvansickle Ah, that makes sense! Is there a way to explicitly scan file extensions that ucg doesn't know about? I would a lot with languages that aren't super-popular, so ucg ends up always showing nothing when I use it on those projects.

[gvansickle]

Hi @kirbyfan64 .

Is there a way to explicitly scan file extensions that ucg doesn't know about?

Well, there's a way to make ucg know about new extensions: the File Type Specification feature. For each new language, you add a line like this to your ~/.ucgrc file:

# Extensions .abc, .xqz, and .def are SomeNewLanguage1 source files
--type-add=SomeNewLanguage1:ext:abc,xqz,def
...[repeat as necessary]...

You do that once, and then ucg will treat files with those extensions just like they were builtins. You can have project .ucgrc files as well, if you have wildly different needs for different project directories.

Now admittedly, if you don't know the extensions (or literal names, e.g. Makefile) beforehand, that feature doesn't do you much good; like I mentioned above, when I have enough of those round tuits, I'll make "the opposite of -k" actually work.

[refi64]

@gvansickle Sweet, that works! Do you want me to close this now?

[gvansickle]

@kirbyfan64 No, I'll close it when I do the final merge.
EDIT: Sorry, I thought we were on the other issue ;-). But still, leave this open. This is the issue I'll use to add the "real opposite-of--k" functionality.