Release v0.023


A quick release to fix a bug in yq that would cause a confusing error. Since yq was using YAML exclusively, but other commands were using other YAML modules, there was a possibility for one command to write output that another command could not parse.

Now this is fixed and yq uses the same formatting routines as everything else.

Full changelog below:

Continue reading Release v0.023...

Release v0.020


This release adds the first initial development of ygrok, the Yertl program for parsing lines of plain text input into documents.

ygrok is perfect for parsing log files and command output. By using simple patterns or named regex captures, we can parse a line of input into a document, which we can then write to a database using ysql, or convert to CSV or JSON with yto.

Lots more development on ygrok to come!

Full changelog below:

Continue reading Release v0.020...

Release v0.021


A breaking change in this release: ysql no longer needs the word query to run a query. With the earlier removal of the write command, the query command was the only way to run a query. Instead of requiring the additional command, now we can save a little bit of typing.

This release has plenty of additions to the ygrok script, allowing it to easily parse logs from Apache HTTPD and Syslog, the output of ls -l and ps. There's also a new -l flag to allow loose matching of input lines.

This release also allows you to add your own ygrok patterns. In addition to adding your own custom patterns, you can list all the existing patterns by doing ygrok --pattern. See the ygrok documentation for more details.

With all the possibilities, ygrok patterns can now be organized into namespaces. This will hopefully make it easier to manage as we add more patterns to the Yertl core.

Full changelog below:

Continue reading Release v0.021...

Release v0.019


This release fixes a few bugs and inconsistencies in the ysql command, in anticipation of a blog post explaining how it works.

We removed the "write" command in favor of making the "query" command accept input on STDIN and handle placeholders. So now you can run "SELECT" queries for each document as it comes in, perhaps to look up users as something else write's YAML with their ID in it.

Full changelog below:

Continue reading Release v0.019...