textar

Overview ¶

Package textar encodes a file list (key-value slice) into a human editable text file and vice versa. This is inspired by https://pkg.go.dev/golang.org/x/tools/txtar but this format can encode any content perfectly without issues. Go's txtar doesn't handle newlines and content containing txtar markers well. In textar each file is encoded via "[SEP] [NAME]\n[CONTENT]\n". SEP is two or more = signs. The first SEP can be arbitrary length, the rest must be the same length. The first line beginning with == determines the separator length. The dynamic SEP-length makes it possible to encode and decode anything perfectly. This is the main advantage over Go's txtar. Furthermore anything before the first SEP is a free form comment. Example:

Some comments here.

=== file1
file1 content.

=== file2
file2 content.
== file3
this is a textar within textar.

The separator here is === so this textar contains file1 and file2. file3 is not part of the main textar, it just shows that file2 could be a textar itself.

The Archive.Format function automatically picks a separator length that is unique and won't conflict with existing file values. Use Parse to parse it back.

See https://github.com/ypsu/textar/blob/main/example/seq.textar for a longer example. See the testdata directory of https://github.com/ypsu/pkgtrim for a more realistic example.