[jlp__inf]

and yess, g/PATTERN/ and /v/PATTERN/ are supzr powerful, creating branch conditions based on wether or not they (do not)match a PATTERN.

based on what you wrote:

i just add trailing \n normalization at the end

, x/(.+\n)+|\n+/ { g/./ x/\n/ c/ / v/./ c/\n/ } $ a/\n/ , x/\n+$/ c/\n/ w file_out.txt q

[piekvorst]

Yeah, sam is not universal, but it can solve complex tasks more simply by just running it a few times. The final example already normalizes \n (v/./ c/\n/) in a single pass, but we can make it even simpler by just writing:

    , x/(.+\n)+/ .,+#0-#1 x/\n/ c/ /
    , x/\n+/ c/\n/

(I haven’t tested it.)

I’m glad that sam worked out for you. You can learn more from [1] and [2]. If you need any help, you’re always welcome to ask me.

[1]: https://ratfactor.com/papers/sam_tut.pdf

[2]: https://9p.io/sources/contrib/steve/other-docs/struct-regex....

[jlp__inf]

thanks, i published an article draft about SAM inspired by your examples (referencing sources / this comment section)

Do not hesitate to pin point improvements.