Re: pkgsrc-wip: misc/DeepSeek-TUI renamed to misc/CodeWhale; editors/fresh 0.3.12

Jonathan Perkin <[email protected]> Sun, 7 Jun 2026 16:32:03 +0100
Newsgroups gmane.os.netbsd.devel.packages,gmane.os.netbsd.devel.pkgsrc.user
Message-ID <[email protected]>
* On 2026-06-07 at 15:13 BST, Alistair Crooks wrote:

>Here's where my views probably differ - Makefiles are, as Jonathan pointed
>out, template vehicles. Patches relate way more to original source trees
>than anything else. Text in DESCR files is usually taken from website or
>original READMEs. Buildlink files are templated as well. For me, not nearly
>as sacrosanct
>
>On reflection, maybe this is what everyone else is saying?

I think we're pretty close?  FWIW here is where I would draw the line:

Not OK: Anything considered "pkgsrc original" that could reasonably be 
expected to have a BSD copyright attached to it, where introduction of 
LLM code could affect that copyright, or documentation (LLM writing 
style is unreadable, at least to me). To be more explicit:

  - bootstrap/
  - doc/
  - mk/
  - pkgtools/digest
  - pkgtools/pkg_install
  - pkgtools/libnbcompat
  - security/netpgpverify
  - ..any other sources shipped under files/ that are part of the
    bootstrap path or related to pkgsrc infrastructure in any way..
  - any package files where there is considerable logic, I'm thinking mostly
    larger builtin.mk files where they are probably at the state they should
    be moved under mk/ anyway.

OK: Anything else.

  - Patches
  - Makefiles (if you really want to, personally I think it would take
    longer to prompt an LLM than to just write it yourself).
  - Using an LLM to debug issues that you then write your own patch or
    fix based on the analysis.

This is for example why I didn't publish the patch Claude wrote for me 
to fix netpgp to work with newer gnupg.  Thankfully Taylor fixed this 
independently!

-- 
Jonathan Perkin                    pkgsrc.smartos.org
Open Source Complete Cloud   www.tritondatacenter.com