new input filter : nibble_swap #62
Conversation
|
(force-pushed because of some whitespace issues I just noticed - nothing else changed) |
There was a problem hiding this comment.
This is looking great @fenugrec! A few minor(I hope) comments and some small line length items from the CI lint tests.
Would you please also update doc/etc/README.man (around line 318) to add the filter
|
Thanks for reviewing. Will update this list as I go
|
|
Ok, the megalint long-line errors should be fixed. About the copyright notice : if I cannot use my handle there, I assigned it to "srecord project". If that is not satisfactory then I will copy the notice from the bit_reverse (or byte_swap) filters which I mostly duplicated except for one line. Note : there's already a |
Fantastic. Looks good.
That is probably a solid approach although people may get a little confused about files with circa 2010 copyright being created in 2023. If they are different, we could have an issue as Peter can no longer assert copyright and his estate has no legal entity. But if this is your preferred approach, I think on balance, we can go with it.
I spotted the existing one too and that was entirely my fault for letting it through. We will need to address this one similarly too unfortunately. The problem is that the GPL relies on assertion/defence of copyright which in many if not all jurisdictions can only be made by a legal entity ie a person or business. While I appreciate the intent with "srecord project", it isn't a legal entity either. I completely support your preference to keep your legal name out of the code base so the best I can offer is if you assign it to me. I don't like the idea of claiming credit for your work (history still shows it comes from you) but unless you have a corporate entity that could hold the copyright, it's the best we have available. Perhaps something like the following? Now I remember why I chose engineering over law :-) |
Good point.
Yeah, getting fast into uncharted waters for me, and of course varies according to local law, but there is definitely some precedent for asserting copyright under a pseudonym (not a new thing in the publishing world). What few references I've found all recommend consulting lawyers , obviously :
I think using "srecord project' would probably be valid in many jurisdictions, I've seen various wording like "works produced in collaboration / in a group or team", but that's another story. As I understand since I published this PR (and the hp64k one) under GPL under a pseudonym, the copyright enforceability is my problem - it says GPL and nothing else, so you are free to include it in srecord under the GPL terms. You don't inherit the responsibility of protecting my copyright, since it's mine ? Anyway, as maintainer, it's entirely up to you how to proceed; if it's a problem then for this PR at least I have no objection to you claiming copyright on it. It's trivial code that you could have produced in less time than it took me to write this reply -- I hardly even feel I deserve to have my own name there !
Yeah, no joke. The only laws I want to deal with are the laws of physics. |
Yep and we have some similar conventions here in Australia. Hence my move to keeping it simple.
Mate, you took the time to contribute and even consider my comments above seriously which I greatly appreciate. In the interests of taking a simple approach that I can apply consistently, would you mind making the updates to something like the following? or perhaps
LOL, much more reasonable set of rules! |
Right. Thought about it some more, and still, I can't see how this would ever be a problem. Even if you or some future maintainer wants to relicense everything, whether there's a real or fake name or an obvious pseudonym in the notice won't change much, as I would still need to be contacted for permission. If someone uses parts of my code and doesn't observe GPL requirements, that's not the project's problem to deal with but entirely mine. Can you think of a scenario where it would cause issues ? |
This comment was marked as outdated.
This comment was marked as outdated.
I gave it some more thought and sought some more advice. Regarding copyright assertion, a handle is essentially a pseudonym (aka pen-name) and this doesn't undermine copyright assertion as long as we can trace it back to a real identity. Your commits are attributed to an email address and possibly even signed giving traceability back to a contact point. It's an email redirect via Sourceforge but better than nothing. I may end up turning on "commits must be signed" at some point. Bottom line is that on balance, I do think we'll be ok. Let's move forward as-is on the copyright assignment. Thanks for helping me think this through. If you wouldn't mind changing the copyrights back to |
I'd be perfectly fine with that if you eventually go for that.
Glad to hear that; thank you for giving this honest thought. Will change back the notices in this PR in a few minutes. |
Simply exchanges upper and lower nibbles of each bytes. This should only be a few opcodes per byte on most CPUs; probably faster than a look-up table (e.g. like the bit-reverse filter).
|
When you get a moment to change the copyright assignments to fenugrec, I can merge this in. |
Oh I thought I alread had, did I miss one ? |
Sorry, my mistake, I mucked something up pulling latest. All good! |
|
awesome. thanks |
Simple, exchange upper and lower halves of each byte, e.g. "A6" -> "6A".
Includes a simple test script.