On Tue, 2021-05-25 at 16:11 +0200, Damjan Georgievski wrote:
Unfortunately it still uses the original code. Here is my test.
$ diff -U0 <(seq 900 1000) <(seq 900 1000 | sed -e "s/999/99999999/" ) --- /dev/fd/63 2021-05-25 09:57:37.954088656 -0400 +++ /dev/fd/62 2021-05-25 09:57:37.954088656 -0400 @@ -100 +100 @@ -999 +99999999
Note the '-999' part.
so your claim is that that single line added for context is copyright- able? I'm not sure judges would agree on that.
1, patch files include some original source code 2, a patch file could include any lines of source code between one and a billion. (well, I admit that a billion is a stretch) 3, There is no hard limit of how many lines of code can be considered as copyright-able in law, AFAIK. GPL's benchmark is 300 lines, but it also says 'Unless, that is, the code is specially important'. [1] 4, If you agree that any lines of source code could be copyright-able, then it boils down to if the patch file is a fair use. Everybody can have his own opinion, while it actually depends on the upstream's feeling, and the US copyright law has four factors to consider. Footnote 4, page 51 of [2]. 5, If we think the code in a patch file is categorically no copyright- able, what if somebody used a patch file you spent a lot of time on and significantly improved a software, but he didn't credit you? When we are protecting other's copyright, we are also protecting our own..... [1] https://www.gnu.org/licenses/gpl-faq.en.html#NoticeInSourceFile [2] https://www.supremecourt.gov/opinions/20pdf/18-956_d18f.pdf Best, Manhong