search

AgentD / squashfs-tools-ng

A new set of tools and libraries for working with SquashFS images
Other
194 stars 30 forks source link

Allow destroy functions to gracefully handle NULL input, add example file to show dumping a single text file #89

Closed bluca closed 2 years ago

bluca commented 2 years ago

Is it OK to add the example as MIT? It would facilitate adoption of the libsqfs library, by allowing users to just take the example and modify it, regardless of the license of their project. I can change it otherwise.

AgentD commented 2 years ago

Hi @bluca,

thanks for the patch set! The changes look OK to me, but I'm still not sure how to deal with the so versioning in the release that includes the changed destroy functions. Although it technically changes function behavior, any code that previously worked still works, it only "breaks" behavior of programs that didn't work earlier. So I guess that's more of a "backwards compatible addition". But that can be dealt with that when the next release comes around.

Is it OK to add the example as MIT? It would facilitate adoption of the libsqfs library, by allowing users to just take the example and modify it, regardless of the license of their project. I can change it otherwise.

I do think example code from manuals should be considered "copy as you please". If an explicit license is added, I think it should be one that reflects that. While the X11/MIT license is fairly nonrestrictive, it does have an attribution clause.

bluca commented 2 years ago

Hi @bluca,

thanks for the patch set! The changes look OK to me, but I'm still not sure how to deal with the so versioning in the release that includes the changed destroy functions. Although it technically changes function behavior, any code that previously worked still works, it only "breaks" behavior of programs that didn't work earlier. So I guess that's more of a "backwards compatible addition". But that can be dealt with that when the next release comes around.

Yeah given it wasn't previously allowed, all existing programs won't have issues. I guess a since version X.Y.Z can be added to the header doc near the release time? Or I can add it right now if you wish and if you know the next version number.

Is it OK to add the example as MIT? It would facilitate adoption of the libsqfs library, by allowing users to just take the example and modify it, regardless of the license of their project. I can change it otherwise.

I do think example code from manuals should be considered "copy as you please". If an explicit license is added, I think it should be one that reflects that. While the X11/MIT license is fairly nonrestrictive, it does have an attribution clause.

Yes good point - after a quick look on the OSI website, I'm going to propose 0BSD: https://opensource.org/licenses/0BSD - is that ok for you? Pushed new version as such.

Also, given technically it's a relicense operation (given the examples have no license specified, afaik they fall under the umbrella license of the project, GPL3), I'm not going to change the existing examples - you seem the sole copyright owner judging from git log (apart from a minor fix that seems small enough not to warrant an explicit ack: 4c10a491a2ddfcdc3226431a71aefd6ca71135d4 ), so to be legally sound it would be better if you changed those yourself, I think.

AgentD commented 2 years ago

Hi @bluca, thanks for the patch set! The changes look OK to me, but I'm still not sure how to deal with the so versioning in the release that includes the changed destroy functions. Although it technically changes function behavior, any code that previously worked still works, it only "breaks" behavior of programs that didn't work earlier. So I guess that's more of a "backwards compatible addition". But that can be dealt with that when the next release comes around.

Yeah given it wasn't previously allowed, all existing programs won't have issues. I guess a since version X.Y.Z can be added to the header doc near the release time? Or I can add it right now if you wish and if you know the next version number.

Yes, adding a note to the doxygen documentation is a good idea, I'll take care of it. My comment was more about the libtool library version semantics.

Is it OK to add the example as MIT? It would facilitate adoption of the libsqfs library, by allowing users to just take the example and modify it, regardless of the license of their project. I can change it otherwise.

I do think example code from manuals should be considered "copy as you please". If an explicit license is added, I think it should be one that reflects that. While the X11/MIT license is fairly nonrestrictive, it does have an attribution clause.

Yes good point - after a quick look on the OSI website, I'm going to propose 0BSD: https://opensource.org/licenses/0BSD - is that ok for you? Pushed new version as such.

Also, given technically it's a relicense operation (given the examples have no license specified, afaik they fall under the umbrella license of the project, GPL3), I'm not going to change the existing examples - you seem the sole copyright owner judging from git log (apart from a minor fix that seems small enough not to warrant an explicit ack: 4c10a49 ), so to be legally sound it would be better if you changed those yourself, I think.

Will do,

thanks!