Non-free files in z88dk
Non-free files in z88dk
Hello z88dk developers.
Fedora noticed (via Debian) that the z88dk codebase contains several non-free files (because of commercial restrictions).
Specifically:
libsrc/oz/ozinterrupt/ozcustomisr.asm
test/machine/Z80/CodesCB.h
test/machine/Z80/CodesED.h
test/machine/Z80/Codes.h
test/machine/Z80/CodesXCB.h
test/machine/Z80/CodesXX.h
test/machine/Z80/ConDebug.c
test/machine/Z80/Debug.c
test/machine/Z80/Tables.h
test/machine/Z80/Z80.c
test/machine/Z80/Z80.h
It seems like these files (or functions, in the case of ozcustomisr.asm) were copied from other places, and are not code that you could simply change the license on. In the case of the test files, the author seems pretty strongly opposed to commercial use of his code, so I'm not sure relicensing is possible there.
Either way, we need for these files to either be relicensed, rewritten to resolve the license issue, or removed. Hopefully, this is possible, otherwise, Fedora will have to remove z88dk from its repositories (Debian will probably move it to their "non-free" repository).
Debian Bug Report: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=709520
Fedora Bug Report: https://bugzilla.redhat.com/show_bug.cgi?id=967408
gNUSense Bug Report: https://savannah.nongnu.org/bugs/?func= ... m_id=39054
Hopefully this is something you can quickly resolve! Thanks,
Tom Callaway
Fedora Legal
Fedora noticed (via Debian) that the z88dk codebase contains several non-free files (because of commercial restrictions).
Specifically:
libsrc/oz/ozinterrupt/ozcustomisr.asm
test/machine/Z80/CodesCB.h
test/machine/Z80/CodesED.h
test/machine/Z80/Codes.h
test/machine/Z80/CodesXCB.h
test/machine/Z80/CodesXX.h
test/machine/Z80/ConDebug.c
test/machine/Z80/Debug.c
test/machine/Z80/Tables.h
test/machine/Z80/Z80.c
test/machine/Z80/Z80.h
It seems like these files (or functions, in the case of ozcustomisr.asm) were copied from other places, and are not code that you could simply change the license on. In the case of the test files, the author seems pretty strongly opposed to commercial use of his code, so I'm not sure relicensing is possible there.
Either way, we need for these files to either be relicensed, rewritten to resolve the license issue, or removed. Hopefully, this is possible, otherwise, Fedora will have to remove z88dk from its repositories (Debian will probably move it to their "non-free" repository).
Debian Bug Report: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=709520
Fedora Bug Report: https://bugzilla.redhat.com/show_bug.cgi?id=967408
gNUSense Bug Report: https://savannah.nongnu.org/bugs/?func= ... m_id=39054
Hopefully this is something you can quickly resolve! Thanks,
Tom Callaway
Fedora Legal
The claims that we received were on behalf of a concerned Fedora user, not from the upstream.
Since we distribute z88dk as part of Fedora, we would be a liable party if the copyright holders approached us for infringement of their license.
Also, with ozcustomisr, it looks like the keyboard driver is the part under the separate license.
Since we distribute z88dk as part of Fedora, we would be a liable party if the copyright holders approached us for infringement of their license.
Also, with ozcustomisr, it looks like the keyboard driver is the part under the separate license.
Found it:
Date: Sun, 7 Oct 2007 14:16:47 -0400 (EDT)
From: Marat Fayzullin <marat@xxxx>
Thanks! I will make a link to your page from the EMUL8 page.
On Sun, 7 Oct 2007, Dominic Morris wrote:
> Hi Marat,
>
> As requested in your licence terms, this is a quick note to let you know that
> I'm using your Z80 core within z88dk (www.z88dk.org) as the basis of a
> testing framework.
>
> Many thanks for making it so easy to use the core.
>
> cheers,
>
> d.
I'll update the makefile to download the source and apply the rabbit patches. In the meantime, just zap since it's not integral.
Date: Sun, 7 Oct 2007 14:16:47 -0400 (EDT)
From: Marat Fayzullin <marat@xxxx>
Thanks! I will make a link to your page from the EMUL8 page.
On Sun, 7 Oct 2007, Dominic Morris wrote:
> Hi Marat,
>
> As requested in your licence terms, this is a quick note to let you know that
> I'm using your Z80 core within z88dk (www.z88dk.org) as the basis of a
> testing framework.
>
> Many thanks for making it so easy to use the core.
>
> cheers,
>
> d.
I'll update the makefile to download the source and apply the rabbit patches. In the meantime, just zap since it's not integral.
There is "happy" and there is "under a free license". Nothing in that email correspondence indicates a license change, and Marat's license is very clear that it may not be used commercially without explicit permission from Marat (and it strongly implies on the website that people need to pay for that right).
But you're right, we could pull the test harness. Its less than ideal, but eh. The ozcustomisr.asm looks like a bigger concern.
But you're right, we could pull the test harness. Its less than ideal, but eh. The ozcustomisr.asm looks like a bigger concern.
I wasn't suggesting that there was a licence change, more that we've not breached any licences by including them in our project (which was my initial reading of things).
I don't really want to get into the usual wanky war over licences , but for our concerns the licence that those test files were shipped under is free enough and after a brief check (and delete/revert cycle) we've not broken any conditions there.
However, I've added in a few makefile targets to allow the test to work in a GPL compatible clean room (i.e. the source isn't distributed but downloaded and patched), so you could add that to the RPM build patch.
The OZ keyboard driver is pretty horrendous in terms of code so unless Stefano can contact the original author (the OZ community seems to have pretty much died in the past 10 years or so) it's an obvious target for a rewrite. However, that will take time so it might be prudent to disable the OZ target in a minimally intrusive way by:
1. Removing that file
2. Removing lib/config/oz700.*
3. Stopping the oz700_clib.lib target from building in libsrc
I don't really want to get into the usual wanky war over licences , but for our concerns the licence that those test files were shipped under is free enough and after a brief check (and delete/revert cycle) we've not broken any conditions there.
However, I've added in a few makefile targets to allow the test to work in a GPL compatible clean room (i.e. the source isn't distributed but downloaded and patched), so you could add that to the RPM build patch.
The OZ keyboard driver is pretty horrendous in terms of code so unless Stefano can contact the original author (the OZ community seems to have pretty much died in the past 10 years or so) it's an obvious target for a rewrite. However, that will take time so it might be prudent to disable the OZ target in a minimally intrusive way by:
1. Removing that file
2. Removing lib/config/oz700.*
3. Stopping the oz700_clib.lib target from building in libsrc
The OZ "wizard" sources come from the "oz development kit". The "OZ custom isr" was a key element for that kit which permitted a smart way to handle that target's memory pages, but since I never got to make it run correctly I already got rid of most of the original components. In the meanwhile Alexander helped me actively to try to solve the problem; it is now a bit difficult to dig into 10 years old email storage to pick his authorization, but I think it can be done.
In any case as Tom noticed the problem is onto the keyboard driver; B. Green was aware of the Alexander's kit being ported to z88dk, but never contributed directly and I don't think he ever authorized that kind of license change.
Long story short, that piece of code is even very limiting and should be rewritten; in any case I'll contact Mr. Pruss ASAP and try to clarify their position about the whole library, including the keyboard driver.
In any case as Tom noticed the problem is onto the keyboard driver; B. Green was aware of the Alexander's kit being ported to z88dk, but never contributed directly and I don't think he ever authorized that kind of license change.
Long story short, that piece of code is even very limiting and should be rewritten; in any case I'll contact Mr. Pruss ASAP and try to clarify their position about the whole library, including the keyboard driver.
I'm still waiting for an updated confirmation From A. Pruss, but in the meanwhile I thought to find the latest HTC released (http://www.lh.co.nz/ozdev/software/win32/csdk/):
http://www.lh.co.nz/ozdev/software/downloads/htc.zip
For what I could see the non-commercial restriction disappeared. Dom, can you double-check ?
http://www.lh.co.nz/ozdev/software/downloads/htc.zip
For what I could see the non-commercial restriction disappeared. Dom, can you double-check ?
ozcustomisr driver removed
This will reduce the (already weak) compatibility to the original Hitech C based OZDEV kit but the z88dk style stdio driver should still be intact since it was still based on my earlier keyboard handlig code.
The ozextra.lib module is now slightly inconsistent in some power management and keyboard parts, but I'm sure we will survive (by the way, did anyone ever use this port on a real Sharp OZ device ?).
This will reduce the (already weak) compatibility to the original Hitech C based OZDEV kit but the z88dk style stdio driver should still be intact since it was still based on my earlier keyboard handlig code.
The ozextra.lib module is now slightly inconsistent in some power management and keyboard parts, but I'm sure we will survive (by the way, did anyone ever use this port on a real Sharp OZ device ?).
I'm looking through this and I see you did this and removed the sources from CVS but then re-uploaded them with the comment that it's ok to have them in the z88dk CVS. This seems contrary to the license to me which forbids distribution of the source.dom wrote:I'll update the makefile to download the source and apply the rabbit patches. In the meantime, just zap since it's not integral.
To be honest, I've scanned through some of the projects that Marat references and all the ones I looked at included the source code.
But, I really can't be bothered with this side issue anymore.
The build now depends on having wget, a working networking connection and Marat's website running and his links never changing.
But, I really can't be bothered with this side issue anymore.
The build now depends on having wget, a working networking connection and Marat's website running and his links never changing.