ROOL Forum: Recent posts

Sun, 12 Apr 2026 22:35:25 +0000

Not sure whether this is a joke, given that there was a cyclone coming through NZ when you posted that…

No I wasn’t aware, as I live in the UK I don’t generally check the weather forecast in New Zealand, my comment was reference to heat and summer weather.

But i hope everyone is okay.

Sun, 12 Apr 2026 22:29:52 +0000

Looking at o.CathLibCPP, does it replace o.c++lib supplied with the DDE?

Sun, 12 Apr 2026 21:53:13 +0000

Andrew so lucky to be in NZ, UK Weather is awful!!

Not sure whether this is a joke, given that there was a cyclone coming through NZ when you posted that…

Sun, 12 Apr 2026 21:41:27 +0000

What I find a bit confusing is the attitude “someone can try taking the ROD sources and making it build in ROOL’s process”. You could, you know, email ROD and ask if we’d do it (for free). If that’s all it’d take, we’d be happy to I’m sure. But instead, the (almost 20k) bounty is assigned elsewhere, and not used to either a) review and assist in merging the ROD work or b) cover any changes required to facilitate its use. I can’t honestly believe it’d cost less to do the work from scratch.

You IMHO risk being disingenuous here as you have ignored the context of the comment about taking a branch, and are potentially implying “why won’t bad nasty evil ROOL merge our code”. I’m sure that’s not the intent, but you risk feeding a narrative, and I’m sure neither of us want to antagonise the other party, nor stoke the flames of division within the small RISC OS community.

To be clear: I was responding to the complaint that ROOL were “gatekeeping RISC OS” and preventing a RISC OS source tree including ROD’s code from being published. That’s the context; we are not; just like Linux, anyone can make a distro.

When it comes to merge requests, there is a team of maintainers; to date, we have received no merge request for the ROD stack. This is likely to be in part because of technical barriers due to the deep technical differences between FreeBSD and OpenBSD. An MR would be… Exciting. It would also be vast, since it amounts to dumping a huge chunk of “completed” work into the source tree without any prior iteration or discussion. Reviewing it properly could take weeks, all done by volunteers in their spare time.

I am sure ROD must have their reasons for having chosen the BSD code base that’s binary-incompatible with RISC OS, despite you then having to go to the trouble of repeatedly adding in “shims” / glue code to try and bridge between the two worlds as bugs arise. Unfortunately, when bugs do arise on top of RISC OS 5, it’s often ROOL who get to field all the oft-weird reports and problems, even though the issue may be in ROD’s glue layer. As Steve said in his talk and as I’ve said in this thread, it doesn’t end well if people expect to be able to throw a huge chunk of “finished” code over the fence, then leave ROOL with the maintenance burden for something that was developed without any to-and-fro. Consider the amount of time it took just to respond to this example and quite small MR. It is a symptom of a much more serious and deep architectural issue.

Here’s a personal opinion: Some people have opined that it’s better to have something that works partly than does not work at all. I understand that. But that’s OK; users have the choice. They can download the ROD stack.

But instead, the (almost 20k) bounty is assigned elsewhere

You seem to be confused about the bounty process.

This is NOT the process:

• ROOL say “here’s a pot of money to achieve Thing”
• Third party at some point announces, “We have fully implemented Thing, pay us now”

This IS the process:

• ROOL say “here’s a pot of money to achieve Thing”
• A third party expresses interest
• ROOL describe the technical details, constraints, requirements
• There’s usually some to-and-fro on that
• If there’s agreement, work proceeds with regular check-ins to make sure both parties are still happy
• Upon agreed completion, it’s handed over and the community “officially” become maintainers

I would point out that the “not how it works” approach is just about true of any and all outsourced/contracted work anywhere, while the “this is the process” approach should look completely familiar and understandable.

Sun, 12 Apr 2026 21:08:34 +0000

For over 300 pages the better option than using my KPDFutil application is to use PDFutils direct.

The command you want is pdfunite so if each page was called page1/pdf page2/pdf etc and you wanted the finished pdf file called complete you would have

pdfunite page1/pdf page2/pdf page3/pdf complete/pdf

That would join the 3 pages togther.

Sun, 12 Apr 2026 20:26:32 +0000

Wait I guess SDFS is different as its treated as a separate module and not mixed with the RISCOS core. isn’t TCP/IP kind of integrated with the main RiscOS core hence why the 2 licences can’t be mixed?

Sun, 12 Apr 2026 20:19:30 +0000

I’m not sure I follow.. Are you saying there is incompatible components to the ROOL Source tree (Apache 2.0 Wise) ?

Sun, 12 Apr 2026 19:51:57 +0000

@ Andrew H.

To bring this thread back to the main topic, I found the following in the GitLab sources, which I vaguely remembered seeing before:

https://gitlab.riscosopen.org/RiscOS/Sources/FileSys/S…

So, while you are correct that the CDDL is not compatible with the Apache 2.0 licence, SDFS is distributed under the CDDL, and it is present in the source repository alongside the rest of the Apache 2.0 licensed material.

Given that, I believe this may simply be a matter of ROD pushing their changes to ROOL’s GitLab. We could then review those changes ourselves and merge them if appropriate. My understanding, based on what both Steve Revill and you have said, is that we can all act as maintainers and that ROOL is not gatekeeping. If that is correct, then it seems to me that this should be a relatively straightforward issue to resolve.

Sun, 12 Apr 2026 19:46:18 +0000

Okay I will do the frosted theme with the other configure icons, ill post those up on the website this week as a 2nd option!

Sun, 12 Apr 2026 19:08:50 +0000

Third video is up now, showing how to create the GitHub repository for your project.

https://www.youtube.com/watch?v=QqvniU_Qni8&list=P…

Sun, 12 Apr 2026 19:08:11 +0000

Thank you, I didn’t realise the STL was part of CathLibCPP, I’ll dig deeper. I had a look at ccdoc, early stages, managed to work out how to create an index but so far haven’t managed to parse any header files. I was trying to run ccdoc in a taskwindow and it kept complaining the Window Manager was in use, or something like that.

P.S. My laptop died and the doxygen config was on the drive along with my local GitHub. So I’m looking around for alternatives which I can run on my old laptop which is very full.