Building RISC OS in mid-2012

I’ve just tried building RISC OS for OMAP3, for the first time in a few months. Shed loads of errors.

Does the build require a very recent Acorn toolset – more recent than was required, say, at the end of 2011?

For example, “Unexpected characters at end of line” is a consequence of ALIGN 4, 0, 10 – is the third argument a recent addition?

You’ll need to be running the latest tools (i.e. objasm 4). I can’t remember if that was the case at the end of 2011, but since the extra ALIGN parameters were an objasm 4 addition, it definitely sounds like your tools are out of date.

I’ve got the new tools. There are/were two problems:

1) Installing the tools by means of the OMAP3Dev.Library.InstallTools doesn’t overwrite older versions of tools. That’s just a minor bug in the flags settings for that particular Obey file.

2) It looks like every attempt to invoke Squash fails with:

AMU: *** File 'squash' not found ***

I haven’t chased this one down yet. There is a file SCSI::HardDisc4.$.RiscOS.OMAP3Dev.mixed.RiscOS.Library.Acorn.Squash of type App, but perhaps it should be copied somewhere else in order to work?

There is a file SCSI::HardDisc4.$.RiscOS.OMAP3Dev.mixed.RiscOS.Library.Acorn.Squash of type App, but perhaps it should be copied somewhere else in order to work?

Did you run OMAP3Dev.Prepare.!Run? That copies a few things around in order to set up the build tree, including copying the Library folder to OMAP3Dev.Library.

Did you run OMAP3Dev.Prepare.!Run?

Yes, but I didn’t run it again after I updated the tools. I’ve just done it again and set off another build. I’ll report back.

I’ve noticed several files called IntMetric0,ff that prevent building until renamed to IntMetric0

A persistent error that I get, and don’t understand, occurs when trying to build SharedRISCOS_Lib:

Cannot open syms.A_Entries1 for I/O redirection

There is only one syms directory; it contains files A_Entries1 and C_Entries, and A_Entries1 is readable after the build process exits.

Cannot open syms.A_Entries1 for I/O redirection

Ah, yes, the old “No name translation” trick…

I got OMAP3 to build late last night.