Harinezumi v0.07

Harinezumi, that little boot utility of the gibberish name¹, has been updated and revised, following a rather interesting use case.

Here’s the link for the attention impaired:

http://www.heyrick.co.uk/software/harinezumi/

Here’s an excerpt from the Versions log showing what’s new and YES YOU REALLY NEED TO UPDATE!

0.07  2017/07/10

      Half a decade later. OMFG.
      
      Removed serious screw up in system detection where
      if the first module was NOT "UtilityModule" (thus
      can't be a known incarnation of RISC OS), it would
      drop out of the check by STACKING instead of
      UNstacking. It's nerdy and technical, so let me
      just say that a weird sort-of-but-not RISC OS would
      have likely frozen solid as soon as Harinezumi was
      started. I guess half a decade of ZERO reports is
      useful to show that it's just a bit of paranoia on
      my part instead of anything likely to be encountered.
      
      Fixed Harinezumi so it works once again on RISC OS 5.
      Somewhere along the way, the "Repeat BootLoad" and
      "Repeat BootRun" commands ended up having "Do"
      prefixed.
      Why?
      "Do" claims to pass its argument to the command
      interpreter, rather like just... not using "Do" at
      all. Is this like the "LET" keyword in BASIC?
      Pointless syntactic sugar?
      At any rate, it broke Harinezumi, so now it's fixed.
      
      FOR THE ABOVE REASON, YOU *MUST* UPDATE YOUR COPY OF
      HARINEZUMI!
      
      An interesting use case turned up recently on the
      ROOL forums. Harinezumi can help diagnose boot
      problems, but THIS machine was running headless
      (that means no monitor or the like, but we'll happily
      go with screaming chicken analogies if you prefer)
      which gets to the interesting problem that if a
      problem *was* uncovered by Harinezumi, rebooting
      with a monitor attached would lead to the *new*
      boot being logged, not the one that failed. Why is
      this important? Because it was the lack of a monitor
      that was triggering the problem. Ding! Interesting!
      
      So what Harinezumi does now is that if the boot
      should fail, then the boot log (either the default
      $.!Boot.BootLog or a custom filename) will be
      copied with the date and time suffixed, for instance
      BootLog -> BootLog_20170710-213017
                       [ yyyymmdd hhmmss ]
      In this way, you can always refer back to see what
      went wrong.
      
      No copying will be done if everything went okay, so
      if your !Boot folder is filling up with loads of log
      files, take it as a warning that something's not
      right.

In case you didn’t spot it earlier, Harinezumi is now licenced under CDDL v1.0. If you’re bored or something, you can go pull a copy off my private server (if it doesn’t respond, try again in a day or two – I switch everything off when thunderstorms occur).

¹ Actually, it’s the Japanese word for hedgehog. So now you know. ;-)

10/10 for using ISO format Date and Time in the log file name, it is so effectively informative!

Chris: Taken to Aldershot. (^_^)

the date and time suffixed, for instance
BootLog → BootLog_20170710-213017

That’s a good idea!

“Do” claims to pass its argument to the command interpreter, rather like just… not using “Do” at all. Is this like the “LET” keyword in BASIC? Pointless syntactic sugar?

As is often the case, the command help text is a bit terse. It passes the argument string to OS_GSTrans, then calls OS_CLI on the result. So it’s a way of forcing system variables to be expanded (along with any other special character sequences), something which OS_CLI doesn’t always do.

As mentioned in the Pi 4 thread, Harinezumi doesn’t want to run on my Pi 3 either. After following the OS 5 setup instructions and rebooting, I get this:

RISC OS 960MB

Cortex-A53 Processor

Acorn SCSIFS

No keyboard present - autobooting
"Running HariKick..."
Error: Trap while in trap handler: Internal error: undefined instruction at &0001AD9C, pc = 0001726C: registers at 0001D7DC (Error number &800E00)
*

The numbers are constant across reboots. Any ideas? I’m on OS 5.28.

Edit: I see that there’s a 0.09 buried in another thread which apparently fixes this (haven’t tried it yet). But https://heyrick.eu/software/harinezumi/ says that 0.07 is the latest. Might want to tweak that :)

says that 0.07 is the latest. Might want to tweak that :)

Oh 💩, I thought I had.

I’ll fix it this evening when I get home…

I see that there’s a 0.09 buried in another thread

I thought it was a new bug, triggered by Pi4, but it seems it’s something Rick fixed a few months ago with 0.09.
That was something I hadn’t seen, then again back in May I wasn’t actually seeing anything apart from light and shade. That’s my excuse and I’m sticking to it.

I can confirm that 0.09 works as expected on both my 3 and 4, with OS 5.28 :)

Harinezumi looks great but my AOC monitor takes so long to lock on to a new signal, the RISC OS boxes are usually at the desktop by the time something appears on screen.