RO 5.20 upgrade for Iyonix

Strictly speaking about the DNS rather than the Hosts file, but…

http://tools.ietf.org/html/rfc4343

Seems like making this case sensitive is simply wrong. No pros and cons about it.

Can anyone make a case for retaining case sensitivity?

Can anyone make a case for retaining case sensitivity?

No they can’t. Frank quoted the appropriate text in which the critical sentence is quite clear on the first page:

Domain Name System (DNS) names are “case insensitive”.

Case sensitivity in URL’s is a different thing. Quite amusing when the server end is done by a Unix guy and the client accessing it is a Microsoft guy who can’t understand that “File1” and “file1” are two different identities.

Quite amusing when the server end is done by a Unix guy and the client accessing it is a Microsoft guy who can’t understand that “File1” and “file1” are two different identities.

…which is one of the reasons I dislike *n*x. That and…

<rant mode="on">The horrid directory structure. The patchy support for hardware where you need to be a genius to recompile your kernel so you can have suspend/resume and WiFi working at the same time.
Don’t even talk to me about what it takes to get my printer “supported” (and how supported is that?).
Video and GPU issues abound. Oh, and since if you’re lucky enough to get some sort of support, it’ll likely be a binary blob. Good luck making that work with the random changes the kernel devs throw at you (learn the word “regression” by heart, you’ll need it – and a big box of paracetamol). Enjoy this: https://bugreports.qt-project.org/browse/QTBUG-32908 ;-)
Talking of which, has anybody in the world of Unix ever heard of the concept of backwards compatibility? I wrote a simple program with Qt2 a long time ago. I wondered about using it with Qt4 on my PVR but… what? Different API? Seriously? For the lulz, I looked briefly at the Qt5 API when it became current. Oh, my, different again. And to think how hard we try here to support ancient APIs – hell, the Wimp will still run BASIC tasks written for RISC OS 2!
Talking of incompatible – different incarnations of RPM, the apt-get method, and… that other one that involves bizarre punctuation.
Speaking of which – I need to drop to the CLI and apt-get to fetch new stuff? In 201x? Really?
I noticed the programs I have written for my PVR (I ported Brandy ;-) ) contain references to libraries with version strings embedded. Am I to assume that this program requires this version of this library? Oh, wait, anything else would be backwards compatibility, right? It is often pointed out as a “strength” that you can download sources and build software directly on your machine. That’s just covering up the appalling mess of an API, the many random tweaks and changes, and how a large binary built for one distro would probably fail on another distro… or, worse, if you “upgrade” to the latest version and find your distro doesn’t look and feel like it did half an hour ago. Oh, and where’s the rollback? How do you get to it if the kernel decides it no longer wants to be your GPU’s best friend and all you see is a wall of darkness…
I noticed when I’m fiddling with Linux, I can’t listen to videos and have system "bing"s – can’t the thing sanely mix audio sources yet?
I don’t hold out any hope for seeing my USB video capture box working. I plugged it in, Ubuntu (the orange/brown one, booted from LiveCD on a USB stick) froze (keyboard LEDs and mouse pointer too). Gee, I thought people boasted about how reliable it’s supposed to be.
There’s something of a lack of distinction between what is actually the “operating system” and what is a user space application; this makes me wonder how consistent the API is across different incarnations of Linux (Ubuntu vs Debian, for instance).>/rant>

Name resolution from the Hosts file is now case sensitive. It used not to be.
I would prefer resolution to be case insensitive. I suspect this may have tripped someone else up too.

Having gone back to August I see it’s specifically when trying to set by hostname from the interfaces dialogue in !InetSetup. This function InHostsFile was last edited in 1999, so don’t try pointing the finger at RISC OS 5.20 – I think you may be dreaming.

And for the those commentators who waded in without going back to August to read the context, yes, general name resolution via the hosts file is case insensitive (the Resolver module was last touched in July 2004).

Well it seems to have been fixed now:
https://www.riscosopen.org/viewer/revisions/logs?ident=1392067653-718648.html

And for the those commentators who waded in without going back to August to read the context, yes, general name resolution via the hosts file is case insensitive (the Resolver module was last touched in July 2004).

Did anyone say “hosts”? I could have sworn the reference was to “DNS rather than the hosts file”
As I pointed out DNS resolve – case insensitive, full URL can be case sensitive. The latter however falls in the remit of the application

Nixiyo, my Iyonix computer, has been running RISC OS 5.20 (softload) satisfactorily for a few weeks. I have attempted to flash the ROM with version 5.20 (Tungsten 5.20 stable (flash prog)); but, apart from producing the directory !Restore518, it does not upgrade Nixiyo to 5.20.

As nobody is responding to my post, perhaps I should give some details.
When I run !ROM520, the window reporting what’s happening includes the following lines (I don’t want to write all of them).

…..
Loading ROM image now …
…..
ROM image checksum good.
….
All ROM tokens found: Patching HAL NVidia STRAP data:
…..
Comparing ROMs
Creating restore file for ROM518
Restore file created

Press SPACE or click mouse to continue
_

End of report. So nothing about erasing ROM518 etc..
Info on the switcher’s menu stiil says 5.18 (16-Jan-12) after a reboot.
Nixiyo has the new !Boot, but I had already added back in some essential files for the applications that I use, testing whether everything worked as per ROOL advice.

More information on upgrade to RISC OS v5.20 problem.
I have tried again to upgrade from RISC OS v5.18 with a fresh copy of Tungsten 5.20 stable (flash prog) with the same results, but have noticed that an empty directory, IDxxxxxxxx (where x^8 is a hexadecimal number), has been created in !Scrap.ScrapDirs plus a file, FlashICP, has been created in !Scrap.ScrapDirs.ScrapDir. The file has only the text content “ROM programming started Sun,23 Feb 00:55:36”. But, no erasing or programming takes place. So the BASIC program, FlashICP, seems to get as far as line 960 and exit.
Since no one else appears to have had this problem, it looks as if there is something wrong in my setup, but I can’t see what.
Is it safe to run !ROM520 from the RAM disc as I would like to try that?