Tiny Core v16.0

[Uncle Warthog]

Juanito & Rich,

I'm seeing the /proc/cmdline there as well; if I'd thought about it a little more, I'd have realized that's what it was.  I still don't see any errors in the terminal used to lauch apps though.

[Juanito]

So far I've been unable to find the problem, which occurs with x86_64 (and aarch64), but not with x86 (which has not moved to fltk-1.4).

tce-load and tce-setup are identical in x86 and x86_64 (aarch64 is still using the previous versions)
Reverting to the previous versions of tce-load and tce-setup does not help.
Reverting to the previous version of apps does not help

If I revert to the fltk-1.3 version of apps with x86_64 (and aarch64) then onboot maintenance works as expected.

Is this somehow a problem created by compiling apps against fltk-1.4?

Edit: building apps against fltk-1.3 and fltk-1.4 with x86 (and armhf), both work.

Something to do with apps and 64-bit?

[CNK]

Before we think about recompiling the kernel, you might want to ask the kernel people why the patch wasn't included?

OK, I'll ask there on the weekend. I figured I'd better test the patch myself before doing that, so I compiled the TC16 kernel with the patch applied, and sure enough it no longer reboots at start-up on my 486s. It does need >16MB of RAM now too though, since on the laptop the boot fails later when it runs out of memory (with error messages though).

Also that kernel boots on my VIA C7-M system. Maybe because I used GCC v12 to build it instead of the newer GCC Juanito used (GCC v14). I haven't tested for hangs during use (that might need rebuilding glibc with an older GCC version too?). I'm not suggesting TC16 revert to older GCC, just posting my observations.

[Rich]

Hi Juanito

... Something to do with apps and 64-bit?

No, something to do with apps and a "comparing ints of different signedness"
warning that I fixed incorrectly.

I changed this:

Code: [Select]

if ( sloc == string::npos ) {
to this:

Code: [Select]

if ( (unsigned)sloc == string::npos ) {
When it should have been this:

Code: [Select]

if ( (long unsigned)sloc == string::npos ) {
I'll push the fix to git.

[Rich]

Hi Juanito
I pushed the fix to git.

[Juanito]

tc-16.x reposted and release notes adjusted