Re: Very long-running buildworld process

On Thu, Jun 12, 2025 at 10:34:50PM -0700, Mark Millard wrote:
> On Jun 12, 2025, at 21:04, bob prohaska <[email protected]> wrote:
> 
> > On Thu, Jun 12, 2025 at 01:28:27PM -0700, Mark Millard wrote:
> >> 
> >> So I'm confused by what you report.
> > 
> > My report was wrong. The host suffering 
> > the loop has neither gdb nor lldb 
> > installed; I must have been looking
> > at the wrong terminal window.  
> 
> Ahh.
> 
> >> This area probably needs to be figured out
> >> first and possibly have devel/gdb installed.
> > 
> > That promises to be more work than expected.
> > Since armv7 is Tier 2, I'll have to compile 
> > the port and learn to set up a package server.
> 
> Do you need to generally be building your own
> packages instead of using the pre-built ones
> that FreeBSD supplies? (Currently only updating
> about once a month, as it takes ampere3 [latest's
> aarch64/armv7 builder machine] around that long
> to cycle through all 4 types of builds for
> latest. Hopefully, pkg 2.2.0 will help with that
> once it is in place.)
> 
Apparently not. Pkg search found lldb-mi-g20231216
(is that the whole thing?) and gdb-15.1_2 . In a
previous search I found neither, again likely looking
at the wrong ssh session.

> > This particular host hasn't the disk capacity 
> > to host a ports tree.
> 
> Room to install just desired pre-built packages?
>
Yes, 
root@www:/usr/src # df
Filesystem  1K-blocks     Used    Avail Capacity  Mounted on
/dev/da0s2a  35532636 16640220 16049808    51%    /
devfs               1        0        1     0%    /dev
/dev/da0s1      51088    17348    33740    34%    /boot/msdos 

> >> Are you more familiar with lldb? gdb? I've
> >> used gdb more and not frequently.
> >> 
> > Neither one. I'll start with gdb.
> >> 
> >> For now I suggest figuring out the lldb and gdb context
> >> and then attaching one of those to the existing process.
> >> After that, figuring out how to look around.
> > 
> > Agreed, I'll begin by trying to get gdb working.
> >> 
> >> attaching will likely report on what symbols or debug
> >> information is found for the process attached to --vs.
> >> if no such is found.
> > 
> > 
> > Stopping and restarting the buildworld process seems to
> > have either delayed or avoided the loop; the machine is
> > now up to the point of:
> > 
> > ...
> > Building /usr/obj/usr/src/arm.armv7/lib/clang/libclang/Sema/SemaExpr.pico
> > Building /usr/obj/usr/src/arm.armv7/lib/clang/libclang/Sema/SemaExprCXX.pico
> > Building /usr/obj/usr/src/arm.armv7/lib/clang/libclang/Sema/SemaExprMember.pico
> > 
> > At this point it looks doubtful the loop is repeatable.
> 
> That means no gdb or lldb session for the specific
> problem. So you may not want to set up a debugger
> at this time.
 
I think it's a prudent exercise to at least have the software
in place and ready to go, in the manner of an insurance policy.

A fresh install of 14.3-RC1 just finished an inaugural meta-mode 
j4 buildworld/kernel on an armv7 Pi2 without failures. FreeBSD 
is definitely making progress.

Thanks very much for your attention and guidance!

bob prohaska