On 2024-09-01, Elie Le Vaillant <eolien55@disroot.org> wrote:
> On Fri Aug 30, 2024 at 11:56 AM CEST, Stuart Henderson wrote:
>> imho you should really be looking for a 64 bit machine if you want to
>> run a web browser.
>
> I am aware that this machine is simply unsufficient for web usage. My
> personal computer needs are quite small. Base, RSS, groff, mpv, and
> links in graphical mode are enough to fulfill my needs.
>
> Still, it works surprisingly well considering that it is a kinda old
> beast.
>
>> pandoc depends on ghc, which is amd64 only.
>
> My bad, I operated based on the lists on posts.to but it is
> apparently outdated.
>
>> When attempting to build chromium and friends on i386, it usually fails
>> for various reasons.
>>
>> In the most recent attempt, iridium and ungoogled-chromium fail with this
>>
>> In file included from ../../v8/src/compiler/turboshaft/int64-lowering-phase.cc:9:
>> ../../v8/src/compiler/turboshaft/int64-lowering-reducer.h:295:24: error: call to member function 'Word32Constant' is a
>> mbiguous
>> 295 | new_index = __ Word32Constant(sizeof(int32_t));
>> | ~~~^~~~~~~~~~~~~~
>
> How did you get these log messages? I was kinda searching for more
I tried to build those ports on i386 and that was the result.
> specific infos on the package building process, but I found nothing in
> the FAQ nor the man pages (maybe I simply searched badly).
Various developers run package builds for various archs, using dpb to
do a full build from scratch each time.
>> Many projects are simply not interested in resource requirements during
>> compilation.
>
> Yes, that's understandable.
>
> Is there a reason why packages are not built on one powerful amd64 server,
> besides the hell of cross-compilation?
OpenBSD policy is to do native builds. It's a good way to make sure that
the OS is exercised on each arch. (And there's no infrastructure in ports
to handle cross compiles).
> Thanks for the valuable information! Are i386 computers more
> power-hungry than amd64 ones? I'd be interested if there's an
> explanation beyond "old hardware was less efficient with power".
> Is it the HDD, the motherboard, the CPU, everything?
Old hw was less efficient, it's everything really.
> Semi-related, but:
>
> I actually have an amd64 laptop, an HP-dbxxxx. I can install OpenBSD on
> it, but when I boot with bsd.mp, it crashes because of a
> malformed/unknown ACPI command. I'll try to look at it more closely and
> send mails with the full information, but I just have one question: how
> does one dump kernel crashes in files, so I can actually send emails?
>
> Thank you for your quite thorough answer!
If you can boot with the non-mp kernel, use sendbug (run as root) to
generate a report which includes the acpi tables (often done by using
sendbug -P and copy to another machine to include in an email if you
don't have email setup on the first machine). If not, if linux runs you
can run acpidump -b from there and those files should still be useful.
Copying down the text from the crash and including that too would be
useful (if there's working serial console you can copy and paste,
otherwise it can just be copied down and retyped).
You can also try looking for firmware updates to see if the manufacturer
fixed anything on their side. IIRC HP fairly often has ACPI tables which
are a bit difficult to support though.
--
Please keep replies on the mailing list.
No comments:
Post a Comment