On Wed, Jan 02, 2019 at 12:25:18PM +0100, Thomas Seeling wrote:
> Hallo,
>
> > I assume those first and last were for packages that aren't in BLFS ?
>
> there were also packages from LFS and BLFS which showed that.
> Imho it's not directly stemming from changes in glibc but in stricter gcc
> checking, more warnings enabled and less assuming what the programmer wanted
> to express with his code.
>
If the current instructions in svn, or the instructions in the
(pre-)release of either book *fail* for those reasons, it is prima
facie a bug in that book. But added warnings don't count, although
they will need to be fixed upstream at some point.
And I don't think there were many failures like that. OTOH, not
updating the versions of packages used in a previous release
definitely will cause such failures. Can you point to any version
of a package in an LFS release that had such problems at the time
(so, nowadays, 8.3 or later) ? If so, we need to review our process.
Similarly, for packages mentioned in the book, but without versions
(e.g. strace, and the dependent perl modules in old BLFS releases,
updating to the current version is usually recommended[¹]. Typically,
I get one or two compile failures each year in strace because
something changed in the latest kernel headers. Fortunately, a new
version is usually available (I only check non-book package versions
irregularly).
> My previous daily driver LFS build was 7.7, and I had one test machine with
> 7.10. Then I started again with LFS 8.0 to get my hands on a gcc 7+ to play
> around with meltdown, spectre etc. fixes. That's where I noticed gcc had
> become much stricter in code verification.
>
LOL, Meltdown was unpleasant and only LFS-8.2+ (plus latest kernels
from the maintained 4.14 or 4.19 series - or current 4.20) will now
likely be good enough (i.e. gcc needs to support retpolines).
Fortunately, for many of us (own hardware, single human user, or
AMD CPU) Meltdown is not disastrous. But Spectre (on desktops) could
be if the workarounds in applications are not up to date. The fun
continues :-(
1. Openssl-1.1.1 excepted for perl modules, of course ,although the
problems do not seem to often mainfest themselves in real usage,
only in the testsuites of perl modules. Again, change is fun.
ĸen
--
The Laird o’Phelps spent Hogmanay declaring he was sober,
Counted his feet to prove the fact and found he had one foot over.
-- Louis MacNeice, Bagpipe Music
--
http://lists.linuxfromscratch.org/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
Do not top post on this list.
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?
http://en.wikipedia.org/wiki/Posting_style
