[tytso at mit.edu: Re: And in 2019? Re: -flto to become more of a routine - any change in opinion since 2011?]

Holger Levsen holger at layer-acht.org
Mon Jul 29 04:09:45 BST 2019


FYI,

(I guess we know already :)

----- Forwarded message from "Theodore Y. Ts'o" <tytso at mit.edu> -----

Date: Sun, 28 Jul 2019 19:44:34 -0400
From: "Theodore Y. Ts'o" <tytso at mit.edu>
To: Steffen Möller <steffen_moeller at gmx.de>
Cc: debian-devel at lists.debian.org
Subject: Re: And in 2019? Re: -flto to become more of a routine - any change in opinion since 2011?
Message-ID: <20190728234434.GA4379 at mit.edu>
List-Id: <debian-devel.lists.debian.org>

On Wed, Jul 24, 2019 at 06:03:21PM +0200, Steffen Möller wrote:
> Hello,
> 
> We just had SuSE embracing LTO
> (https://www.linuxtoday.com/infrastructure/opensuse-enables-lto-by-default-for-tumbleweed-smaller-faster-binaries.html).
> I am not sure about the progress on issues summarised in
> http://blog.regehr.org/archives/1180 that Ian pointed to. But since I
> last asked in 2016 we have more pedantic compiler settings and more CI -
> and LTO, as much as compilers have improved on that, does not need to be
> applied everywhere. Any change in opinion?

I'm currently compiling e2fsprogs with LTO for Debian --- and I'm
seriously considering ditching that change.  The reason why is because
LTO breaks reproducible builds, and so it makes it harder when I'm
verifying whether a particular packaging change (say, moving to a new
debhelper compat level) is going to make any changes to the binary ---
because using LTO pretty much guarantees that it will.

Yeah, the binaries are a little bit smaller, and presumably a little
bit more CPU efficient, but 99% of the time, e2fsprogs binary are I/O
bound, not CPU bound, and the fact that my package builds aren't
reproducible is !@#?! annoying.

						- Ted


----- End forwarded message -----

-- 
tschau,
	Holger

-------------------------------------------------------------------------------
               holger@(debian|reproducible-builds|layer-acht).org
       PGP fingerprint: B8BF 5413 7B09 D35C F026 FE9D 091A B856 069A AA1C
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://alioth-lists.debian.net/pipermail/reproducible-builds/attachments/20190729/bb63cda9/attachment.sig>


More information about the Reproducible-builds mailing list