[Reproducible-commits] [blog] 02/02: 69: FIXME: explain a bit why and give numbers

[Holger Levsen]

This is an automated email from the git hooks/post-receive script.

holger pushed a commit to branch master
in repository blog.

commit 909fc9d61d5d14403339e563f293f50fd5ab7caa
Author: Holger Levsen <holger at layer-acht.org>
Date:   Sat Aug 20 01:08:51 2016 +0200

    69: FIXME: explain a bit why and give numbers
---
 drafts/69.mdwn | 21 ++++++++++++++++++++-
 1 file changed, 20 insertions(+), 1 deletion(-)

diff --git a/drafts/69.mdwn b/drafts/69.mdwn
index 808a992..b4c3c53 100644
--- a/drafts/69.mdwn
+++ b/drafts/69.mdwn
@@ -35,7 +35,6 @@ diffoscope development:
 	< h01ger> |  bd21db708fe91c01ba1c9cb35b9d41a7c9b0db2b 62288 diffoscope_59_all.deb
 	< h01ger> |  366200bf2841136a4c8f8c30bdc87057d59a4cdd 20146 trydiffoscope_59_all.deb
 
-
 bugs:
 	lamby pinged 164 bugs he filed more than 90 days ago which have a patch and no maintainer reaction
 		TODO: describe number of replies, call for more replies
@@ -45,4 +44,24 @@ t.r-b.o
 		- show percentage of results in the last 24/48h (h01ger)
 		- switch python database backend to SQLAlchemy (Valerie)
 		- vary build path everywhere except when testing stretch (h01ger) - FIXME: explain why + expectations…
+some notes for writing this FIXME:
+enable build path variation for unstable and experimental
+	[00:50] <      h01ger> | so, more thinking brought me to: enable build path variation for exp. and sid, disable it for testing. besides being more logical and keeping the higher percentages (testing 
+          is more reproducible than sid…) this has another advantage: we keep the less varying diffoscope output available (for packages which vary outside buildpath) which should make analysis 
+          easier. and debian can decide whether
+[00:50] <      h01ger> |  we want to keep build pathes in .buildinfo files or ...
+[00:50] <      h01ger> | ... not… 
+[00:51] <      h01ger> | and probabyl we as debian should keep them, and we as reproducible debian project should test with build path variation… 
+[00:51] <      h01ger> | and maybe then we should do two kinds of rebuilds: with a different path than the original archive build and (only) if that fails, with the same path
+[00:53] < HW42> sounds good
+[00:53] <      h01ger> | because besides academia we also have another goal: enable users to indipendantly verify builds. and if they only can do this in /home/holger, than this is not so good, but 
+          maybe acceptable with s#/home/holger/#/tmp/buildd/#g
+
+^^ 	rewrite
+	include numbers for sid + exp
+	include date when enabling this
+unstable/amd64	24693 	21794 / 88.2%	1753 / 7.1%	972 / 3.9%	65 / .2%	95 / .3%	10 / 0%
+unstable/i386	24693 	21182 / 85.7%	2349 / 9.5%	972 / 3.9%	76 / .3%	103 / .4%	10 / 0%
+unstable/armhf	24693 	20889 / 84.6%	2050 / 8.3%	1126 / 4.5%	199 / .8%	296 / 1.1%	129 / .5%
+
 

-- 
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/reproducible/blog.git