[sagemath] 01/01: Add a comment explaining why we don't use `make ptestlong`

[Ximin Luo]

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

infinity0 pushed a commit to branch master
in repository sagemath.

commit 176ceda66653f68b04544a6e2dd942c5576278a3
Author: Ximin Luo <infinity0 at debian.org>
Date:   Sat Aug 13 15:32:35 2016 +0200

    Add a comment explaining why we don't use `make ptestlong`
---
 debian/rules | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/debian/rules b/debian/rules
index ae95009..00c18b9 100755
--- a/debian/rules
+++ b/debian/rules
@@ -41,6 +41,14 @@ prune:
 
 ptestlong:
 	cd sage && local/bin/sage-starts
+	# We do the below instead of `$(MAKE) --directory=sage ptestlong` because
+	# it has a bug where it can't detect a previous successful build. So it
+	# will try to rebuild some sagelib files, then the docbuild, then those
+	# files again, even after a "make all" that was previously successful. This
+	# is not incorrect, but is a massive waste of time. OTOH debhelper is able
+	# to detect a previous successful (override_)dh_auto_build, so if we run
+	# `debian/rules build` twice it will go directly to (override_)dh_auto_test
+	# (and then this target) without trying (override_)dh_auto_build again.
 	cd sage && ./sage -t -p --all --long --logfile=logs/ptestlong.log
 
 reset: clean

-- 
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/debian-science/packages/sagemath.git