On 07/05/2011 08:12 PM, Scott Schaefer wrote:
On 07/05/2011 05:49 AM, Holger Levsen wrote:
Hi Scott,
On Montag, 4. Juli 2011, Scott Schaefer wrote:
awesome than you already are, you could include debian/changelog
entries
in your patches too :-)
Um ... OK. I will have to learn how to do that. I will add to my
"Learn About ..." list :-)
sudo apt-get install devscripts
man dch
dch
:-)
Ah ... That will get us into long discussion about host of topics
concerning development workflows, packaging/release workflows, patch
approvals, testing, policies, reposibilities, etc., etc. For now, I
would like to continue to rely on you to update the changelog.
it also seems to contain unwanted noise about force-confdef :/
I didn't see that in my (limited) testing. I will re-test (after mod
described below) and look specifically for anything about confdef. In
meantime, do you recall precisely what it reports ?
You misunderstood me: it's in the patch you posted, not in the output
of the
code running...
Hm ......
This is against svn r907 ...
$ svn log|head
------------------------------------------------------------------------
r907 | holger | 2011-06-27 03:17:41 -0400 (Mon, 27 Jun 2011) | 1 line
piuparts-report.py: correct a typo from the patch for #523950.
------------------------------------------------------------------------
And piuparts.py:
$ cksum piuparts.py
1279739423 86368 piuparts.py
Applies with fuzz, since patch is against version with
"dpkg_divert.patch" already applied
$ patch -p1 < .hg/patches/apt_proxy.patch
patching file piuparts.py
Hunk #1 succeeded at 677 (offset -5 lines).
Applies cleanly with "dpkg_divert.patch" popped
$ hg qpop
popping apt_proxy.patch
$ patch -p1 < .hg/patches/apt_proxy.patch
patching file piuparts.py
I have run result with both HTTP_PROXY set and unset, with both
--dpkg-force-confdef and without.
I will test against the piuparts.conf I have from piatti for
master/slave testing.
Or perhaps I'll just get the revision for apt-config coded, and test
the revised one :-)
I already found 'apt-config' (which addresses all three of my
"issues"). I have it on my list to execute it and pull any
Acquire::Proxy value from there if HTTP_PROXY is not in os.environ. So
I'll send revised patch in next 3-4 days.
Cool!
cheers,
Holger
_______________________________________________
Piuparts-devel mailing list
piuparts-de...@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/piuparts-devel
Revised patch attached. Should apply cleanly against SVN r 907.
Uses env HTTP_PROXY in preference to apt_config value, if present.
diff --git a/piuparts.py b/piuparts.py
--- a/piuparts.py
+++ b/piuparts.py
@@ -677,14 +677,32 @@
def create_apt_conf(self):
"""Create /etc/apt/apt.conf inside the chroot."""
+ lines = [
+ 'APT::Get::Assume-Yes "yes";\n',
+ 'APT::Install-Recommends "0";\n',
+ 'APT::Install-Suggests "0";\n',
+ ]
+ lines.append('APT::Get::AllowUnauthenticated "%s";\n' % settings.apt_unauthenticated)
+ if "HTTP_PROXY" in os.environ:
+ proxy = os.environ["HTTP_PROXY"]
+ else:
+ proxy = None;
+ pat = re.compile(r"^Acquire::http::Proxy\s+\"([^\"]+)\"", re.I);
+ p = subprocess.Popen(["apt-config", "dump"],
+ stdout=subprocess.PIPE)
+ stdout, _ = p.communicate()
+ if stdout:
+ for line in stdout.split("\n"):
+ m = re.match(pat, line)
+ if proxy is None and m:
+ proxy = m.group(1)
+ if proxy:
+ lines.append('Acquire::http::Proxy "%s";\n' % proxy)
+ if settings.dpkg_force_confdef:
+ lines.append('Dpkg::Options {"--force-confdef";};\n')
+
create_file(self.relative("etc/apt/apt.conf"),
- 'APT::Get::AllowUnauthenticated "%s";\n' % settings.apt_unauthenticated +
- 'APT::Get::Assume-Yes "yes";\n' +
- 'APT::Install-Recommends "0";\n' +
- 'APT::Install-Suggests "0";\n')
- if settings.dpkg_force_confdef:
- append_to_file(self.relative("etc/apt/apt.conf"),
- 'Dpkg::Options {"--force-confdef";};\n')
+ "".join(lines))
def create_dpkg_conf(self):
"""Create /etc/dpkg/dpkg.cfg.d/piuparts inside the chroot."""
