Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.bugs.dist > #1261020 > unrolled thread

Bug#817079: autopkgtest: *3* copies for running a test?!

Back to article view | Back to linux.debian.bugs.dist

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Bug#817079: autopkgtest: *3* copies for running a test?! Paul Gevers <elbrus@debian.org> - 2025-09-11 22:10 +0200

#1261020 — Bug#817079: autopkgtest: *3* copies for running a test?!

[Multipart message — attachments visible in raw view] — view raw

Control: tags -1 confirmed

On Tue, 8 Mar 2016 01:18:59 +0100 Rene Engelhard <rene@debian.org> wrote:
> Then I run
> 
> $ adt-run -B -s -d --unbuilt-tree . --timeout-copy=10000 --- null

In confirm that running $(autopkgtest -B -d --shell . -- null) in an 
unbuild tree gets me this:

paul@toba /tmp $ diff -r autopkgtest.output.dva1_zh9/tests-tree 
autopkgtest.9IxRhQ/ubtree-./
paul@toba /tmp $ diff -r autopkgtest.output.dva1_zh9/tests-tree 
autopkgtest.9IxRhQ/build.nnc/real-tree/

The logic in runner/autopkgtest for build trees makes one copy less 
before running dpkg-source. From the documentation in the code (lacking 
in this regard) it's unclear why the ubtree (copy in the testbed) is 
kept around and not e.g. $(mv)-ed and used for the real-tree. I'm 
assuming that the code is as it is because the code to create the build 
dir is the same for all build kinds and the path is not known yet when 
the copydown happens.

The tests-tree is the copy from the testbed to the host, potentially 
after building binaries. I didn't check but I think this copy is made to 
pass it to clean testbeds after testbed resets. It's indeed a bit much 
for the null backend, but the result of common code.

Paul

[toc] | [standalone]

Back to top | Article view | linux.debian.bugs.dist

csiph-web