[Bug 1694697] Re: build-depends keeps OR flag if end of or group is ignored

[Julian Andres Klode]

Fixed in upstream release 1.4.5. Now we only need to wait for it to sync
:)

** Description changed:

  [Impact]
  If the last alternative(s) of an Or group is ignored, because it does
  not match an architecture list, we would end up keeping the or flag,
  effectively making the next AND an OR.
  
  For example, when parsing (on amd64):
  
      debhelper (>= 9), libnacl-dev [amd64] | libnacl-dev [i386]
   => debhelper (>= 9), libnacl-dev |
  
  Which can cause python-apt to crash.
  
  Even worse:
  
       debhelper (>= 9), libnacl-dev [amd64] | libnacl-dev [i386], foobar
    => debhelper (>= 9), libnacl-dev [amd64] | foobar
  
  [Test case]
- TODO
+ On amd64:
+ 
+ cat > segv.dsc  << EOF
+ Format: 3.0 (native)
+ Source: foobar
+ Binary: foobar
+ Architecture: all
+ Version: 1
+ Maintainer: Joe Sixpack <joe at example.org>
+ Build-Depends: build-essential [amd64] | build-essential [fancy]
+ Standards-Version: 3.9.8
+ EOF
+ cat > failure.dsc  << EOF
+ Format: 3.0 (native)
+ Source: foobar
+ Binary: foobar
+ Architecture: all
+ Version: 1
+ Maintainer: Joe Sixpack <joe at example.org>
+ Build-Depends: build-essential [amd64] | build-essential [fancy], a-non-existing-package
+ Standards-Version: 3.9.8
+ EOF
+ 
+ 
+ (1) apt-get build-dep -s ./segv.dsc should succeed instead of crash
+ (2) apt-get build-dep -s ./failure.dsc should complain about "Depends: a-non-existing-package but it is not installable" instead of succeeding.
  
  [Regression Potential]
  apt-get build-dep and friends can now fail where they succeeded previously for packages that employ architecture-limited alternatives in their build depends, as in the second example given above, because now additional packages need to be installed (which is correct, though).
  
  [Other info]
  By setting the previous alternatives Or flag to the current Or flag
  if the current alternative is ignored, we solve the issue.

** Description changed:

  [Impact]
  If the last alternative(s) of an Or group is ignored, because it does
  not match an architecture list, we would end up keeping the or flag,
  effectively making the next AND an OR.
  
  For example, when parsing (on amd64):
  
      debhelper (>= 9), libnacl-dev [amd64] | libnacl-dev [i386]
   => debhelper (>= 9), libnacl-dev |
  
  Which can cause python-apt to crash.
  
  Even worse:
  
       debhelper (>= 9), libnacl-dev [amd64] | libnacl-dev [i386], foobar
    => debhelper (>= 9), libnacl-dev [amd64] | foobar
  
  [Test case]
  On amd64:
  
  cat > segv.dsc  << EOF
  Format: 3.0 (native)
  Source: foobar
  Binary: foobar
  Architecture: all
  Version: 1
  Maintainer: Joe Sixpack <joe at example.org>
  Build-Depends: build-essential [amd64] | build-essential [fancy]
  Standards-Version: 3.9.8
  EOF
  cat > failure.dsc  << EOF
  Format: 3.0 (native)
  Source: foobar
  Binary: foobar
  Architecture: all
  Version: 1
  Maintainer: Joe Sixpack <joe at example.org>
  Build-Depends: build-essential [amd64] | build-essential [fancy], a-non-existing-package
  Standards-Version: 3.9.8
  EOF
  
- 
  (1) apt-get build-dep -s ./segv.dsc should succeed instead of crash
  (2) apt-get build-dep -s ./failure.dsc should complain about "Depends: a-non-existing-package but it is not installable" instead of succeeding.
+ 
+ This is the same test as run by CI and autopkgtests, so if they pass the
+ tests passed. You can also run apt-get build-dep -s dq for a real life
+ example that should not segfault.
  
  [Regression Potential]
  apt-get build-dep and friends can now fail where they succeeded previously for packages that employ architecture-limited alternatives in their build depends, as in the second example given above, because now additional packages need to be installed (which is correct, though).
  
  [Other info]
  By setting the previous alternatives Or flag to the current Or flag
  if the current alternative is ignored, we solve the issue.

** Description changed:

  [Impact]
  If the last alternative(s) of an Or group is ignored, because it does
  not match an architecture list, we would end up keeping the or flag,
  effectively making the next AND an OR.
  
  For example, when parsing (on amd64):
  
      debhelper (>= 9), libnacl-dev [amd64] | libnacl-dev [i386]
   => debhelper (>= 9), libnacl-dev |
  
- Which can cause python-apt to crash.
+ Which can cause python-apt and apt-get build-dep to crash.
  
  Even worse:
  
       debhelper (>= 9), libnacl-dev [amd64] | libnacl-dev [i386], foobar
    => debhelper (>= 9), libnacl-dev [amd64] | foobar
  
  [Test case]
  On amd64:
  
  cat > segv.dsc  << EOF
  Format: 3.0 (native)
  Source: foobar
  Binary: foobar
  Architecture: all
  Version: 1
  Maintainer: Joe Sixpack <joe at example.org>
  Build-Depends: build-essential [amd64] | build-essential [fancy]
  Standards-Version: 3.9.8
  EOF
  cat > failure.dsc  << EOF
  Format: 3.0 (native)
  Source: foobar
  Binary: foobar
  Architecture: all
  Version: 1
  Maintainer: Joe Sixpack <joe at example.org>
  Build-Depends: build-essential [amd64] | build-essential [fancy], a-non-existing-package
  Standards-Version: 3.9.8
  EOF
  
  (1) apt-get build-dep -s ./segv.dsc should succeed instead of crash
  (2) apt-get build-dep -s ./failure.dsc should complain about "Depends: a-non-existing-package but it is not installable" instead of succeeding.
  
  This is the same test as run by CI and autopkgtests, so if they pass the
  tests passed. You can also run apt-get build-dep -s dq for a real life
  example that should not segfault.
  
  [Regression Potential]
  apt-get build-dep and friends can now fail where they succeeded previously for packages that employ architecture-limited alternatives in their build depends, as in the second example given above, because now additional packages need to be installed (which is correct, though).
  
  [Other info]
  By setting the previous alternatives Or flag to the current Or flag
  if the current alternative is ignored, we solve the issue.

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to apt in Ubuntu.
https://bugs.launchpad.net/bugs/1694697

Title:
  build-depends keeps OR flag if end of or group is ignored

Status in apt package in Ubuntu:
  Fix Committed

Bug description:
  [Impact]
  If the last alternative(s) of an Or group is ignored, because it does
  not match an architecture list, we would end up keeping the or flag,
  effectively making the next AND an OR.

  For example, when parsing (on amd64):

      debhelper (>= 9), libnacl-dev [amd64] | libnacl-dev [i386]
   => debhelper (>= 9), libnacl-dev |

  Which can cause python-apt and apt-get build-dep to crash.

  Even worse:

       debhelper (>= 9), libnacl-dev [amd64] | libnacl-dev [i386], foobar
    => debhelper (>= 9), libnacl-dev [amd64] | foobar

  [Test case]
  On amd64:

  cat > segv.dsc  << EOF
  Format: 3.0 (native)
  Source: foobar
  Binary: foobar
  Architecture: all
  Version: 1
  Maintainer: Joe Sixpack <joe at example.org>
  Build-Depends: build-essential [amd64] | build-essential [fancy]
  Standards-Version: 3.9.8
  EOF
  cat > failure.dsc  << EOF
  Format: 3.0 (native)
  Source: foobar
  Binary: foobar
  Architecture: all
  Version: 1
  Maintainer: Joe Sixpack <joe at example.org>
  Build-Depends: build-essential [amd64] | build-essential [fancy], a-non-existing-package
  Standards-Version: 3.9.8
  EOF

  (1) apt-get build-dep -s ./segv.dsc should succeed instead of crash
  (2) apt-get build-dep -s ./failure.dsc should complain about "Depends: a-non-existing-package but it is not installable" instead of succeeding.

  This is the same test as run by CI and autopkgtests, so if they pass
  the tests passed. You can also run apt-get build-dep -s dq for a real
  life example that should not segfault.

  [Regression Potential]
  apt-get build-dep and friends can now fail where they succeeded previously for packages that employ architecture-limited alternatives in their build depends, as in the second example given above, because now additional packages need to be installed (which is correct, though).

  [Other info]
  By setting the previous alternatives Or flag to the current Or flag
  if the current alternative is ignored, we solve the issue.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/apt/+bug/1694697/+subscriptions