If you define two triggers in a SPEC that are triggered from the same source, rpm-4.6 generates invalid trigger sections in the RPM, and the rpm tool crashes when processing the resultant RPM.
Version-Release number of selected component (if applicable):
rpm-4.6.0-1.fc10.x86_64
(this bug does not affect rpm-4.4.2.3-3.fc9.i386 as with F9)
How reproducible:
Always
Steps to Reproduce:
1. Create a spec file with duplicate trigger sources (see attached)
2. Generate the rpm using F9 and F10
3. Query the rpm using F9 and F10
Actual results:
F9 behaves as expected.
F10 generates an invalid RPM, and afterwards the rpm tool crashes when trying to process it.
Expected results:
Additional info:
F9 behaves as expected: the RPM contains multiple triggers keyed to the same trigger source.
As a matter of discussion, I am not asking that either RPM version support multiple triggers with the same source -- in my case this was a bug in my SPEC file. I *would* expect, though, that if it is unsupported, that RPM would report an error or warning, like it does for duplicate %post sections.
Description of problem:
If you define two triggers in a SPEC that are triggered from the same source, rpm-4.6 generates invalid trigger sections in the RPM, and the rpm tool crashes when processing the resultant RPM.
Version-Release number of selected component (if applicable):
rpm-4.6. 0-1.fc10. x86_64
(this bug does not affect rpm-4.4. 2.3-3.fc9. i386 as with F9)
How reproducible:
Always
Steps to Reproduce:
1. Create a spec file with duplicate trigger sources (see attached)
2. Generate the rpm using F9 and F10
3. Query the rpm using F9 and F10
Actual results:
F9 behaves as expected.
F10 generates an invalid RPM, and afterwards the rpm tool crashes when trying to process it.
Expected results:
Additional info:
F9 behaves as expected: the RPM contains multiple triggers keyed to the same trigger source.
$ rpm -qp --triggers foo-0.0-0.i386.rpm
triggerin scriptlet (using /bin/sh) -- some-package
:
triggerin scriptlet (using /bin/sh) -- some-package
:
F10 generates an invalid RPM -- only the first trigger has a valid trigger source.
F10's rpm crashes (segfault) when analyzing this RPM (rpm -qp --triggers XXX).
F9's rpm can analyze both RPMs, but the RPM generated by F10 appears to be invalid (as described above)
F10 can handle the RPM generated by F9; the multiple triggers appear correctly.
$ rpm -qp --triggers foo-0.0- 0.x86_64. rpm
triggerin scriptlet (using /bin/sh) -- some-package
:
trigger: scriptlet (using /bin/sh) --
:
[note that the second trigger appears invalid]
As a matter of discussion, I am not asking that either RPM version support multiple triggers with the same source -- in my case this was a bug in my SPEC file. I *would* expect, though, that if it is unsupported, that RPM would report an error or warning, like it does for duplicate %post sections.