- 27 Feb, 2019 1 commit
-
-
Sean Whitton authored
-
- 15 Feb, 2019 1 commit
-
-
Sean Whitton authored
Mistakenly, the 'next' branch diverged from 'master'. This merge commit nullifies the divergent commit so that 'next' be fast-forwarded to 'master', like this: % git checkout master % git merge -s ours next % git checkout next % git merge --ff-only master
-
- 13 Feb, 2019 2 commits
-
-
Sean Whitton authored
-
- 10 Feb, 2019 1 commit
-
-
Sean Whitton authored
-
- 23 Dec, 2018 11 commits
-
-
Sean Whitton authored
-
Sean Whitton authored
-
Sean Whitton authored
-
Sean Whitton authored
-
Sean Whitton authored
-
Sean Whitton authored
Due to vendor-specific patch series change.
-
Sean Whitton authored
-
Sean Whitton authored
-
Sean Whitton authored
Wording: Sean Whitton <spwhitton@spwhitton.name> Seconded: Simon McVittie <smcv@debian.org> Seconded: gregor herrmann <gregoa@debian.org> Closes: #833401 -
Sean Whitton authored
Wording: Sean Whitton <spwhitton@spwhitton.name> Seconded: Russ Allbery <rra@debian.org> Seconded: gregor herrmann <gregoa@debian.org> Closes: #850156 -
Sean Whitton authored
-
- 08 Dec, 2018 1 commit
-
-
Sean Whitton authored
-
- 26 Nov, 2018 1 commit
-
-
Jonathan Dowland authored
Replace occurences of virtual-package-names-list.txt with virtual-package-names-list.yaml in other files that reference the older name (except changelog), e.g. Makefile and translations.
-
- 23 Nov, 2018 6 commits
-
-
Jonathan Dowland authored
Convert the plain-text virtual-package-names-list to a structured, machine-readable document, encoded in YAML. Preserve all the information in the text version of the document, namely the virtual package name, the description, the ChangeLog and whether or not the virtual-package-provider normally provides an alternative. The preamble, section headings and ChangeLog are preserved in the form of YAML comments. The structured document should be just as legible for humans as the prior version. The preamble is mildly trimmed to remove a description of the format of the text document (obsolete) and some superfluous attribution. We expand the encoding of the alternative to include the binary path(s) that is expected to be provided. Rationale: The current list cannot easily be machine-parsed. It's also likely out of date, containing virtual package names that are not in use, and missing package names that are in use. The pre-amble seems out-of-place and might better fit in the policy document text. By re-encoding in a machine-readable format, it shall be much easier to: • automate checking the status of the virtual package names herein against the names in use within the archive (e.g. from a UDD query). • check that a package which provides a virtual package name also registers paths in the alternatives system, if this document suggests it should (in e.g. lintian or some other QA tool) I envisage many future possibilities, too, such as encoding more properties about the "interface" of a program name registered as an alternative, such as command-line arguments that should (or must) be accepted, etc. The YAML schema is as follows (here in "Kwalify" syntax): map: schemaVersion: {type: text, required: True} virtualPackages: seq: - map: name: {type: str, required: True} description: {type: str, required: True} alternatives: seq: - {type: str, required: True} Future work: • simple tooling in the policy source tree to validate this file against the schema • house-keeping the list based on the names in use in-archive • lintian check for virtual package/alternatives correspondence • move the preamble out of this document and into a policy section • move the ChangeLog somewhere else • explore extensions to e.g. encode more properties about the nature of program names registered as alternatives -
Sean Whitton authored
-
Sean Whitton authored
-
Sean Whitton authored
-
Sean Whitton authored
-
Sean Whitton authored
-
- 17 Nov, 2018 6 commits
-
-
Sean Whitton authored
-
Sean Whitton authored
-
Sean Whitton authored
-
-
Sean Whitton authored
-
According to #913154, there is consensus, that `/etc/init.d/skeleton' is not suitable location for example init.d script, and actually duplicates information, provided by init-d-script(5) manpage.
-
- 11 Nov, 2018 3 commits
-
-
Sean Whitton authored
-
Sean Whitton authored
-
Sean Whitton authored
-
- 03 Nov, 2018 4 commits
-
-
Sean Whitton authored
-
Sean Whitton authored
Suggested-by:Santiago Vila <sanvila@unex.es>
-
Sean Whitton authored
-
Sean Whitton authored
-
- 02 Nov, 2018 3 commits
-
-
Sean Whitton authored
-
Sean Whitton authored
-
-