vcswatch reports that
this package seems to have a new changelog entry (version
2.100, distribution
UNRELEASED) and new commits
in its VCS. You should consider whether it's time to make
an upload.
Here are the relevant commit messages:
commit 5d5c7a2c08a3509a6f3f4967cf0879aa721371e4
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Tue Jun 11 00:27:14 2019 +0100
Change command list to not look for help text
commit b4409cb8447b4f49a0b3ceb4df2ed67de1acacd4
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Tue Jun 11 00:14:57 2019 +0100
help: Invoke man instead of looking for plain text files
commit 198fbe68b49af12e9c3c88e5600a312667db89b0
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Tue Jun 11 00:10:00 2019 +0100
Convert help text into manual pages
commit 181da254838c529eda0e9de8507dd78d85cb5228
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Tue Jun 11 00:19:41 2019 +0100
Create source subdirectory for manual pages
I'm going to add a lot more of them.
commit 930103b0947765b952ea80bb6b381a13080c2008
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Tue Jun 11 00:05:54 2019 +0100
kernel-wedge(1): Fix style errors
- Use bold and italic for constant text and parameters in commands
- Escape hyphens in command names
- Put date and "Kernel-wedge" in footers
commit 0f65d185685b292b395bbe6aae39579561aff2e1
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Mon Jun 10 23:00:13 2019 +0100
kernel-wedge(1): Fix title
commit 41fda774d3e439fbc256d835a4d5cc7acfa3a6d7
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Tue Jun 11 00:02:35 2019 +0100
kernel-wedge(1): Remove Joey Hess's email address
He is no longer a maintainer and surely does not wish to receive
questions about kernel-wedge.
commit cc28e0cf10901775f80fae91b93b6eceee0e61fd
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Mon Jun 10 22:45:37 2019 +0100
kernel-wedge(1): Update reference to README file
commit 55785881f503a0c1c96257a8cd6994de8cd32b8d
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Mon Jun 10 23:53:03 2019 +0100
Remove mass-build, which is no longer useful
commit 05602361669edb225b6596489da33ad217941887
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Mon Jun 10 00:35:20 2019 +0100
preprocess: Correct "excludes and excludes" in help text
This was presumably meant to be "includes and excludes". But the
inclusion (#include) and exclusion (foo -) mechanisms are not really
counterparts, so expand the wording to make that clearer.
commit f2052b95b405e7889b494c8a63b8c557d1babef2
Author: Holger Wansing <hwansing@mailbox.org>
Date: Sun Apr 21 10:13:52 2019 +0200
Add reference to bugreport
commit 1fa9f7edd7bbf290fa44bf1b07cacfb838c99343
Author: Cyril Brulebois <kibi@debian.org>
Date: Sat Apr 20 22:35:10 2019 +0200
Remove Christian Perrier from Uploaders.
With many thanks for all his contributions over the years!
commit 5a173faf40d4d43694d9a3d5f3b47961b970a090
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Tue Feb 12 19:40:04 2019 +0000
debian/tests: Add test cases for preprocess
commit f28ccd73b3180447e9ff8532423c44ddcd0ec358
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Tue Feb 12 18:12:38 2019 +0000
preprocess: Check for missing modules after processing exclusions
When filename lookup was done in copy-modules, inclusion and
subsequent exclusion of a missing module did not result in an error as
it would not be included in the output of preprocess. Some module
lists for Linux depend on that. However my earlier change
"copy-modules, preprocess: Move all module filename lookup into
preprocess" makes the inclusion an immediate error.
Instead of checking for missing modules in find_modules(), make
loadlist() update a hash of missing modules and check that it's empty
when the top-level loadlist() returns. To support this check, make
find_modules() return a dummy filename for modules that are actually
built-in.
commit fee50b0bc6b944ecbcfb35e81d4bf775db0689ce
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Tue Feb 12 17:58:48 2019 +0000
preprocess: Factor out wildcard detection from find_modules()
commit c1abf54a4cf9830276f1b88ad035a829cab3c17b
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Tue Feb 12 17:38:51 2019 +0000
preprocess: Move flag parsing out of find_modules()
commit 879d9a310f9bb3e66a437bf32f80509580317c0c
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Fri Feb 8 03:18:20 2019 +0000
preprocess: Remove support for modules with a '.o' extension
Linux 2.6+ and FreeBSD 3+ both use '.ko'.
commit 6c80f96af46138681958dd2cb8750acb94ba62fa
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Fri Feb 8 02:57:50 2019 +0000
Remove strip-modules, which has never done anything useful
strip-modules was changed to a no-op shortly after its introduction.
For Linux, at least, debug information is already stripped during the
package build process. Stripping again would remove any signatures
and possibly other important metadata.
commit 52aba5b5e0f786d87a6824e0a9e40d5cd82b1449
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Thu Feb 7 03:31:24 2019 +0000
preprocess: Treat '-' and '_' as equivalent in Linux module names
Avoid the sort of problem seen in:
https://salsa.debian.org/kernel-team/linux/merge_requests/76
commit 3827f1ee9f53540b104c592a8a2695f78d8629ed
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Thu Feb 7 02:50:13 2019 +0000
copy-modules, preprocess: Move all module filename lookup into preprocess
The interpretation of module lists is awkwardly split between the
copy-modules and preprocess commands.
My introduction of wildcards made this much worse: preprocess finds
full filenames for modules matched by a wildcard, but then it throws
away the directory parts and leaves copy-modules to do a full tree
scan to find them again.
Move all responsibilty for filename lookup, including checking for
missing modules, into the preprocess command.
commit 0f0bbc0d6412a1fe000ce9511992e43d2feacd37
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Thu Feb 7 01:22:35 2019 +0000
preprocess: Rename and invert $checkdir to $optional
commit e19326dc3b9d26218e475fb54694a6940dc3eedc
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Thu Feb 7 01:17:00 2019 +0000
preprocess: Only read the modules directory tree once
Read the modules directory at startup and make a list of all module
files. When searching for modules later, use the list instead of
reading directories again.
This actually makes things slower because it removes the optimisations
that allowed us to limit which directories we search. But it will
enable further changes that should make up for that.
commit 3672b6bcd0fa382eff9eaa368da5af9b9a999502
Author: Ben Hutchings <ben@decadent.org.uk>
Date: Thu Feb 7 00:35:24 2019 +0000
preprocess: Fix handling of wildcards with no directory part
If we find a line that has a wildcard character but no directory
part, we end up using a pattern that doesn't include the module
base directory and so will never match any of the files we find.
Linux doesn't need this case (all modules are organised into
subdirectories) and kFreeBSD hasn't yet taken advantage of wildcards.
But it should now work.
commit 4d9ac0561fbe6f7a8942255b5c395bb0958401bb
Author: Holger Wansing <hwansing@mailbox.org>
Date: Sun Jan 13 17:12:59 2019 +0100
Remove trailing whitespaces from changelog file, to fix lintian tag
![[bug history graph]](https://qa.debian.org/data/bts/graphs/k/kernel-wedge.png){kind=link}