FS#58633 - [iputils] Remove deprecated sysfsutils dependency
Attached to Project:
Arch Linux
Opened by Leonid Isaev (lisaev) - Thursday, 17 May 2018, 10:34 GMT
Last edited by Tobias Powalowski (tpowa) - Thursday, 12 July 2018, 04:53 GMT
Opened by Leonid Isaev (lisaev) - Thursday, 17 May 2018, 10:34 GMT
Last edited by Tobias Powalowski (tpowa) - Thursday, 12 July 2018, 04:53 GMT
|
Details
Sysfsutils has been dead for over 10 years now and most
packages that depend on it have moved on to accessing /sys
directly. Here is Debian TODO list on deprecating
libsysfs:mpitt@debian.org">
https://bugs.debian.org/cgi-bin/pkgreport.cgi?tag=libsysfs-deprecation;users=mpitt@debian.org
.
[core]/iputils is one such package: pls see the attached diff (PKGBUILD.patch) against current SVN trunk that removes sysfsutils dep and also gets rid of an unused variable for make(1). Additionally, modern-day iputils (i) do not require perl-sgmls for manpages and use docbook-xsl instead (commit be0657da8842e90db446b10fbb9335d9033c54f2); and (ii) sign commits with a github key 4AEE18F83AFDEB23. Attached PKGBUILD.future takes care of these changes -- it was checked to produce a working package from commit 5f5a34c1b4d418ea68c27c00d3a0c173dd5560d6. Thanks! |
This task depends upon
Closed by Tobias Powalowski (tpowa)
Thursday, 12 July 2018, 04:53 GMT
Reason for closing: Fixed
Additional comments about closing: 2018.06.29-2
Thursday, 12 July 2018, 04:53 GMT
Reason for closing: Fixed
Additional comments about closing: 2018.06.29-2
In order to update to a newer snapshot, we'd need to switch libidn => libidn2 as well:
https://github.com/iputils/iputils/commit/665fe4d30ff8b6649454ae25e4768b2f23605332
doc moving to xml:
https://github.com/iputils/iputils/commit/be0657da8842e90db446b10fbb9335d9033c54f2
We could get rid of docbook-utils too.
removing sysfsutils:
https://github.com/iputils/iputils/commit/f2060f992541b58cb0407dc95d9e40babeae76a5
Regarding github key, I don't know. I guess the signature indicates that a commit was made using _the account_ (at github). If you believe that a successful login proves the identity of a committer, then the signature is meaningful.
PGP doesn't work that way, by definition it's silly (and Github is silly) to try signing on someone else's behalf.
The key is under Github's control, not the committer's control.
...
Also snapshot != release. There have been no new releases, bu we use snapshots anyway...