On 10/5/22 08:54, Panu Matilainen wrote:
On 10/4/22 16:56, Scott Andrews wrote:
On Tue, 4 Oct 2022 10:43:28 +0300
Panu Matilainen <[email protected]> wrote:
On 10/3/22 17:10, Scott Andrews wrote:
On Mon, 3 Oct 2022 15:35:24 +0300
Panu Matilainen <[email protected]> wrote:
On 9/21/22 19:48, Scott Andrews wrote:
%prep
%setup -q -n %{name}-%{version}-release
sed 's:/bin/sh:/usr/bin/sh:' -i macros.in
sh autogen.sh --noconfigure
%build
_options=( --prefix=/usr
--program-prefix=
--sysconfdir=/etc
--sharedstatedir=/var/lib
--localstatedir=/var
--with-crypto=libgcrypt
--with-gnu-ld
--with-archive
--with-cap
--with-acl
--without-selinux
--enable-zstd
--enable-sqlite
--enable-python
--disable-openmp
--disable-dependency-tracking
--disable-silent-rules
--disable-nls
--disable-rpath)
./configure ${_options[@]}
%{make_build}
make[3]: Entering directory
'/home/rpi/BUILD/rpm-4.18.0-release/docs/man' no -s -t man
fr/rpm.8.md -o fr/rpm.8 make[3]: no: No such file or directory
make[3]: *** [Makefile:883: fr/rpm.8] Error 127
make[3]: Leaving directory
'/home/rpi/BUILD/rpm-4.18.0-release/docs/man' make[3]: *** Waiting
for unfinished jobs.... make[3]: Entering directory
'/home/rpi/BUILD/rpm-4.18.0-release/docs/man' no -s -t man
ja/rpm.8.md -o ja/rpm.8 make[3]: no: No such file or directory
make[3]: *** [Makefile:883: ja/rpm.8] Error 127
make[3]: Leaving directory
'/home/rpi/BUILD/rpm-4.18.0-release/docs/man' make[3]: Entering
directory '/home/rpi/BUILD/rpm-4.18.0-release/docs/man' no -s -t
man ja/rpm2cpio.8.md -o ja/rpm2cpio.8 make[3]: no: No such file or
directory make[3]: *** [Makefile:883: ja/rpm2cpio.8] Error 127
make[3]: Leaving directory
'/home/rpi/BUILD/rpm-4.18.0-release/docs/man' make[3]: Entering
directory '/home/rpi/BUILD/rpm-4.18.0-release/docs/man' no -s -t
man ja/rpmbuild.8.md -o ja/rpmbuild.8 make[3]: no: No such file or
directory
These are attempts to run 'pandoc' but it's not present, which is
where the "no" comes from. But pandoc isn't needed building rpm
from the source tarballs, the manual pages are shipped in the
tarball. Unless of course the .md files have been touched.
Here's what those timestamps look on my system after unpacking the
4.18.0 tarball:
-rw-r--r--. 1 pmatilai pmatilai 34668 Jun 28 10:16
docs/man/fr/rpm.8 -rw-r--r--. 1 pmatilai pmatilai 30566 Apr 7
14:13 docs/man/fr/rpm.8.md -rw-r--r--. 1 pmatilai pmatilai 1270
Jun 28 10:16 docs/man/ja/rpm2cpio.8 -rw-r--r--. 1 pmatilai
pmatilai 658 Apr 7 14:13 docs/man/ja/rpm2cpio.8.md
-rw-r--r--. 1 pmatilai pmatilai 72339 Jun 28 10:16
docs/man/ja/rpm.8 -rw-r--r--. 1 pmatilai pmatilai 34450 Apr 7
14:13 docs/man/ja/rpm.8.md
Also, the beta1 tarball had a bug wrt this, might want to
double-check that.
- Panu -
I cloned the rpm repository and then created a archive of the
4.18.0 tag
Ah, that explains. That's not *at all* the same as the release
tarball, especially as long as we're on autotools-based build system.
When extracting the archive all the timestamps are dated as follows:
./fr:
total 40K
drwxr-xr-x 2 scott scott 4.0K Oct 3 09:55 .
drwxr-xr-x 8 scott scott 4.0K Oct 3 09:55 ..
-rw-r--r-- 1 scott scott 30K Sep 20 08:07 rpm.8.md
./ja:
total 64K
drwxr-xr-x 2 scott scott 4.0K Oct 3 09:55 .
drwxr-xr-x 8 scott scott 4.0K Oct 3 09:55 ..
-rw-r--r-- 1 scott scott 658 Sep 20 08:07 rpm2cpio.8.md
-rw-r--r-- 1 scott scott 34K Sep 20 08:07 rpm.8.md
-rw-r--r-- 1 scott scott 8.5K Sep 20 08:07 rpmbuild.8.md
-rw-r--r-- 1 scott scott 1.6K Sep 20 08:07 rpmgraph.8.md
This if from the tarball on rpm.org:
./fr:
total 76K
drwxr-xr-x 2 scott scott 4.0K Oct 3 10:02 .
drwxr-xr-x 8 scott scott 4.0K Oct 3 10:02 ..
-rw-r--r-- 1 scott scott 34K Jun 28 03:16 rpm.8
-rw-r--r-- 1 scott scott 30K Apr 7 07:13 rpm.8.md
./ja:
total 164K
drwxr-xr-x 2 scott scott 4.0K Oct 3 10:02 .
drwxr-xr-x 8 scott scott 4.0K Oct 3 10:02 ..
-rw-r--r-- 1 scott scott 1.3K Jun 28 03:16 rpm2cpio.8
-rw-r--r-- 1 scott scott 658 Apr 7 07:13 rpm2cpio.8.md
-rw-r--r-- 1 scott scott 71K Jun 28 03:16 rpm.8
-rw-r--r-- 1 scott scott 34K Apr 7 07:13 rpm.8.md
-rw-r--r-- 1 scott scott 18K Jun 28 03:16 rpmbuild.8
-rw-r--r-- 1 scott scott 8.5K Apr 7 07:13 rpmbuild.8.md
-rw-r--r-- 1 scott scott 3.7K Jun 28 03:16 rpmgraph.8
-rw-r--r-- 1 scott scott 1.6K Apr 7 07:13 rpmgraph.8.md
I am on an ARM platform and pandoc is not available
Is there a fix for building from the git version?
Not at this time. Build from the release tarball instead, there's
even a blurb about this in the GH releases page:
https://github.com/rpm-software-management/rpm/releases (I did just
now expand on the rationale a bit)
- Panu -
_______________________________________________
Rpm-maint mailing list
[email protected]
http://lists.rpm.org/mailman/listinfo/rpm-maint
_______________________________________________
Rpm-maint mailing list
[email protected]
http://lists.rpm.org/mailman/listinfo/rpm-maint
With your help I have managed to create a sed to get by this issue
In my spec file to build from the github sources
./configure ${_options[@]}
sed '/^SUBDIRS =/ s/docs//' -i Makefile
%{make_build}
It prevents the docs/man pages from being built and allows me to
use the github sources.
I am OK with out the man pages.
Ok. Pandoc being the exotic thing it is, it may well make sense to make
this an actual configure option that makes bootstrapping from git much
more lightweight.
Turns out we already have everything needed to make it conditional,
except the conditional itself:
https://github.com/rpm-software-management/rpm/commit/1b8f7a182fe917ed5af5086d715cae529540a4d3
Explicit --with/--without switches for this and doxygen would be better
but leaving that exercise for another day :)
Thanks for testing + reporting,
- Panu -
_______________________________________________
Rpm-maint mailing list
[email protected]
http://lists.rpm.org/mailman/listinfo/rpm-maint
