The command for unmounting the kernel vfs is now
umount -Rv $LFS
So we have first to use that command in teardown (change in
kernfs.xsl). Second, since the command tries to umount
$LFS, but the Makefile is on $LFS, it fails. Ignore the
error in the Makefile (change in master.sh)
- major number can be on 4 digits, so use it in packInstall
- minor number (for package manager) can be greater than 10
- code on 2 digits
- change various files accordingly
- Prefer SysV over Sys V in Config.in
- do not generate the package list for lfs, it is not needed anymore
- prevent an (ignored) error in LFS/master.sh
- log new files under the scriptlet name (allows to log files for the same
package from different builds)
- keep the full name in log_new_file functions (fix a undetected bug in
CLFS/HLFS)
If we stop a build before the "do-housekeeping" target is executed, there
may be symlinks of the type /tools->/mnt/lfs/tools. If we want to restart,
for example after fixing something else, and we run the "creatingtoolsdir"
target (or similar one in CLFS) again, it will fail. It's better to clean
those symlinks when cleaning the build directory.
Most books have something like "exec /bin/bash" in .bash_profile for the
building user. This results in stopping and waiting for commands when
running a a login shell. The current jhalfs way of removing .bash_profile is
suboptimal, since a user may want to login as $(LUSER) and expect that the
environment is set (without having to source .bashrc). So remove the
"-i" flag in "sudo -u $(LUSER) ..."
In create-sbu_du.sh, we use "du -s" for getting the size of the installed
files at the end of the build, and we compare to the "KB" from the previous
package to get the Installed size of the last package. But the du command
excludes an absolute dir, while it should be relative. This lead to bogus
stats.
- Replace the menu system with the Kconfiglib, which has an ISC license
- Remove farce and any reference to it
- Rewrite the copyright notice, add the LICENSE files
- Adapt Config.in and a few other programs to the new menu system
* Various improvements from shellcheck around quoting, the use of
read, `` vs $() and other logical comparisons
* Provide a function for sourcing/loading files so we can remove a lot
of duplication. Move loading of some functions from
common/common-functions to ./jhalfs for consistency
* Provide some functions for standardized messaging
* Remove other duplication in the code, especially checking VERBOSITY
* Standardize indentation. Two spaces seemed the most prevalent so I
went with that.
Patch by Jeremy Huntwork, with slight modifications.
- add sect1info sections to packagemanager.xml.xxx + various fixes
- add templates in LFS/lfs.xsl for script start and script end:
- script start: define variables containing package information
print disk usage
unpack and change dir
- script end: print disk usage
remove build dir
- remove the corresponding operations from master.sh
- add variables LOCAL (y is hw clock set to local time) and LOG_LEVEL (sysv)
- generate /etc/sysconfig/console (sysv) or /etc/vconsole.conf (systemd)
- generate /etc/adjtime (only systemd)
- remove UNICODE and FONTMAP variables (jhalfs makes unicode default)
with LFS pacakges. But packages installed only in tools should not be in
this database. As a side note, it would be cleaner to use an xsl stylesheet
for that.
a tag in git, and not checking "Rebuild files": git pull fails because
the local repo is in "detached head" state.
- fix an issue with iputils patch in clfs-3.0.0
Add instructions to copy to SRC_ARCHIVE in func_install_blfs, so that files
are kept in host repo, and not downloaded again, should a new build be
started.
when the user fstab and/or kernel config are the same as the ones in
$BUILDDIR/sources: avoid to cp on itself
- set SRC_ARCHIVE=/dev/null in BLFS tools download script if SRC_ARCHIVE is
empty
- Remove the download part of the book instruction for cacerts
- Separate download and md5sum instructions in "download-script" generation,
so that packages with no md5sum can still be downloaded.
- the tzdata DESTDIR should not be tzdata, because some package managers use
$PACKAGE as their working directory, and this may conflicts with this one.
tzdata DESTDIR is now 001-tzdata
- when reading md5sums, the current code assumes that there is only one
<para><literal> node reachable from ../.. (relative to <ulink>). This is true
for the "official" book, but not necessarily for the snippets written for
package management. The following change only assumes that the md5sum
<para><literal> for a given <ulink> is the first one after <para><ulink>,
which does not change anything for the book, and is better for user
written snippets.
