- Factor out the dev-tools into a stand-alone, no-installed program
- Built the dev-tools if --enable-testing is specified
- Remove the --enable-dev-tools configure option
- Allow suffix on the binary name
- Update symlinks
- Cleanup Makefile
I hadn't realised that check_file_exists() also checked that it was
a regular file, which we don't want for the couple of uses I recently
added.
This patch adds an optional arg must_be_regular_file, and defaults
it to true, preserving the original behaviour. The recent additions
have this set to false.
If the actual metadata version is 1, then the dirty bitset will not be
present. But the superblock was trying to write it because of the
overridden md version.
* [file_utils] spin-off syscall-related file operations
1. Eliminate the potential circular dependency between
persistent-data/block.h and persistent-data/file_utils.h,
if the former one wants to include the latter.
2. Avoid namespace pollution by removing the "using namespace std"
declaration in block.tcc.
3. Correct the header hierarchy: base/xml_utils.h now no longer
depends on the higher-level persistent-data/file_utils.h
* [file_utils] support block files in get_file_length()
When check_check is used with --clear-needs-check-flag option,
device needs to be opened in RW exclusive mode, but for this
existing RO exlusive open must be closed.
LVM2 will consider version 0.5.3 as a version without flag support
for cache_check since it cannot successfully pass cache_check.
TODO: when 'RO' parts detects there is nothing to 'clear' it should
completely skip RW open to avoid udev rule processing.
Template array exist in the namespace persistent_data as well as in std
of C++11. Explicitly use the one from persistent_data.
This fixes compilation bugs with CXXFLAGS=-std=gnu++11 together with
gcc 4.8.3 and boost 1.55.
Class shared_ptr exist in the namespace std for C++11 as well as in boost.
Explicitly use the one from boost in order to be compatible.
This fixes compilation bugs with CXXFLAGS=-std=gnu++11 together with
gcc 4.8.3 and boost 1.55.
We had a cycle from transaction_manager <-> space_map, and also from
the ref_counters back up to the tm.
This prevented objects being destroyed when various programs exited.
From now on we'll try and only use a shared ptr if ownership is
implied. Otherwise a reference will be used (eg, for up pointers).