| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
| |
Moreover, make testMagicBytestringToDict test passing again.
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
| |
... in addition to yes/no.
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
|
|
| |
Based on:
- facility: libqb/lib/log_format.c:facilitynames
- priority: corosync/exec/logsys.c:prioritynames
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
| |
Also add a unit test targeting this issue.
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
| |
Also fix a typo in simpleconfig pseudo-grammar.
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
| |
(which is implicitly assumed and was present before a rewrite)
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
|
| |
Thanks to more tolerant "pretty-printer" and also with respect to what
target structuring will be optimal for the parser.
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Beside being much more legible, it is also (almost unnoticably) faster,
mainly due to not increasing the indentation step by step along the
return from recursion.
"previous apply_aggregation_preserving_passing_depth attempt was
unsuccessful" -- oh, well.
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
| |
(introduced in Corosync v2.3.4)
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
... otherwise original quotation is not preserved in, say,
ccs2pcscmd(-flatiron) command due to command format usage in cmd-wrap
filter (--noop=cmd-wrap would serve as a workaround before this fix).
Also add a respective unit test.
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
|
|
| |
First, it's a convention, second, it looks way better when outputting
the list on stdout (line is not interleaved with subsequent shell
prompt).
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
|
| |
This is for upcoming cluster.conf (Pacemaker) + CIB -> sequence of pcs
commands conversion.
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
| |
"pcs" should be reserved for actual pcs CLI tool commands.
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
| |
Also extend corosync.rng (target=needle format) with missing key{,file}.
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
|
| |
It is usually not used internally to allow "1 function, more roles"
anyway.
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
|
| |
Meant to be easily omissible/suppressable with "--noop=pcs2pcsfinal"
provided on command-line or an equivalent.
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
So the progress for ccs-to-pcs (in isolation) currently is:
- ccs2ccsflat
- ccs-revitalize
- ccsflat2pcsprelude
- pcsprelude2pcscompact
. TODO: service arrangement(?)
- pcscompact2pcs
. TODO: stonith-enabled=false if no fencing present
and other finalizations
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Also make distinguishing more restricted internal variants of the
formats more explicit (ccsflat, pcsprelude) -- this is known to
cause a little regression (format checking too strict to cope with
this) and the fix for that is forthcoming.
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is a preparation work for being able to pick up plugin of choice
nicely without the necessity of scanning the whole name/path space
followed with some initialization routines across all plugins even if
they won't be used for sure. Such heavy-weight approach is fine for
using clufter as CLI tool (covering potentially any available use-case),
not so much for the use as a library with presumably a priori determined
task to perform.
As generally one module can (and often is practical!) contain several
modules, we are introducing following convention to allow our proposed
filtering to have a good false-positives/convenience ratio:
module shall contain only plugins the name of which contains
the same prefix as the module's filename does, where the prefix
is assumed the part before first occurrence of '_'
('-' for stringified/decorated plugin names)
As a concrete example, formats/ccs.py module contains ccs and ccs_flat
plugins, hence when ccs_flat is to be picked with this fine granularity
and using default search path, only formats/ccs.py file will match
the prefix test and hence only ccs and ccs_flat will be prepared
for further use (one false positive in this case is a result of the
mentioned trade-off).
This commit normalizes the plugin-module allocation to follow
the convention, reflects these changes in tests, etc. and adds a note
to the HACKING doc file.
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
| |
(Thanks to Tomas Jelinek)
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
| |
self(self.PROTO, ...) -> self.PROTO(...)
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
+ reflect that on the formats' level (also bring a convention for that)
+ make plugin_registry.probe fnc more tolerant towards non-classes
as we want to differentiate between protocol on instance basis
rather than based on dedicated classes (because we want to
pragmatically maintain interchangeability with plain strings!)
+ make extra steps in protocol.py so the mentioned interchangeability
actually works
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
| |
... using already present corosync.rng
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
| |
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|
|
|
|
|
|
|
|
|
| |
which required finalizing some missing functionality:
- XML's 'etree' -> simpleconfig's 'struct' (filters/XML2simpleconfig)
- simpleconfig's 'struct' produced/pretty-printed as 'bytestring'
(formats/simpleconfig)
Signed-off-by: Jan Pokorný <jpokorny@redhat.com>
|