Instead of using two different modules to generate UUIDs and GUIDs, use
a single implementation, which we use uniformly in all of our codebase,
so that we can drop this from our dependency tree.
Switch to use the new API, which simplifies the code.
Change-Id: I37013c063f283cd27b8ffc48129b35288ce733a5
support loading the full CDR graph, featuring all relations:
-cdr_groups
-cdr_tags
-cdr_mos
-cdr_status
-cdr_relations
-cdr_presentity
-cdr_cash_balances
-cdr_time_balances
aside various fixes, NGCP::BulkProcessor::Closure is introduced,
which provides uniform code execution:
-perl code (from YAML)
-perl code (from string)
-javascript (from string)
an identical symbol table is exposed to either
language env, including the SQL connector api.
Change-Id: If10422df33d996fb6a6a6b53d0ead28ea1cef755
This should make it easier to see the application version in contrast to
the system modules version. We initialize this at build time from the
package version, and fallback to the module version if the application
has no usable version.
Change-Id: I4b4e1a926aea517b2b3b46261794e187af0c9865
Let's try to make the experience a bit smoother, by tuning a bit the
dependency relationships for the obsolete transitional package.
We add versioned Breaks, and a version constraint to the existing
Replaces, against the obsolete package, so that we force an upgrade
in case we just install the libngcp-bulkprocessor-perl package.
We add a strict versioned dependency from the obsolete package to
the new one, so that we force a matching version, in case people
try to install the package standalone.
Change-Id: I2642b4ec00ccbf9d744790c52903c84f53cc9a0b
Which is also around 300% faster:
Rate had-hoc sprintf
had-hoc 3508772/s -- -75%
sprintf 14084507/s 301% --
Change-Id: I5ebce0058473f9de08c381739646a20e42308a8f
List all used modules, remove unused ones and add missing ones. Move
some optional dependencies into suggests.
Change-Id: I66966af1fb32e97cb44c15fe9aec9cac96e36050
Sipwise Jenkins Builder
Rene Krenn
Guillem Jover