Lire 2.1 will get released late April 2003
Status: Implemented, to be integrated
Responsible: Francis Lacoste
Prerequisites: Section , “Complete DLF Converter API”
Est. Man-hours: 120
Currently, one log file can only generate reports about one superservice. Removing that limitations (i.e. one DLF converter can generate DLF for multiple schemas from the same log file) enables us to solve the "cross-superservice" problem. A report can contain subreport from all the supported schemas (login, daemon, proxy, etc.)
See the thread Tackling the Cross-Superservice Problem for a more in-depth discussion. This also includes a discussion on why this approach is better than the previous multiple-event types DLF approach (as proposed in Proposals to LogReport).
This feature is implemented as part of the new DLF converter API, by changing the internal of the DLF conversion process from a pipe paradigm to a store paradigm The changes needed to support this in the report generation process are minimal once the DLF API is fully integrated.
Status: Started
Responsible: None
Prerequisites: Section , “Complete DLF Converter API”
Est. Man-hours: 20
Currently, the analysers that produce the derived schema and extended schema DLF data work in the report generation process. Those should be moved into a separate process between the log normalisation process and the report generation process. This would permit some optimisations in the data format and is necessary to support other report generation backends.
As part of the integration of Lire::DlfStore into lr_dlf2xml work is being done on some items which are related to separation of the analysis process. This is needed to integrate the new DLF API into lr_dlf2xml.
Status: Experimental Code in CVS
Responsible: None
Prerequisites: Section , “Separation of the Analysis Process”
Est. Man-hours: 120
An important target for Lire is to develop an SQL based reporting engine which should offer more scalable reporting and probably make it worthwhile to use the framework in an interactive development.
Arnaud Gaillard has written some proof-of-concept code to make Lire work with an SQL backend.