JCOP FWWG Meeting,
11 June 2019
JCOP Fw and WinCC OA news:
- Latest patch P011 for WinCC OA 3.16, as well as the latest version of CtrlRDBAccess were released and available on the downloads pages.
- The WinCC OA 3.16 downloads page do not specify the recommended version/patch. Explained that this was partially intentional, i.e. not to confuse people to apply 3.16 to production systems; in addition we wanted to wait for the FeaturePack2 as well as the extensions necessary for NextGenArchiver (they should come in the patch at the end of June). It was requested to mark one of the version anyway as "recommended for develpers". Riku to follow up
- FwCaen-8.3.2 released with a couple of fixes; 8.3.3 will come with yet more fixes relatively soon.
ISO->UTF migration
- Rafal presented the problematics: see slides in Indico
- Question: what is it that PARA displays? - the printable UTF content, hence it is useless for char/dyn_char; we will request ETM to have an alternative representation (e.g. hex or int)
- Proposal for the tool that could check for the problematic strings in the project (could be considered for development); should it run on "live" projects or on ASCII exports
- as agreed with ATLAS, for Linux systems such tool could be relatively trivial to achieve: a shell command (e.g. running iconv) plus diff, etc; would probably be an acceptable solution for ATLAS already
- Clarified also that there is yet another problem with panels, that could have separate strings for e.g. button titles or lists for ISO and UTF. Framework components were reviewed in 2018 (fw-8.2.0) to manually remove the incorrect ones; still if anyone made a copy-and-paste of e.g. old buttons of DE&N, unreadable or nonsense labels may be visible instead of expected ones; deserves a separate presentation
- What is the impact on archived strings? We need to review in detail, but for "problematic" strings we might see two differentt representations in the history then (one before and one after the migration); in addition this may depend on the configuration of the client AND the configuration of char-encoding on the database; as of now, all the Oracle DBs were created with ISO/Latin1 encoding; only recently a UTF test database were created by IT/DB for testing. Conversion of the whole database is out of question, as it would require rewriting all the recorded data
- What about the CtrlRDBAccess? Efforts were made by Pedro a year ago to test/enhance it; to be further tested with ConfDB, yet it was brought to the state where the behaviour should be predictable and "sane".
Proposal for new way of CTRL library loading
See Piotr's presentation for details.
- It was suggested to have an automated tool that would scan the scripts and libs (panels are not possible) using WCCOActrl -syntax to detect the unknown function calls, and then suggest necessary edits, or even apply such edits (insert the necessary #uses statements). Needs to be seen how much effort it would be.
- It was also pointed out that the CodePurger tool of BE-ICS developers could be used to scan the copies of the projects that experiments delivered to search for unresolved calls, and suggest initial list of edits. There are some LHCb projects that are missing for that though.
- We need to consider how to detect/correct FSM's DU Type scripts that would need to be edited; they are stored in datapoints...
- What about dpFunctions? Very limited, and we would leave them loaded by the LoadCtrlLib anyway
XML panel format
Another proposal: migrate to XML panel format, keeping the original ".pnl" filename extensions. This should also ease the transition of panels to UTF, following recommendations from the 3.16 release notes. No objection raised
Recent Improvements in CTRL++
Look at the presentation. Disclaimer: due to binary-compatibility, not all of the features will be back-portable to WinCC OA 3.16. Details to come.
AOB
Next JCOP FWWG likely at the beginning of July to report on the ETM Users Days (ie. Simatic SCADA Conference Europe).
There are minutes attached to this event.
Show them.