Re: DBD/DB validator?

[Michael Davidsaver <mdavidsaver@bnl.gov>]

I'd second the call for a syntax check of .db files.  The fact that this
gets discovered at runtime is terrible.

I can't count the number of times I've missed (and seen others miss) a
closing quote, and then miss the error from dbLoadRecord because it
scrolled off the top of the screen.


The complicate is that all the parsers in Base are designed to work on
full definitions.

On a .db file with only record definitions dbExpand from 3.14 gives:

> /usr/lib/epics/bin/linux-x86_64/dbExpand -I. cacount.db >/dev/null
> Record "cnt1" is of unknown type "calc"

dbdExpand.pl from 3.15 doesn't recognize the "record" keyword.

> ./bin/linux-x86_64/dbdExpand.pl -I. cacount.db 
> dbdExpand.pl: Syntax error in 'record(calc, "cnt1") {'


So I can see two courses.

1) Modify the makefiles so that the validator can be given .dbd file(s)
as well.  This would allow (an updated) "dbdExpand ... >/dev/null" to
serve as a check.

2) Write a new utility which parses, but doesn't follow include
statements, then does as much checking as possible.


Michael



On 04/23/2015 05:56 AM, Ralph Lange wrote:
> Dear all,
> 
> Ages ago I remember there was a tool called 'dbst' that could be used to 
> validate DBD/DB setups and that could dump DBs in a sorted, normalized 
> way (only non-default value fields, standard indentation) to make them 
> easily comparable with diff and such.
> 
> Is dbst still around? I suppose there are more recent ways nowadays. 
> Perl/Python? Web-Service?
> (I do like the idea of using the same library as the IOC, though...)
> 
> What is a good way to do database validation?
> 
> Thanks a lot,
> ~Ralph
>