I think we should probably start adding some more comments to the
PlcResponseCode enum to describe in which szenario which value should be used.
Chris
Von: Jinlin Hong <myhongjin...@gmail.com>
Datum: Montag, 23. September 2024 um 10:16
An: dev@plc4x.apache.org <dev@plc4x.apache.org>
Betreff: Re: [DISCUSS] Which return code to give for items that don't exist?
Then I agree with your view, +1. 'return INVALID_ADDRESS if the address is
syntactically invalid and NOT_FOUND, if the address is formally correct,
but the resource could not be found.'
Christofer Dutz <christofer.d...@c-ware.de> 於 2024年9月23日 週一 下午4:03寫道：
> The thing is however, that S7 for example has “NOT FOUND” if an address
> could not be found. And that actually represents the reality a lot better
> .. because the addresss is actually valid, just not on that device.
>
> Chris
>
> Von: Jinlin Hong <myhongjin...@gmail.com>
> Datum: Montag, 23. September 2024 um 09:59
> An: dev@plc4x.apache.org <dev@plc4x.apache.org>
> Betreff: Re: [DISCUSS] Which return code to give for items that don't
> exist?
> Hi Chris,
>
>
> I think it makes sense to keep INVALID_ADDRESS consistent with the PLC
> logic, meaning it would represent the case when a resource can’t be found.
> Then we could use something like ERROR_ADDRESS_FORMAT to indicate a syntax
> error. I just checked plc4go and saw errors.Errorf("Invalid address format
> for address '%s'", tagAddress) being used. Looking forward to hearing
> thoughts from the rest of the team.
>
>
> Jinlin Hong
>
> Christofer Dutz <christofer.d...@c-ware.de> 於 2024年9月23日 週一 下午3:22寫道：
>
> > While working on this I noticed, we need to update the ReadRequest a bit
> …
> > When we call “build” on the ReadRequest builder and that passes all
> > address strings to the tag-handler. If there’s only one invalid address,
> > the entire build() call fails.
> >
> > I think we need to add a:
> >
> > PlcResponseCode getTagResponseCode(String tagName);
> >
> > To check if a tag in the ReadRequest was OK, if it’s not to simply relay
> > that to the result.
> >
> > Or what do you think would be a better solution?
> >
> > Chris
> >
> > Von: Lukas Ott <ott.lukas...@gmail.com>
> > Datum: Sonntag, 22. September 2024 um 15:30
> > An: dev@plc4x.apache.org <dev@plc4x.apache.org>
> > Betreff: Re: [DISCUSS] Which return code to give for items that don't
> > exist?
> > +1 for "return INVALID_ADDRESS if the address is syntactically invalid
> and
> > NOT_FOUND, if the address is formally correct, but the resource could not
> > be found." sounds logical and understandable to me.
> >
> > Luk
> >
> > Christofer Dutz <christofer.d...@c-ware.de> schrieb am So., 22. Sept.
> > 2024,
> > 14:38:
> >
> > > Hi all,
> > >
> > > I have been working a bit on streamlining the return results from our
> > > drivers to be a bit more consistent.
> > > Now I noticed that there is quite a huge difference between
> > > INVALID_ADDRESS and NOT_FOUND.
> > >
> > > Most drivers return an INVALID_ADDRESS if for example a symbolic
> address
> > > is used, and this does not exist on the PLC.
> > > However the address IS valid, it’s just that the PLC can’t find a
> > resource
> > > with that given name. Similar with S7, if I specify a block number that
> > > does not exist or a byte offset that exceeds the size of the block.
> > >
> > > I think we should probably return INVALID_ADDRESS if the address is
> > > syntactically invalid and NOT_FOUND, if the address is formally
> correct,
> > > but the resource could not be found.
> > >
> > > What do you think?
> > >
> > > Chris
> > >
> > >
> > >
> >
>
