Zeoslib Portal

Thoughts: MySQL should perfectly be able to support IMultipleResults. Even if the patch make the driver running, i would start from the premisse there is a "unsupported" parameter set before. i.e. SetOleCommandProperties f.e. SupportsMultipleResultSets := DatabaseInfo.SupportsMultipleResul...

Hello Norbert
https://sourceforge.net/p/zeoslib/code-0/7086/ should fix it. Can you confirm it?

According the invalid sizes:
Could you plz test https://sourceforge.net/p/zeoslib/code-0/7073/

Nasty, i added a version which does not use OCI_ATTR_DISP_SIZE:
https://sourceforge.net/p/zeoslib/code-0/7073/
All tests are passing. Hope i didn't break anything else?

Hope https://sourceforge.net/p/zeoslib/code-0/7076/ will resolve that issue?
Thanks for pointing this out again? The testcases do not benchmark stl...

@Krill,
still not had time to write totorials. IMHO it should work like the FPC/FireDac/Unidac transaction components.
Did you test it?

Side notes: If you're using the component, avoid the Conenction [StartTransaction/Commit/Rollback] and use the TZTransaction component instead.
Was this helpfull?

Just let you know: Have no time to look to that bug. Guess ou are right. I will write a second post if i got it resolved..

@aehimself
According your patch i just wanna say: Awesome job! That rocks! Thank you very much.

You guys are right, hope the quick fix of https://sourceforge.net/p/zeoslib/code-0/7033/ will resolve the issue?

I already have doCachedLobs disabled in the TZQuery object to make sure all data is available when the grid starts updating. Isn't it exactly vice versa? CachedLobs are local memory lobs on client side. UnCached lobs are fetched from Server to Client on demand(If the server supports a nice api for ...

I thought that we agreed that the ADO driver can raise EZSQLExceptions/EDatabaseError and EOleExceptions as well as long as we document it? Why do we do that kind of rewrapping? We really do need to document design decisions. Well we can document everything ... does it make sence?? Is it handy for ...

Thx guys,

i added same logic to ODBC too.
Mark could you provide the Connecton string as an example for everbody?

@Jan the SAFCALL convention it's selve isn't the problem in first order. The point is the compiler always adds the error testing around each call even if it's not required in some circumstances. f.e. if we get objects from a Ole-Collection: For I := 0 to AdoStatement.(OleCheck)Params.Count(OleCheck)...

Hello Mark, yet the OleDB (ODBC_A/ODBC_W too) protocol is tested against SQLServer only. It's very stable and fast. Honestly i never had time to test it using other providers. We will add this limitation for the release notes of Zeos8. Later we can invest some time to get the protocols running for o...

@aehimself,

of course it can be done. Yet it has a low priority for me. Propose you open a feature request ticket: https://sourceforge.net/p/zeoslib/internaltickets/ so we won't forget it.