[sqlite] System.Data.SQLite version 1.0.110.0 released

System.Data.SQLite version 1.0.110.0 (with SQLite 3.27.2) is now available on the System.Data.SQLite website: https://system.data.sqlite.org/ Further information about this release can be seen at: https://system.data.sqlite.org/index.html/doc/trunk/www/news.wiki Please post on the

Re: [sqlite] Handling ROLLBACK

On 3-3-2019 16:27, Jesse Rittner wrote: It didn't work correctly because the sequence of commands you sent doesn't make sense. BEGIN; ... RELEASE point1; ... ROLLBACK; ... END; First you began an explicit transaction. Then you tried to release a savepoint that you never created. (Hence the

Re: [sqlite] Handling ROLLBACK

It didn't work correctly because the sequence of commands you sent doesn't make sense. BEGIN; ... RELEASE point1; ... ROLLBACK; ... END; First you began an explicit transaction. Then you tried to release a savepoint that you never created. (Hence the "no such savepoint" error message.) Then

Re: [sqlite] Handling ROLLBACK

On 3-3-2019 15:33, Simon Slavin wrote: On 3 Mar 2019, at 2:29pm, Luuk wrote: Conclusion: RESTORE does not end TRANSACTION ? Your statement is correct. However, RESTORE is a partner of SAVEPOINT. My question does not consider SAVEPOINTs. Simon.

Re: [sqlite] Handling ROLLBACK

On 3 Mar 2019, at 2:29pm, Luuk wrote: > Conclusion: RESTORE does not end TRANSACTION ? Your statement is correct. However, RESTORE is a partner of SAVEPOINT. My question does not consider SAVEPOINTs. Simon. ___ sqlite-users mailing list

Re: [sqlite] Handling ROLLBACK

On 3-3-2019 15:01, Richard Damon wrote: On Mar 3, 2019, at 8:32 AM, Simon Slavin wrote: To summarize, the list feels that this is an incorrect model BEGIN; ... first set of commands ROLLBACK; ... second set of commands END; whereas this is how things are meant to work:

Re: [sqlite] Handling ROLLBACK

On Mar 3, 2019, at 8:32 AM, Simon Slavin wrote: > > To summarize, the list feels that this is an incorrect model > >BEGIN; >... first set of commands >ROLLBACK; >... second set of commands >END; > > whereas this is how things are meant to work: > >BEGIN; >... first

[sqlite] Documentation change request

talks about PRAGMA read_uncommitted = boolean; Explicitly stating that the argument should be boolean. However the explanation of the effect of the PRAGMA talks about its argument as if it is SERIALIZABLE. No boolean value to indicate

Re: [sqlite] Handling ROLLBACK

To summarize, the list feels that this is an incorrect model BEGIN; ... first set of commands ROLLBACK; ... second set of commands END; whereas this is how things are meant to work: BEGIN; ... first set of commands ROLLBACK; BEGIN; ... second set of

Re: [sqlite] Handling ROLLBACK

My observation (on the current tip version 3.28.0) of Schrodingers Transactions is that if there is (for example) a transaction in progress and that is COMMIT or ROLLBACK, then the changes are either committed or rolled back and the explicit transaction is ended (that is, autocommit becomes