Re: [sqlite] bug? suggestion? add full version number to files, so tools can warn of forward-incompatibility

2016-12-30 Thread Stephen Chrzanowski
I didn't interpret your statement properly. I thought you were referring to a warning thrown by the library itself, not the SQLite CLI. For the CLI, since its application based, I can see the use. If it were to state the DB version versus the CLI version of the library on load, that'd be cool, b

Re: [sqlite] bug? suggestion? add full version number to files, so tools can warn of forward-incompatibility

2016-12-30 Thread Bennett Haselton
On 12/30/2016 5:25 PM, Stephen Chrzanowski wrote: IMO, the responsibility of checking database versions should be owned by the application, not the library. The logic that the application can or cannot, should or should not use the database is an application decision. If the library just were to

Re: [sqlite] bug? suggestion? add full version number to files, so tools can warn of forward-incompatibility

2016-12-30 Thread Richard Hipp
On 12/30/16, Bennett Haselton wrote: > My other suggestion was that if you open a database file with a > *newer* version of the library than the one that was used to create it, > you can also warn, "This file was created using SQLite version a.b.c, > but you're attempting to save it in format SQL

Re: [sqlite] bug? suggestion? add full version number to files, so tools can warn of forward-incompatibility

2016-12-30 Thread Stephen Chrzanowski
IMO, the responsibility of checking database versions should be owned by the application, not the library. The logic that the application can or cannot, should or should not use the database is an application decision. If the library just were to provide what version of SQLite made or last modifie

Re: [sqlite] bug? suggestion? add full version number to files, so tools can warn of forward-incompatibility

2016-12-30 Thread Bennett Haselton
On 12/29/2016 9:20 PM, Richard Hipp wrote: On 12/30/16, Bennett Haselton wrote: Presumably it's not possible for a tool to output a detailed message like "Your file was generated by SQLite library version 3.6.2, but this tool only supports versions up to 3.5.1", OK. I have your request to en

Re: [sqlite] bug? suggestion? add full version number to files, so tools can warn of forward-incompatibility

2016-12-30 Thread Jim Borden
For what it's worth the file stores the version number of library that most recently opened the file as part of the 100 byte header (the last four bytes specifically). It's just not in the first 16 byte magic portion. Changing this would immediately cause all previous versions to report databas

Re: [sqlite] bug? suggestion? add full version number to files, so tools can warn of forward-incompatibility

2016-12-30 Thread Bennett Haselton
It was SQLite 3.3.6 -- I know, more than 10 years old, but it's the latest version that CentOS 5.5 will update to automatically, and I was strongly advised against updating individual components to anything more recent than what "yum update" would do by default. In any case it wasn't worth tak

Re: [sqlite] bug? suggestion? add full version number to files, so tools can warn of forward-incompatibility

2016-12-29 Thread Richard Hipp
On 12/30/16, Bennett Haselton wrote: > > Presumably it's not possible for a tool to output a detailed message > like "Your file was generated by SQLite library version 3.6.2, but this > tool only supports versions up to 3.5.1", OK. I have your request to enhance the error message that is issued

Re: [sqlite] bug? suggestion? add full version number to files, so tools can warn of forward-incompatibility

2016-12-29 Thread Richard Hipp
On 12/29/16, Bennett Haselton wrote: > Yesterday I spent some time trying to solve a SQLite problem that turned > out to be due to using an old version of sqlite3, so that when I tried > to open a database created with a newer SQLite library, it would output > "Error: file is encrypted or is not a

[sqlite] bug? suggestion? add full version number to files, so tools can warn of forward-incompatibility

2016-12-29 Thread Bennett Haselton
Yesterday I spent some time trying to solve a SQLite problem that turned out to be due to using an old version of sqlite3, so that when I tried to open a database created with a newer SQLite library, it would output "Error: file is encrypted or is not a database". Later I found out this was be