Re: [Firebird-devel] Odd difference for IN (..) and = ANY (..) in SELECT vs WHERE

2020-05-29 Thread Mark Rotteveel
On 29-05-2020 22:18, Vlad Khorsun wrote:   Try select not (null in (select 'a' from RDB$DATABASE)) from RDB$DATABASE select not (null = any (select 'a' from RDB$DATABASE)) from RDB$DATABASE and they also returns FALSE.   I didn't check details yet, but guess that expression result is NULL bu

[Firebird-devel] [FB-Tracker] Created: (CORE-6322) IN predicate and quantified comparison predicates behave incorrectly with NULL

2020-05-29 Thread Mark Rotteveel (JIRA)
IN predicate and quantified comparison predicates behave incorrectly with NULL -- Key: CORE-6322 URL: http://tracker.firebirdsql.org/browse/CORE-6322 Project: Firebird Core

Re: [Firebird-devel] Does Firebird 4 Beta 2 really need libncurses6?

2020-05-29 Thread Tony Whyman
On 29/05/2020 14:33, Alex Peshkoff via Firebird-devel wrote: On 29.05.2020 16:22, Tony Whyman wrote: Just tried to check out Beta 2 on Linux Mint 19.3 (= ubuntu bionic 18.04) and found that the binaries have been compiled to use libncurses6. Ubuntu 18.04 uses libncurses5. You need ubuntu 19.10

Re: [Firebird-devel] Odd difference for IN (..) and = ANY (..) in SELECT vs WHERE

2020-05-29 Thread Vlad Khorsun
29.05.2020 22:43, Mark Rotteveel wrote: I was just testing some claims in the Firebird Language Reference against Firebird 3 and 4, and ran into the following oddity: The result of `null in (non-empty list or query)` and `null = any (non-empty query)` should be `null` (see SQL:2016-2, 8.4 predi

Re: [Firebird-devel] Error message 'Install incomplete' in FB4

2020-05-29 Thread Mark Rotteveel
On 29-05-2020 22:07, Leyne, Sean wrote: The default list should only contain options which are available without additional configuration or initialization. I disagree. The default should represent best practice, most sensible and reasonable values. 1- SRP is not a commonly used protocol I

Re: [Firebird-devel] Odd difference for IN (..) and = ANY (..) in SELECT vs WHERE

2020-05-29 Thread Mark Rotteveel
On 29-05-2020 22:07, Mark Rotteveel wrote: Similar problems apply for `(null = any (..))`, this should produce TRUE when the subquery is empty, but NULL when non-empty. I meant to say this for `(null = all (..))`... Mark -- Mark Rotteveel Firebird-Devel mailing list, web interface at https:

Re: [Firebird-devel] Error message 'Install incomplete' in FB4

2020-05-29 Thread Leyne, Sean
> > The default list should only contain options which are available > > without additional configuration or initialization. > > I disagree. > > The default should represent best practice, most sensible and reasonable > values. 1- SRP is not a commonly used protocol 2- If not for the extra con

Re: [Firebird-devel] Odd difference for IN (..) and = ANY (..) in SELECT vs WHERE

2020-05-29 Thread Mark Rotteveel
On 29-05-2020 21:43, Mark Rotteveel wrote: I was just testing some claims in the Firebird Language Reference against Firebird 3 and 4, and ran into the following oddity: The result of `null in (non-empty list or query)` and `null = any (non-empty query)` should be `null` (see SQL:2016-2, 8.4

[Firebird-devel] Odd difference for IN (..) and = ANY (..) in SELECT vs WHERE

2020-05-29 Thread Mark Rotteveel
I was just testing some claims in the Firebird Language Reference against Firebird 3 and 4, and ran into the following oddity: The result of `null in (non-empty list or query)` and `null = any (non-empty query)` should be `null` (see SQL:2016-2, 8.4 and 8.9 ). This is the case for select (

Re: [Firebird-devel] Error message 'Install incomplete' in FB4

2020-05-29 Thread Jiří Činčura
> The default list should only contain options which are available > without additional configuration or initialization. I disagree. The default should represent best practice, most sensible and reasonable values. Extra configuration, especially for security/admin isn't a bad thing. BTW, I ha

Re: [Firebird-devel] Provider module unload

2020-05-29 Thread Vlad Khorsun
29.05.2020 18:27, Dimitry Sibiryakov wrote:   Hello, All.   Is provider module unloaded as soon as the last connection to it is closed or there is some delay in case that next connection come soon? Delay of 60 sec, IIRC. This is standard delay of Plugin Manager before unload any non-refer

Re: [Firebird-devel] Firebird 4.0 Beta 2 : 1) fbclient install on Linux 2) gsec

2020-05-29 Thread Pro Turm
I see. I would appreciate most, the answer of the question below (it is also under-documented ...) how should the correct symbolic link to the fbclient (libfbclient.so ?) look like, when the ./install.sh is executed ? Am Fr., 29. Mai 2020 um 17:17 Uhr schrieb Mark Rotteveel < m...@lawinegevaar.nl

[Firebird-devel] Provider module unload

2020-05-29 Thread Dimitry Sibiryakov
Hello, All. Is provider module unloaded as soon as the last connection to it is closed or there is some delay in case that next connection come soon? -- WBR, SD. Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel

Re: [Firebird-devel] [FB-Tracker] Created: (CORE-6321) GBAK can't restore database due to 'do not recognize table attribute'

2020-05-29 Thread Dimitry Sibiryakov
29.05.2020 17:13, Hugo Eyng | Datamais wrote: One or more records could be writing, in someway, in the FBK with a null value. It happens when a NOT NULL column is created when a table has records. Quoted errors has nothing to do with data. -- WBR, SD. Firebird-Devel mailing list, web i

Re: [Firebird-devel] Firebird 4.0 Beta 2 : 1) fbclient install on Linux 2) gsec

2020-05-29 Thread Mark Rotteveel
On 2020-05-29 14:01, Pro Turm wrote: 2) No, I meant executing gsec independently from the install.sh As gsec was deprecated with Firebird 3 and might be removed from future Firebird versions, it is probably good to get in the habit of using the SQL user management statements instead of gsec.

Re: [Firebird-devel] [FB-Tracker] Created: (CORE-6321) GBAK can't restore database due to 'do not recognize table attribute'

2020-05-29 Thread Hugo Eyng | Datamais
Hello. Did you try early creating the DB structure and changing "field1 INTEGER Not Null" to "field1 INTEGER " and after that restore with -REP option? One or more records could be writing, in someway, in the FBK with a null value. It happens when a NOT NULL column is created wh

[Firebird-devel] [FB-Tracker] Created: (CORE-6321) GBAK can't restore database due to 'do not recognize table attribute'

2020-05-29 Thread JIRA
GBAK can't restore database due to 'do not recognize table attribute' - Key: CORE-6321 URL: http://tracker.firebirdsql.org/browse/CORE-6321 Project: Firebird Core Issue Type

Re: [Firebird-devel] Does Firebird 4 Beta 2 really need libncurses6?

2020-05-29 Thread Alex Peshkoff via Firebird-devel
On 29.05.2020 16:22, Tony Whyman wrote: Just tried to check out Beta 2 on Linux Mint 19.3 (= ubuntu bionic 18.04) and found that the binaries have been compiled to use libncurses6. Ubuntu 18.04 uses libncurses5. You need ubuntu 19.10 if you don't want to go through the bother of backporting lib

Re: [Firebird-devel] Firebird 4.0 Beta 2 : 1) fbclient install on Linux 2) gsec

2020-05-29 Thread Pro Turm
How should the correct symbolic link look like? Am Fr., 29. Mai 2020 um 14:09 Uhr schrieb Alex Peshkoff via Firebird-devel < firebird-devel@lists.sourceforge.net>: > On 29.05.2020 15:02, Pro Turm wrote: > > error messages from my application like : "Dynamic linking error: > > fbclient.so not foun

[Firebird-devel] Does Firebird 4 Beta 2 really need libncurses6?

2020-05-29 Thread Tony Whyman
Just tried to check out Beta 2 on Linux Mint 19.3 (= ubuntu bionic 18.04) and found that the binaries have been compiled to use libncurses6. Ubuntu 18.04 uses libncurses5. You need ubuntu 19.10 if you don't want to go through the bother of backporting libncurses6. By contrast the most recent d

[Firebird-devel] [FB-Tracker] Created: (CORE-6320) Replace Util methods to get interface pointer by legacy handle with plain functions

2020-05-29 Thread Alexander Peshkov (JIRA)
Replace Util methods to get interface pointer by legacy handle with plain functions --- Key: CORE-6320 URL: http://tracker.firebirdsql.org/browse/CORE-6320 Project: Fireb

Re: [Firebird-devel] Firebird 4.0 Beta 2 : 1) fbclient install on Linux 2) gsec

2020-05-29 Thread Alex Peshkoff via Firebird-devel
On 29.05.2020 15:02, Pro Turm wrote: error messages from my application like : "Dynamic linking error: fbclient.so not found" Looks like installer creates symbolic link to client lib in wrong dir. That's certainly a bug to  be fixed. Right now you may create symlink in right dir yourself.

Re: [Firebird-devel] Firebird 4.0 Beta 2 : 1) fbclient install on Linux 2) gsec

2020-05-29 Thread Pro Turm
error messages from my application like : "Dynamic linking error: fbclient.so not found" Am Fr., 29. Mai 2020 um 14:01 Uhr schrieb Pro Turm : > 2) No, I meant executing gsec independently from the install.sh > > 1) is an independent question: > > do I need to install fbclient or do something extr

Re: [Firebird-devel] Firebird 4.0 Beta 2 : 1) fbclient install on Linux 2) gsec

2020-05-29 Thread Pro Turm
2) No, I meant executing gsec independently from the install.sh 1) is an independent question: do I need to install fbclient or do something extra while installing the Beta 2 http://web.firebirdsql.org/downloads/prerelease/v40beta2/Firebird-4.0.0.1963-Beta2.amd64.tar.gz ? Because after installati

Re: [Firebird-devel] Firebird 4.0 Beta 2 : 1) fbclient install on Linux 2) gsec

2020-05-29 Thread Mark Rotteveel
On 2020-05-29 09:27, Pro Turm wrote: 2) Ubuntu 20.04 Firebird Beta2 Installed firebird with sydba and password ending with "!!" (without quotation marks). gsec couldnt start with a password ending with "!!". Did you manually execute gsec, or is the install script still relying on gsec? With

Re: [Firebird-devel] Firebird 4.0 Beta 2 : 1) fbclient install on Linux 2) gsec

2020-05-29 Thread Pro Turm
Interesting enough with the same bash, beta 1 gsec worked. How about the 1st more important question? fbclient install with beta 2 ? Am Fr., 29. Mai 2020 um 10:23 Uhr schrieb Sergey Mereutsa : > Hello, > > IIRC, exclamation mark(s) "!!" has special meaning in shells and should be > protected usi

Re: [Firebird-devel] Firebird 4.0 Beta 2 : 1) fbclient install on Linux 2) gsec

2020-05-29 Thread Sergey Mereutsa
Hello, IIRC, exclamation mark(s) "!!" has special meaning in shells and should be protected using backslash or your password should be passed in single quotes. I faced this in Bash :) -- Best regards, Sergeymailto:s...@dqteam.com On Fri, 29 May 2020 at 10:27, Pro

[Firebird-devel] Firebird 4.0 Beta 2 : 1) fbclient install on Linux 2) gsec

2020-05-29 Thread Pro Turm
1) Executing the current ./install.sh on Linux would not install properly the fbclient, i.e. paths are not set correctly in order for a linker (e.g. ldconfig) and thus executables needing the fbclient cannot find it. How is the installation/setting of the environment of/for the fbclient on Linux e