2019 WISH LIST
---
- A list of libraries that are known to be successfully added to SQLite.
- Have a base instance, select some check-boxes, press SUBMIT, something
somewhere somehow generates your .DLL, .SO, etc. so that you have
everything you want built in
- A
On May 23, 2019, at 4:28 PM, R Smith wrote:
>
> it is very easy to add things to the base distro, but extremely hard to
> impossible to ever take it away again, which means one should only ever "add"
> with great caution.
Easy fix: -DSQLITE_OMIT_EXTENDED_MATH_LIBRARY
Include it by default,
On 2019/05/23 1:26 PM, J. King wrote:
On May 23, 2019 6:46:52 a.m. EDT, R Smith wrote:
This is SQLite. Perhaps some of us could collaborate on a fork called
SQLbloat //..
I find this a little condescending. There's a lot of reasons to like SQLite, and the aspect that
sways me more than
On Thursday, 23 May, 2019 14:39, Jens Alfke wrote:
>> On May 22, 2019, at 8:16 PM, Keith Medcalf
>wrote:
>> Basically, User Defined Types (UDT) were implemented in a fashion
>analgous to a C++ class (remember that at this time C++ was just a
>pre-processor for C and a C++ class was nothing
On Thursday, 23 May, 2019 08:35, Dominique Devienne wrote:
>On Thu, May 23, 2019 at 3:50 PM Jose Isaias Cabrera
> wrote:
>
>> I have been working network for a long time, and I have never seen
>> any application that takes "zeroed left-filled" IP addresses. Just
>> sharing...
>> Thanks.
>
> On May 22, 2019, at 8:16 PM, Keith Medcalf wrote:
>
> Basically, User Defined Types (UDT) were implemented in a fashion analgous to
> a C++ class (remember that at this time C++ was just a pre-processor for C
> and a C++ class was nothing more than a struct and mangled function names to
>
On Wed, 22 May 2019 17:56:23 -0700
Jens Alfke wrote:
> > On May 22, 2019, at 3:55 PM, James K. Lowden
> > wrote:
> >
> > I've always thought user-defined types were unnecessary except as a
> > convenience.
>
> User-defined types are quite important if you?re doing fancy stuff in
>
On Wed, 22 May 2019 21:16:04 -0600
"Keith Medcalf" wrote:
> Basically, when you declared something as a UDT you were giving a
> "blob" a type-domain. Whenever you tried to do something with a UDT
> a "mangled function name" was generated that took that blob as the
> first argument and you
On Wed, May 22, 2019 at 03:43:13PM -0400, Richard Hipp wrote:
> On 5/22/19, Sebastian Kemper wrote:
> >
> > Hello Richard,
> >
> > I have run-tested this with qemu on two targets until now:
> >
> > arm_arm1176jzf-s_vfp
> > armeb_xscale
> >
> > Both worked fine. Unfortunately I still don't have
On 23 May 2019, at 12:02pm, Dominique Devienne wrote:
> On Thu, May 23, 2019 at 12:37 PM Simon Slavin wrote:
>
>> IP addresses are easy: convert to the form aaa.bbb.ccc.ddd where each of the
>> four parts is always three digits long.
>
> Sure. But representing it as 001.001.001.001 for
On Thu, May 23, 2019 at 5:14 PM Jen Pollock wrote:
> This is getting pretty far off topic, but I think at least some tools
> will interpret values with leading zeroes as octal, which means 001 is
> the same as 1, but 010 isn't 10, it's 8.
>
Good catch! That's indeed what's happening. Win7 BTW.
This is getting pretty far off topic, but I think at least some tools
will interpret values with leading zeroes as octal, which means 001 is
the same as 1, but 010 isn't 10, it's 8.
On Thu, May 23, 2019 at 04:35:02PM +0200, Dominique Devienne wrote:
> On Thu, May 23, 2019 at 3:50 PM Jose Isaias
Dominique Devienne, on Thursday, May 23, 2019 10:35 AM, wrote...
>On Thu, May 23, 2019 at 3:50 PM Jose Isaias Cabrera
>wrote:
>
>Works for me with a .001 at least, as shown below. But that wasn't really
>the point I was making, FWIW. --DD
>
>C:\Users\ddevienne>ping 192.168.223.001
>
>Pinging
On Thu, May 23, 2019 at 3:50 PM Jose Isaias Cabrera
wrote:
> I have been working network for a long time, and I have never seen any
> application that takes "zeroed left-filled" IP addresses. Just sharing...
> Thanks.
>
Works for me with a .001 at least, as shown below. But that wasn't really
J. King, on Thursday, May 23, 2019 07:26 AM, wrote...
>On May 23, 2019 6:46:52 a.m. EDT, R Smith wrote:
>
>>This is SQLite. Perhaps some of us could collaborate on a fork called
>>SQLbloat and put out standard libs/code/precompileds for versions of
>>sqlite with everything - bbq sauce and all,
Dominique Devienne, on Thursday, May 23, 2019 07:02 AM, wrote...
>On Thu, May 23, 2019 at 12:37 PM Simon Slavin wrote:
>> On 23 May 2019, at 3:55am, Keith Medcalf wrote:
>> > Technically, COLLATE only works on TEXT. Most people declare their own
>> types as binary blobs and the programmer has
Thank you for bringing topic back to original intent.
Clearly much is missing from the latest SQL standard. Debates aside, the
URL is equally lacking.
Yes, "own" extensions solve much of my needs.
On Thu, May 23, 2019, 7:26 AM J. King wrote:
> On May 23, 2019 6:46:52 a.m. EDT, R Smith wrote:
Don't forget IPv6 addresses.
On Thu, May 23, 2019, 3:37 AM Simon Slavin wrote:
> On 23 May 2019, at 3:55am, Keith Medcalf wrote:
>
> > Technically, COLLATE only works on TEXT. Most people declare their own
> types as binary blobs and the programmer has to keep track of what is in
> there and
On May 23, 2019 6:46:52 a.m. EDT, R Smith wrote:
>This is SQLite. Perhaps some of us could collaborate on a fork called
>SQLbloat and put out standard libs/code/precompileds for versions of
>sqlite with everything - bbq sauce and all, for when you don't need
>Lite
>- then you can specify that
This doesn't solve the problem. The database must work regardless of whether it
is used within my own app or any other database viewer that might not have the
extension functions available. In the first case, medians, standard deviations,
etc. are included in the view, in the second case the
This ain't quite helpful. Sqlite also supports FTS3/4/5 which I would
personally consider as bloat. So what is the actual "bloat" in supporting a
function that can just check whether a function exists or not. That would
probably add only some bytes to the library.
- Original Message -
On Thu, May 23, 2019 at 12:37 PM Simon Slavin wrote:
> On 23 May 2019, at 3:55am, Keith Medcalf wrote:
> > Technically, COLLATE only works on TEXT. Most people declare their own
> types as binary blobs and the programmer has to keep track of what is in
> there and how to work with it.
>
> So
Ok, but you can't have it both ways. Either you want to write generic
use-everywhere SQL, in which case you need to stick to the generic
included-everywhere functions,
OR
You want to write specialized specific queries that use either your own
or other UDF's, in which case you must be able to
On 23 May 2019, at 3:55am, Keith Medcalf wrote:
> Technically, COLLATE only works on TEXT. Most people declare their own types
> as binary blobs and the programmer has to keep track of what is in there and
> how to work with it.
So it would seem that rather than define a function which
On 23 May 2019, at 7:57am, Thomas Kurz wrote:
> CREATE VIEW foo AS SELECT {if has stddev then stddev(...) else null} FROM ...
'if' in SQL language is CASE.
Near the beginning of your code, try to execute a function that contains
stddev(). Make a note of whether it compiles without errors or
That doesn't make any difference. Then I could use the extensions-functions.c
loadable module as well. My database has to work equally well no matter what
dll and/or extension is used or not.
Regardless of my application, the problem stays the same when you open the
database in any other
Put your function into a loadable extension and load it during application
startup so that it is always available to your code. This does not require
checking the sqlite3.dll
-Ursprüngliche Nachricht-
Von: sqlite-users [mailto:sqlite-users-boun...@mailinglists.sqlite.org] Im
Auftrag
I want to define VIEWs that work equally well regardless of whether a default
sqlite3.dll or a custom build with built-in extension-functions.c is used.
Defining my own function as an extension module which checks that case would
additionally require checking whether a default sqlite3.dll or a
On Thu, May 23, 2019 at 11:09 AM Keith Medcalf wrote:
> >Keith, as can be seen below, those pragma_*list eponymous vtables are
> >you referring to as not built-in.
>
> Actually they are. They are the eponymous vtables for the corresponding
> pragma's:
>
> pragma function_list;
> pragma
On Thursday, 23 May, 2019 02:08, Dominique Devienne wrote:
>On Thu, May 23, 2019 at 7:39 AM Keith Medcalf
>wrote:
>> You can check if what you need is available on a connection and
>either load it if needed or just abort:
>> sqlite> select * from pragma_function_list order by 1, 2;
>> name
Just execute the SQL containing the function. If the function does not exist
then you will get an error thrown when you attempt to prepare the statement
containing the function that does not exist ... and your application can parse
the error message and do the needful. Of course, just
On Thu, May 23, 2019 at 7:39 AM Keith Medcalf wrote:
> You can check if what you need is available on a connection and either
> load it if needed or just abort:
>
> sqlite> select * from pragma_function_list order by 1, 2;
> name builtin
> -
Just write a function that takes a function name, a default value and an
unspecified number of arguments.
function_present(,[,])
The implementation will then check if the named function is available;
If so, prepare, execute and return the result of "SELECT ()"
If not, just return .
This is
Ok, thank you for that hint. But it is still very unconvenient. How can I
define a view based on your suggestion? I want to have something like
CREATE VIEW foo AS SELECT {if has stddev then stddev(...) else null} FROM ...
- Original Message -
From: Keith Medcalf
To: SQLite mailing
34 matches
Mail list logo