On 12/11/14 19:27, Adriano dos Santos Fernandes wrote:
> On 11/12/2014 14:06, Alex Peshkoff wrote:
>> On 12/11/14 18:32, Adriano dos Santos Fernandes wrote:
>>
Installing multiple ICU versions on Windows could be a challenge for
many users.
>>> http://site.icu-project.org/download
>>>
>>>
On 11/12/2014 14:06, Alex Peshkoff wrote:
> On 12/11/14 18:32, Adriano dos Santos Fernandes wrote:
>
>>> Installing multiple ICU versions on Windows could be a challenge for
>>> many users.
>> http://site.icu-project.org/download
>>
>>> So IMHO a gfix command really looks like a handy tool.
>> A de
On Thursday 11 December 2014 17:07:50 Adriano dos Santos Fernandes wrote:
> On 11/12/2014 13:42, Dmitry Yemanov wrote:
> > 11.12.2014 18:32, Adriano dos Santos Fernandes wrote:
> >> Copy back and forth and rebuilding indexes every time?
> >
> > A database created with incompatible (at least differe
On 11/12/2014 13:42, Dmitry Yemanov wrote:
> 11.12.2014 18:32, Adriano dos Santos Fernandes wrote:
>
>> Copy back and forth and rebuilding indexes every time?
> A database created with incompatible (at least different) ICU version
> simply cannot be accessed on the other host. This was my experien
On 12/11/14 18:32, Adriano dos Santos Fernandes wrote:
>> Installing multiple ICU versions on Windows could be a challenge for
>> many users.
> http://site.icu-project.org/download
>
>> So IMHO a gfix command really looks like a handy tool.
> A developer who cannot download a zip from the above ur
11.12.2014 18:32, Adriano dos Santos Fernandes wrote:
> Copy back and forth and rebuilding indexes every time?
A database created with incompatible (at least different) ICU version
simply cannot be accessed on the other host. This was my experience with
FB3 before Beta. I had to ask for a backu
On 11/12/2014 13:21, Dmitry Yemanov wrote:
> 11.12.2014 18:06, Adriano dos Santos Fernandes wrote:
>
>> Support for multiple ICU versions (and its maybe different data tables)
>> exists, user just need to install them!
> If I would be a support person dealing with customers databases, I'd
> rather
11.12.2014 18:06, Adriano dos Santos Fernandes wrote:
> Support for multiple ICU versions (and its maybe different data tables)
> exists, user just need to install them!
If I would be a support person dealing with customers databases, I'd
rather get fired instead of installing whetever ICU versi
On 11/12/2014 13:06, Dmitry Yemanov wrote:
> 11.12.2014 17:42, Alex Peshkoff wrote:
>
>> Have you ever seen in gstat -h? :)
> BTW, is it really about database creation?
>
> IIRC, ICU version is stored in collation specific attributes. If one
> creates one ICU collation, then upgrade ICU and create
11.12.2014 17:42, Alex Peshkoff wrote:
> Have you ever seen in gstat -h? :)
BTW, is it really about database creation?
IIRC, ICU version is stored in collation specific attributes. If one
creates one ICU collation, then upgrade ICU and create another collation
- wouldn't he/she be in the same
Support for multiple ICU versions (and its maybe different data tables)
exists, user just need to install them!
I don't like artificial commands that exposes internal details to users.
I'd instead test and document a bit what should be done.
Adriano
---
11.12.2014 15:24, Alex Peshkoff wrote:
> I do not think it's good idea to check all tables metadata on connect.
Isn't ICU version used on database creation written to header?
--
WBR, SD.
--
Download BIRT iHub F-Ty
On 12/11/14 17:34, Dimitry Sibiryakov wrote:
> 11.12.2014 15:24, Alex Peshkoff wrote:
>> I do not think it's good idea to check all tables metadata on connect.
> Isn't ICU version used on database creation written to header?
>
No.
Have you ever seen in gstat -h? :)
--
On 12/11/14 17:14, Dimitry Sibiryakov wrote:
> IMHO, the best solution for the problem would be an error on connect + an
> option for gfix.
>
I do not think it's good idea to check all tables metadata on connect.
At least this makes connect much slower.
What about using gfix instead SQL - g
IMHO, the best solution for the problem would be an error on connect + an
option for gfix.
--
WBR, SD.
--
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your
On 12/11/14 16:54, Dimitry Sibiryakov wrote:
> 11.12.2014 14:51, Alex Peshkoff wrote:
>> On 12/11/14 15:47, Dmitry Yemanov wrote:
>>> ALTER DATABASE USE ICU
>> Or may be
>> ALTER DATABASE USE CURRENT ICU
>> ?
> Why not simple "FIX ICU" or "FIX [ALL] COLLATIONS"?..
>
May be because except colla
11.12.2014 14:51, Alex Peshkoff wrote:
> On 12/11/14 15:47, Dmitry Yemanov wrote:
>> ALTER DATABASE USE ICU
> Or may be
> ALTER DATABASE USE CURRENT ICU
> ?
Why not simple "FIX ICU" or "FIX [ALL] COLLATIONS"?..
--
WBR, SD.
--
11.12.2014 14:46, Alex Peshkoff wrote:
> Issuing error on connect is bad option - may be one needs to access
> other tables w/o UTF columns?
Postponing error detection to undefined time doesn't seem as a good idea to
me either.
--
WBR, SD.
On 12/11/14 15:47, Dmitry Yemanov wrote:
> ALTER DATABASE USE ICU
Or may be
ALTER DATABASE USE CURRENT ICU
?
--
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Bu
On 12/11/14 13:07, marius adrian popa wrote:
> I guess compression can be further improved with lz4/snappy
Once again: I've tried lz4. It's really much faster - when compressing
some big (100Mb) binary with debug info it's 15 times faster than gzip,
but compression ratio is worse:
-rw-r--r-- 1
On 12/11/14 16:34, Dimitry Sibiryakov wrote:
> 11.12.2014 13:47, Dmitry Yemanov wrote:
>> Does user need to know what is the "current" ICU? Or "current" would
>> mean the one that can be found? What to do if multiple ICU versions are
>> installed and accessible?
> I have another question: how u
11.12.2014 16:39, Alex Peshkoff wrote:
>
> I like
> ALTER DATABASE RESET ICU
> or even
> ALTER DATABASE RESET INTERNATIONAL
> IMHO it's no big use explicitly setting VERSION here
> What do you mean here - no idea.
[VERSION] meant to explicitly specify ICU version to use (from
multiple available
On 12/11/14 15:47, Dmitry Yemanov wrote:
> 11.12.2014 15:32, Alex Peshkoff wrote:
>
>> I've almost finished with SQL command fixing database to make it usable
>> with ICU version not matching one with which database was created. This
>> helps in a cases like one described in
>> http://tracker.fireb
11.12.2014 13:47, Dmitry Yemanov wrote:
> Does user need to know what is the "current" ICU? Or "current" would
> mean the one that can be found? What to do if multiple ICU versions are
> installed and accessible?
I have another question: how user would know that (s)he must execute this
command
11.12.2014 15:32, Alex Peshkoff wrote:
> I've almost finished with SQL command fixing database to make it usable
> with ICU version not matching one with which database was created. This
> helps in a cases like one described in
> http://tracker.firebirdsql.org/browse/CORE-4571 . With mentioned
> i
I've almost finished with SQL command fixing database to make it usable
with ICU version not matching one with which database was created. This
helps in a cases like one described in
http://tracker.firebirdsql.org/browse/CORE-4571 . With mentioned
improvement one will be able to issue special S
I guess compression can be further improved with lz4/snappy
http://www.firebirdnews.org/testing-the-firebird-3-protocol-enhancements
--
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instan
On 12/11/14 00:22, Dmitry Yemanov wrote:
> Mark et al,
>
> > 1) It looks like Firebird no longer returns the number of rows requested
> > by the client. Jaybird (wire protocol) by default requests 400 rows and
> > in the past Firebird always returned the requested number of rows,
> > unless
28 matches
Mail list logo
