nonomura escreveu no dia segunda, 29/10/2018 à(s)
05:47:
>
> Luis,
>
> Thank you very much!
> I'll examine it further and let yall know if found something.
>
> Hiro
You're Welcome
I believe the problem is in code that decides if the index is a full match
or a partial match, somehow in the case
Luis,
Thank you very much!
I'll examine it further and let yall know if found something.
Hiro
(2018/10/27 21:34), Luis Forra wrote:
I made the changes that you suggested:
and tested in Linux x64, didn't solve the problem
Firebird-Devel mailing list, web interface at
nonomura escreveu no dia sexta, 26/10/2018 à(s)
03:00:
> >
> >
> > >
> > >>
https://github.com/Alexpux/firebird-git-svn/blob/master/src/common/unicode_util.cpp#L1334
> > >
> > >
> > > Just a heads up: that is an outdated clone of an old repository that
> > > was synced from subversion and
https://github.com/Alexpux/firebird-git-svn/blob/master/src/common/unicode_util.cpp#L1334
Just a heads up: that is an outdated clone of an old repository that
was synced from subversion and hasn't been updated since 2015.
The Firebird repository is on
Adriano
You don't need to be on the development team. You know, you fix, you
send pull request.
Please put aside those nonsense and pointless arguments.
Shall we discuss in a constructive way?
Hiro
Firebird-Devel mailing list, web interface at
On 24-10-2018 08:20, nonomura wrote:
The source code cited below clearly tells the root of the problem that I
reported.
[..]
https://github.com/Alexpux/firebird-git-svn/blob/master/src/common/unicode_util.cpp#L1334
Just a heads up: that is an outdated clone of an old repository that was
On 25/10/2018 00:01, nonomura wrote:
>
> (2018/10/24 17:51), Adriano dos Santos Fernandes wrote:
>> As I said for you *many* times: if UNICODE (sensitive) has problems, as
>> it has,
> Nobody seems supporting your assumption.
>
Ok then.
>> how can you changing code under the insensitive flags
(2018/10/24 17:51), Adriano dos Santos Fernandes wrote:
As I said for you *many* times: if UNICODE (sensitive) has problems, as
it has,
Nobody seems supporting your assumption.
how can you changing code under the insensitive flags condition
will fix the problem?
If I were one of the
>
>Adriano dos Santos Fernandes escreveu no dia quarta,
24/10/2018 à(s) 11:52:
> >
> As I said for you *many* times: if UNICODE (sensitive) has problems, as
> > it has, how can you changing code under the insensitive flags condition
> > will fix the problem?
>
UNICODE collation doesn't
24.10.2018 12:51, Adriano dos Santos Fernandes wrote:
if UNICODE (sensitive) has problems, as
it has
I cannot remember that in this thread anyone said that it has problems.
Subject is "Case (and accent) *insensitive* ICU collations".
--
WBR, SD.
Firebird-Devel mailing list, web
As I said for you *many* times: if UNICODE (sensitive) has problems, as
it has, how can you changing code under the insensitive flags condition
will fix the problem?
Adriano
On 24/10/2018 03:20, nonomura wrote:
> The source code cited below clearly tells the root of the problem that
> I
The source code cited below clearly tells the root of the problem that I
reported.
The problem should be resolved by setting sortCollator's UCOL_STRENGTH
property as well as compareCollator.
There are two collater objects relating to this problem; Those are
sortCollator(*A) and
(2018/10/19 20:18), Dimitry Sibiryakov wrote:
AFAIK, sorting key is just an array of bytes and sorting comparison
routine treat it as an array of bytes. Key generation routine
transform any data into an array of bytes and after that operation
source data type is irrelevant.
I understood
19.10.2018 14:03, Adriano dos Santos Fernandes wrote:
ICU generates sort key for a single string. For multi segment (columns)
index/sort, Firebird call ICU (or any collation) for each column and
join the generated fully (with all levels/strengths) sort keys.
AFAIK, sorting key is just an
(2018/10/19 19:50), Adriano dos Santos Fernandes wrote:
I already showed to you that UNICODE (sensitive) has the problem, and it
does not nor can't use this condition.
If you wrote that way, I have no words other than that I could not see
the problem in UNICODE (sensitive) and nobody
On 19/10/2018 09:42, nonomura wrote:
> (2018/10/19 19:03), Adriano dos Santos Fernandes wrote:
>> ICU generates sort key for a single string. For multi segment
>> (columns) index/sort, Firebird call ICU (or any collation) for each
>> column and join the generated fully (with all levels/strengths)
(2018/10/19 19:03), Adriano dos Santos Fernandes wrote:
ICU generates sort key for a single string. For multi segment
(columns) index/sort, Firebird call ICU (or any collation) for each
column and join the generated fully (with all levels/strengths) sort
keys. To generate interleaved keys and
On 19/10/2018 06:24, nonomura wrote:
> Adriano,
>
> I've read your comment again, carefully, and understood what you were
> thinking as
> "the root of the problem".
>
> But I could not understand how it is relating to the case that I
> reported.
>
> ICU library generates sort keys for those three
Adriano,
I've read your comment again, carefully, and understood what you were
thinking as
"the root of the problem".
But I could not understand how it is relating to the case that I reported.
ICU library generates sort keys for those three collations and Firebird
uses it.
Firebird
>>I'm really trying to communicate with you here and in the tracker
I hate to tell you this but I don't think so.
You always wrote without reading my comment carefully with your
prejudice, just like my first reply to you here :)
>>you seem to answer every point of mine as rewriting what I
I'm really trying to communicate with you here and in the tracker but
you seem to answer every point of mine as rewriting what I wrote giving
another meaning.
Or it seems you're so expert on the subject, so please send pull
request. It's very easy.
Adriano
On 18/10/2018 17:06, nonomura wrote:
>the root of the problem of multi-level and multi-segment with it.
Sorry, it's like I missed the very point you meant.
Ok, you will see that it is not the problem of multi-level and
multi-segment if you examine the exact sort key in that site.
The root of the problem should be much simpler as
Hi,
>then
When? :)
Sort keys seem to be stored fully(body,case,accent) for utf8 fields.
And compared fully where case or case and accent should not be compared.
That's what I wrote.
Regards,
Hiro
Please visit this site:
(http://demo.icu-project.org/icu-bin/collation.html )
and try to see
>No. Base letter is most important.
Sorry, I just meant that base letter's importance does not change anytime.
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
>I really don't understand why you emphasize so much on
case-/accent-insensitivity
Because it is a bug. that's all.
>Case is less important than base letter.
No. Base letter is most important.
a1 A1 a2 A2 ... This is correct result as they are ordered by CSAS in
icu collation rule.
If case
Hi,
maybe naive. But why keys ar not stored in uppercase then?
Regards,Karol Bieniaszewski
nullFirebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel
I really don't understand why you emphasize so much on
case-/accent-insensitivity, and say UNICODE is ok, as we can easily
demonstrate the root of the problem of multi-level and multi-segment
with it.
Yes, there is inconsistencies with insensitive with/without unique, but
root of the problem
Hi there,
I would like to ask someone who can check and correct if bug was found
to see the relating part of source code regarding sorting on UTF8 + ICU
collations.
The symptom and what I tested and confirmed have commented in CORE-5940.
The following is the summary:
1. UNICODE_CI and
28 matches
Mail list logo