firebird-test — Firebird engine testing, test flight/suite discussion/information

RE: [Firebird-test] Blob segmentation in gtcs.gdb

[Mark O'D. <Mar...@lu...>]

Hi Frank

If we need to we can store ib-tests in our CVS tree just in a different
module so it's not part of the interbase download.  Ultimately it's not a
bad idea to have our own copy, and it's pretty easy to do.  

So if you eventually decide you need to make a change, I think we'll put
import it into our CVS tree.

Cheers

Mark


-----Original Message-----
From: Sch...@t-...
To: fir...@li...
Sent: 11/10/00 3:27 AM
Subject: [Firebird-test] Blob segmentation in gtcs.gdb

Hi,
after a long frustrating day, I finally found out
why the TCS fails to work with a gtcs.gdb created
by my import routines.

Some routines expect the blobs in INITS.OUTPUT to
contain a segment per line.

I now can either modify my import routines to
store the blobs in a way that the existing tcs can
handle them or change the tcs code to work with blobs 
regardless of segment size.

At the moment I would prefer the second solution, 
because I think it would be good to be able
to insert data into the tcs database with any
tool capable to do it.


Any comments?

Frank
-- 
"Fascinating creature, phoenixes, they can carry immensely heavy loads,
 their tears have healing powers and they make highly faithful pets."  
         - J.K. Rowling

http://sourceforge.net/projects/firebird
_______________________________________________
Firebird-test mailing list
Fir...@li...
http://lists.sourceforge.net/mailman/listinfo/firebird-test

Re: [Firebird-test] Blob segmentation in gtcs.gdb

[<Sch...@t-...>]

"Leyne, Sean" schrieb:
> 
> I agree!
> 
> The TCS should not depend on segment size - it would be the first
> application that I had heard of which does.
> 

I'm not so sure about this, BLOB_edit and the likes store
blobs exactly in this way, and there are these suspicious
BLOB_text_dump and BLOB_text_load in gds.h.

Frank


-- 
"Fascinating creature, phoenixes, they can carry immensely heavy loads,
 their tears have healing powers and they make highly faithful pets."  
         - J.K. Rowling

http://sourceforge.net/projects/firebird

RE: [Firebird-test] Blob segmentation in gtcs.gdb

[Leyne, S. <sl...@at...>]

I agree!  

The TCS should not depend on segment size - it would be the first
application that I had heard of which does.


Sean

 -----Original Message-----
From: 	Sch...@t-...
[mailto:Sch...@t-...] 
Sent:	Thursday, November 09, 2000 11:27 AM
To:	fir...@li...
Subject:	[Firebird-test] Blob segmentation in gtcs.gdb

Hi,
after a long frustrating day, I finally found out
why the TCS fails to work with a gtcs.gdb created
by my import routines.

Some routines expect the blobs in INITS.OUTPUT to
contain a segment per line.

I now can either modify my import routines to
store the blobs in a way that the existing tcs can
handle them or change the tcs code to work with blobs 
regardless of segment size.

At the moment I would prefer the second solution, 
because I think it would be good to be able
to insert data into the tcs database with any
tool capable to do it.


Any comments?

Frank
-- 
"Fascinating creature, phoenixes, they can carry immensely heavy loads,
 their tears have healing powers and they make highly faithful pets."  
         - J.K. Rowling

http://sourceforge.net/projects/firebird
_______________________________________________
Firebird-test mailing list
Fir...@li...
http://lists.sourceforge.net/mailman/listinfo/firebird-test

[Firebird-test] Blob segmentation in gtcs.gdb

[<Sch...@t-...>]

Hi,
after a long frustrating day, I finally found out
why the TCS fails to work with a gtcs.gdb created
by my import routines.

Some routines expect the blobs in INITS.OUTPUT to
contain a segment per line.

I now can either modify my import routines to
store the blobs in a way that the existing tcs can
handle them or change the tcs code to work with blobs 
regardless of segment size.

At the moment I would prefer the second solution, 
because I think it would be good to be able
to insert data into the tcs database with any
tool capable to do it.


Any comments?

Frank
-- 
"Fascinating creature, phoenixes, they can carry immensely heavy loads,
 their tears have healing powers and they make highly faithful pets."  
         - J.K. Rowling

http://sourceforge.net/projects/firebird

[Firebird-test] RE: patch run

[Mark O'D. <Mar...@lu...>]

 
Yep,

should be about a 24hr turnaround time.

Cheers

Mark

-----Original Message-----
From: Griffin, Patrick J.
To: Mark O'Donohue
Sent: 11/9/00 4:15 PM
Subject: RE: patch run

If you can, I have loaded some tests into the firebird tree
interbase/firebird/patrickgriffin/tcs_tests   which exercise my patch.

Could you give them a try?

Thanks,
 
...pat
p.s. The messages at the bottom look normal to me. 

-----Original Message-----
From: Mark O'Donohue
To: 'gr...@su...'
Sent: 11/8/00 6:25 PM
Subject: FW: patch run


Hi Pat

Ok the SS on redhat linux 6.2  run through TCS.  The only thing strange
was
the following bits.

You may have to remind me of any other tests you wanted me to do.

These are the original set of tests from ib-test's I haven't updated it
with
Franks latest ones as yet.


Cheers

Mark

-----Original Message-----
From: Mark O'Donohue
To: Mark O'Donohue
Sent: 11/9/00 8:41 AM
Subject: ss run


**** tcs:  Test "FLD0_V40_1008       " with version <= V006000000 not
found
Running metaseries VECTOR_12HOUR      
Running series RI_TESTS             started on Thu Nov  9 03:32:42 2000
Running global test C_DSQL_RI_INIT       ... passed
Running global test C_DSQL_RI_ALT_ADD    ... passed
Running global test C_DSQL_RI_ALT_DROP   ... passed
Running global test C_DSQL_RI_COL_REF    ... passed

**** tcs:  Test "C_DSQL_RI_COL_CHECK " with version <= V006000000 not
found

-- 
Your database needs YOU!
http://sourceforge.net/projects/firebird

Re: [Firebird-test] Adding ib-tests to CVS

[<Sch...@t-...>]

Mark O'Donohue schrieb:
> 
> Sch...@t-... wrote:
> 
> > Nice book. I thought this sort of things could be done only
> > by members of the admin group.
> 
> Im not sure, In theory all you need is CVS access, but I keep finding
> extra little things that you need to be admin to do.

I will try what can be done without beeing an admin :-)


> >
> > I have seen this after I uploaded the exported tests to sourceforge,
> > so they are now in interbase/firebird/fsg/TCS
> 
> But that is just your extra stuff not the whole ib-tests thing is that
> right?

I have no intention to copy the whole tcs stuff from ib-tests to
the firebird tree as long I don't need to make some massive changes
to the source of the TCS. 
But the extra stuff that I uploaded today was about 3 MB so this would
be better placed outside of the main tree.

> 
> > BTW shouldn't we move all the private directories that now
> > are in the main tree to an extra module?
> 
> Yep I was thinking of adding a "personal" module with user
> subdirectories for self storage,
> 
> But I still think that we also need one within interbase itself to
> distribute those "little goodies" to do with build etc, before they
> become widely accepted (probably meaning compiling on a win32 platforms ;-).

Agreed.

Frank

-- 
"Fascinating creature, phoenixes, They can carry immensely heavy loads,
 their tears have healing powers and they make highly faithful pets."  
         - J.K. Rowling

Re: [Firebird-test] Adding ib-tests to CVS

[Mark O'D. <mar...@lu...>]

Sch...@t-... wrote:

> Mark O'Donohue schrieb:
> 
>> Hi Frank,
>> 
>> When you feel you have a comfortable directory structure for the TCS
>> test suite from ib-tests.
>> And perhaps after a bit of discussion on structure, we can import it in
>> as a new module into firebird.   Then we have it under CVS control and
>> can add things to it etc.
>> 
>> The commands to do it are fairly easy and are explained in:
>> 
>> http://cvsbook.red-bean.com/
> 
> Nice book. I thought this sort of things could be done only
> by members of the admin group. 

Im not sure, In theory all you need is CVS access, but I keep finding 
extra little things that you need to be admin to do.

>> Having it as a seperate module from interbase has some advantages.
>> 
>> There's no rush, but when your ready let us know and we'll see what we can do.
> 
> I have seen this after I uploaded the exported tests to sourceforge,
> so they are now in interbase/firebird/fsg/TCS

But that is just your extra stuff not the whole ib-tests thing is that 
right?

> BTW shouldn't we move all the private directories that now 
> are in the main tree to an extra module?

Yep I was thinking of adding a "personal" module with user 
subdirectories for self storage,

But I still think that we also need one within interbase itself to 
distribute those "little goodies" to do with build etc, before they 
become widely accepted (probably meaning compiling on a win32 platforms ;-).



Cheers

Mark

> Frank
> 

Re: [Firebird-test] Adding ib-tests to CVS

[<Sch...@t-...>]

Mark O'Donohue schrieb:
> 
> Hi Frank,
> 
> When you feel you have a comfortable directory structure for the TCS
> test suite from ib-tests.
> And perhaps after a bit of discussion on structure, we can import it in
> as a new module into firebird.   Then we have it under CVS control and
> can add things to it etc.
> 
> The commands to do it are fairly easy and are explained in:
> 
> http://cvsbook.red-bean.com/

Nice book. I thought this sort of things could be done only
by members of the admin group. 

> 
> Having it as a seperate module from interbase has some advantages.
> 
> There's no rush, but when your ready let us know and we'll see what we can do.

I have seen this after I uploaded the exported tests to sourceforge,
so they are now in interbase/firebird/fsg/TCS

BTW shouldn't we move all the private directories that now 
are in the main tree to an extra module?

Frank

-- 
"Fascinating creature, phoenixes, They can carry immensely heavy loads,
 their tears have healing powers and they make highly faithful pets."  
         - J.K. Rowling

[Firebird-test] Firebird 0.9CS (& SS) pas TCS test suite

[Mark O'D. <mar...@lu...>]

Ok
So here is a message that really went astray, I sent it on the 23/10/00 
but it only went to Tom (Coleman) when it was supposed to go here.

I only found out this was missing when I went hunting back for the results.


Updates from this are that super now passes SQL_EVENT_13 - I haven't 
changed anything - I think it just needed a little time.


Anyway enjoy - original title was "FirebirdCS passed (SS nearly so)".  ;-).

Mark



Hi All

Well, the linux firebird 0.9 classic release passed through the TCS test
suite (ib-tests) without any problems.  So we haven't buggered things up
there too much, as yet ;-).

The firebird 0.9 super release passed through just about everything
except possibly the test SQL_EVENT_13 and a number of tests where things
were marked as read only.

I did a ^C to kill test SQL_EVENT_13 as the script was taking some time 
.  I may have killed this test prematurely.  since the
output log seems to make some sense.

So a few tests to have a look at. I've attached the logs for both runs
(compressed they are only 2K each), so if anyone knows better than me
what they mean, id appreciate it.


One amusing thing is that the meta test suite VECTOR_12HOUR only takes
about 20 minutes to run.  Im not sure if this means that cpu's are a lot
faster these days, ... or the number of test cases in it has been
reduced quite a bit, or im missing something.

Anyway its good to have something to bounce our changes off.

So I suppose we need to get a README organised, and figure out if we are
aiming for a  Firebird 1.0 or 6.01.

I can put that one up as a questionare alongside the one about
reindenting the source code ;-).

Anyway, feedback on the installs or tests results would be welcome.

Cheers

Mark

-- 
Your database needs YOU!
http://sourceforge.net/projects/firebird

[Firebird-test] Testing Pats gbak patch.

[Mark O'D. <mar...@lu...>]

Hi Pat

Sorry for taking so long to test your patch on super server, but Im 
having some problems with the build.  I was reluctant to rm -rf and 
start again, which is my normal procedure since I'd applied some of the 
patches manually.   But Im back to that point.

Im back onto it now.

Cheers

Mark

-- 
Your database needs YOU!
http://sourceforge.net/projects/firebird

[Firebird-test] Adding ib-tests to CVS

[Mark O'D. <mar...@lu...>]

Hi Frank,

When you feel you have a comfortable directory structure for the TCS 
test suite from ib-tests. 
And perhaps after a bit of discussion on structure, we can import it in 
as a new module into firebird.   Then we have it under CVS control and 
can add things to it etc.

The commands to do it are fairly easy and are explained in:

http://cvsbook.red-bean.com/

Having it as a seperate module from interbase has some advantages.

There's no rush, but when your ready let us know and we'll see what we can do.


Cheers

Mark


-- 
Your database needs YOU!
http://sourceforge.net/projects/firebird

Re: [Firebird-test] What does this mean?

[Ann H. <har...@ne...>]

>>Mark O'Donohue wrote:
>
>I presume these are the tests then that require a good cleanup
>and rethink anyway.  I assume the chances of Inprise doing this
>are about zero.

I don't know what Inprise is likely to do.  I do know that the
crashme test that MySQL has would be a good candidate to include
in our suite.

>It's interesting, if this was one of the main cost items, I
>wonder how they would plan to go in selling it or use of it.

Price is what a buyer will pay.  The cost to Inprise of
licensing the test suite was much less of an issue than
the value of the tests to us.

>How would you rate it, I would guess that the TCS suite
>gets rid of the obvious bugs, but how much value does
>running the QA suite add to say running TCS on all
>client/server platforms, and putting in some volume/stress tests.

None.  If we run both the development & production builds
against what Inprise has released and create some sort of
multi-user high contention test, Firebird releases should
be at least as stable as Inprise releases.  We can pick
up and validate crashme and the other open source database
tests that are kicking around.

The fact that there was no test in the nightly runs or
the QA suite that tried to create a second generator page
suggests something of the completeness of those tests.

Regards,

Ann

[Firebird-test] Automatic mailings

[<Sch...@t-...>]

Hi all,
is it somehow possible to suppress the automatic Firebird-checkins
mail when adding new files to the tree, or to reduce it
to only mail the names and not the content of the files.

I think of adding the tests to the tree in the near
future and I don't want them all queue up in my mailbox, 
and probably nobody else would want to see them in
the mail.

Frank

-- 
"Fascinating creature, phoenixes, They can carry immensely heavy loads,
 their tears have healing powers and they make highly faithful pets."  
         - J.K. Rowling

[Firebird-test] Bulk import and export into TCS

[<Sch...@t-...>]

Hi all,
I have placed some files in firebird/fsg/TCS to allow 
import and export of complete test databases in plain
text format. (You will need IBPerl -0.8p3 from www.karwin.com
to use this.)

I gave up the attempt to integrate this into tcs.e
(learned a lot of GPRE and c though) because
this would have resulted in changes
in the tcs source tree and the database layout.
BTW I'm impressed what can be done in perl
in a relatively short time.

I would like you to test this stuff.
If you like it, we could use this to place the
complete test database under cvs control, so 
that we can share a common test pool without
mailing every new test.


cheers 
Frank

-- 
"Fascinating creature, phoenixes, They can carry immensely heavy loads,
 their tears have healing powers and they make highly faithful pets."  
         - J.K. Rowling

[Firebird-test] Simple utiliity to load tests into TCS

[Griffin, P. J. <gr...@su...>]

In case anyone is interested, and I'm sure there are better ways, here is
how I load tests into TCS on my AIX unix box:

$ cat ~/bin/load_test.ksh 
#!/bin/ksh
set -x
function load
{
        export EDITOR="cat $1 >"
        {
        echo dt $1
        echo ct $1
        if [[ -n $2 ]] ; then
            echo dts $1 $2
            echo at $1 $2
        fi
        echo quit
        } | tcs
}

series=""

while getopts ":s:" opt; do
    case $opt in
        s )     series=$OPTARG ;;
        \? )    print "usage: load_test.ksh [-s [series name]] files..."
    esac
done
shift $(($OPTIND -1))

for filename in "$@" ; do
    load $filename $series
done



...pat

[Firebird-test] Re: [Firebird-admin] What does this mean?

[Mark O'D. <mar...@lu...>]

Ann Harrison wrote:

> At 07:55 AM 10/25/2000 +1100, Mark O'Donohue wrote:
> 
>>> It appears that we have the nightly test stuff which belonged to 
>>> engineering
>>> and generally worked.
>> 
>> 
>> Thanks for that, so this is kind of a smoke test, So what is in this 
>> other
>> stuff that takes months to run?
> 
> 
> The suite used by the QA group was full of test and result errors -
> bad design and bad initialization.  Each of those tests was run once
> for each server/client platform pair.  Each failure was researched and
> documented.  So, a test which printed 'today' in date format would be
> analyzed a dozen times, and each time reported as an error in test
> design - "test does not return a consistent answer."  Each analysis
> required reading the test, running it manually, reading the expected
> result, comparing it with the actual result both of the automated test
> and the manual test, and writing a little report.
> 
>> I presume a lot of it is manual, which explains the time required to
>> run it, and probably explains Inprises reluctance to run it again.
> 
> 
> There are some manual tests and some load tests that require special
> hardware.  You don't want to know my real opinion about all this.


I presume these are the tests then that require a good cleanup and 
rethink anyway.  I assume the chances of Inprise doing this are about zero. 

Do you think Inprise will ever invest time in rerunning this QA test 
suite again?

It's interesting, if this was one of the main cost items, I wonder how 
they would plan to go in selling it or use of it.  It would appear that 
if it requires a lot of effort to implement then they wouldn't be able 
to sell it for too much.  I suppose if Jim is right and they effectively 
dump IB as soon as their current contracts wind up, then they'll try and 
sell it then.

How would you rate it, I would guess that the TCS suite gets rid of the 
obvious bugs, but how much value does running the QA suite add to say 
running TCS on all client/server platforms, and putting in some 
volume/stress tests.


Cheers

Mark

[Firebird-test] TCS installation for linux

[<Sch...@t-...>]

Hi,
I have uploaded my updated installation for the test control system to 
the Firebird FTP area.

ftp://firebird.sourceforge.net/pub/firebird/TCS/TCS_linux.tar.gz

Included is a script to get a working test control system with one 
keystroke, updated tests for version 6.x servers and some files that
are necessary to run the tests.

Most of the stuff should work on other platforms as well.

happy testing

Frank

-- 
"Fascinating creature, phoenixes, They can carry immensely heavy loads,
 their tears have healing powers and they make highly faithful pets."  
         - J.K. Rowling

[Firebird-test] New test: BUG_117138

[<Sch...@t-...>]

Hi,
here is another test, be careful, it will crash the engine.

I will have no time the next two weeks to look further into
it, so here is what I have found out so far:

The test will output the following plan if it lives long enough:

PLAN SORT (MERGE (SORT (MG NATURAL)(FM MAINMENU NATURAL)
PLAN JOIN (FM S NATURAL1,FM M INDEX (RDB$FOREIGN6))
PLAN ,è÷¿çè*F

dsql/dsql.c:get_rsb_item apparently tries to read behind the end of
of the explain buffer to get a relation name. 

I seems to me that there is a missing end in the buffer returned
by isc_request_info, but I'm not sure about this.

Here is the buffer and my interpretation:
 
  1  BEGIN       
  2    TYPE      
  6    SORT      
  1    BEGIN     
  2      TYPE    
  8      BOOLEAN 
  1      BEGIN   
  2        TYPE     
 11        MERGE    
  2        COUNT 2         
  1        BEGIN   
  2          TYPE  
  6          SORT  
  1          BEGIN 
  3            RELATION 
  2            LENGTH 2 
 77            M
 71            G
  2            TYPE 
  4            SEQUENTIAL 
  0          END  
  0        END             
  1        BEGIN   
  2          TYPE   
  6          SORT   
  1          BEGIN   
  2            TYPE 
  9            UNION 	
  2            COUNT 2 
  1            BEGIN 
  3              RELATION
 11              LENGTH 11 
 70              F
 77              M
 32            
 77              M
 65              A
 73              I
 78              N
 77              M
 69              E
 78              N
 85              U
  2              TYPE
  4              SEQUENTIAL
  0            END
  1            BEGIN 
  2              TYPE
  5              CROSS 
  2              COUNT 2 
  1              BEGIN 
  3                RELATION
  4                LENGTH 4
 70                F
 77                M
 32            
 83                S
  2                TYPE
  4                SEQUENTIAL 
  0              END
  1              BEGIN
  2                TYPE
  8                BOOLEAN
  1                BEGIN 
  3                  RELATION
  4                  LENGTH 4
 70                  F
 77                  M
 32            
 77                  M
  2                  TYPE
  2                  INDEXED
  4                  INDEX 
 12                  LENGTH 12
 82                  R
 68                  D
 66                  B
 36                  $
 70                  F
 79                  O
 82                  R
 69                  E
 73                  I
 71                  G
 78                  N
 54                  6
  0                END
  0              END
  0            END 
  0          END 
  0        END 
  0      END 
  0    END 
     ^here should be another END



cheers 
Frank

-----------------------------------

/*
**  Name:               BUG_117138
**  Series:             BUGS
**  Meta-Series:
**  Dependencies:       None
**  Description:        This test will kill IB
**  Expected result:    See source
**
**  Author:             FSG
**  Bugs:               This test crashes the server.
**  Change history:
**
*/
$ CREATE cf_test.sql
create database "WHERE_GDB:cf.gdb";

CREATE DOMAIN D_GLOBAL_ID AS VARCHAR(15) NOT NULL ;
CREATE DOMAIN D_LONG_DESC AS VARCHAR(200);
CREATE DOMAIN D_GROUP AS INTEGER  DEFAULT 0 CHECK ((value is not null));

CREATE DOMAIN D_GLOBAL_REF AS VARCHAR(15);
CREATE DOMAIN D_ICON AS SMALLINT CHECK (((value is null) or (value
between 0 and 8)));

CREATE TABLE KNOWLEDGESTREAMS (
        STREAM_ID D_GLOBAL_ID NOT NULL,
        NAME D_LONG_DESC,
        CONTENT_GROUPS D_GROUP,
        CONSTRAINT PK_KNOWLEDGESTREAMS PRIMARY KEY (STREAM_ID)
);

CREATE TABLE MAINMENU (
        MENU_ID D_GLOBAL_ID NOT NULL,
        PARENT_ID D_GLOBAL_REF,
        DESCRIPTION D_LONG_DESC,
        CONTENT_GROUP D_GROUP NOT NULL,
        ICON D_ICON,
        CONSTRAINT PK_MAINMENU PRIMARY KEY (MENU_ID)
);

ALTER TABLE MAINMENU ADD CONSTRAINT FK_MAINMENU FOREIGN KEY (PARENT_ID)
REFERENCES MAINMENU(MENU_ID) ON DELETE CASCADE ON UPDATE CASCADE;

CREATE TABLE MENU_GROUPS (
        MENU_ID D_GLOBAL_ID NOT NULL,
        CONTENT_ID D_GLOBAL_ID NOT NULL
);

CREATE INDEX MENU_GROUPS_IDX1 ON MENU_GROUPS (MENU_ID);
CREATE INDEX MENU_GROUPS_IDX2 ON MENU_GROUPS (CONTENT_ID);

CREATE TABLE STREAMMENU (
        STREAM_ID D_GLOBAL_ID NOT NULL,
        PARENT D_GLOBAL_ID NOT NULL,
        CONSTRAINT PK_STREAMMENU PRIMARY KEY (PARENT, STREAM_ID)
);

ALTER TABLE STREAMMENU ADD CONSTRAINT FK_STREAMMENU_PARENT FOREIGN KEY
(PARENT) REFERENCES MAINMENU(MENU_ID) ON DELETE CASCADE;

ALTER TABLE STREAMMENU ADD CONSTRAINT FK_STREAMMENU_STREAM_ID FOREIGN
KEY (STREAM_ID) REFERENCES KNOWLEDGESTREAMS(STREAM_ID) ON DELETE
CASCADE;

CREATE VIEW FULLMENU (
    CODE,
    PARENT,
    DESCRIPTION,
    LINK,
    CONTENT_GROUP
) AS
select menu_id,parent_id,description,cast(null as
varchar(100)),content_group from mainmenu
union all
select m.stream_id, m.parent, s.name
,cast('/servlets/uk.co.wmeng.intelus.KnowledgeStream?ACTION=DISPLAY&ID='
|| s.stream_id as varchar(100)),content_groups from streammenu m join
knowledgestreams s on s.stream_id = m.stream_id
;

/* And now IB will die and take the TCS with it */

set plan;
Select distinct fm.code, fm.Description, fm.link
from fullmenu fm join menu_groups mg on fm.code = mg.menu_id;

exit;

$ ISQL -e -m -input cf_test.sql
$ DELETE cf_test.sql
$ RUN drop_gdb WHERE_GDB:cf.gdb

[Firebird-test] New test: GBAK_NON-ENGLISH_01

[Griffin, P. J. <gr...@su...>]

I don't know if the win1251 characters will make it across the list, but I'm
testing a potential patch to fix [ Bug #112177 ] error with non-English
column defaults.

Please consider this new test  GBAK_NON-ENGLISH_01

/*
**  Name:               gbak_non-english_01
**  Series:             cf_isql
**  Meta-Series:
**  Dependencies:       None
**  Description:        
**
**  Expected result:    GBAK shouldn't abort.
**                      
**  Author:             Patrick J. P. Griffin based on cf_isql_10 by Ravil
A. Desai
**  Bugs:               [ Bug #112177 ] error with non-english column
defaults
**  Change history:
**
*/

$ CREATE cf_test.sql
set names win1251;
create database "WHERE_GDB:cf.gdb";

create table bug_bugbase ( 
s varchar(20) character set win1251 default 'RôbértNélsôn' not null collate
win1251,
i smallint default '1234' not null
); 

insert into bug_bugbase (s) values('BruceNélsôn');
insert into bug_bugbase (s) values('narf');
insert into bug_bugbase (i) values(4321);

exit;

$ CREATE cf_test2.sql
connect "WHERE_GDB:cf.gdb";
select * from bug_bugbase;
exit;

$ CREATE cf_test3.sql
set names win1251;
connect "WHERE_GDB:cf.gdb";

insert into bug_bugbase (i) values(1);
insert into bug_bugbase (i) values(2);
insert into bug_bugbase (i) values(3);

exit;

$ ISQL -m -input cf_test.sql
$ GBAK -b WHERE_GDB:cf.gdb cf.gbak
$ RUN drop_gdb WHERE_GDB:cf.gdb
$ GBAK -r cf.gbak WHERE_GDB:cf.gdb
$ ISQL -e -m -input cf_test2.sql
$ ISQL -e -m -input cf_test3.sql
$ ISQL -e -m -input cf_test2.sql
$ RUN drop_gdb WHERE_GDB:cf.gdb
$ DELETE cf_test.sql cf_test2.sql cf_test3.sql cf.gbak





Thanks,
...pat

Re: [Firebird-test] TCS

[<Sch...@t-...>]

Mark O'Donohue schrieb:
> 
> Hi Frank (and Pat ? and whoever).
> 
> Yep the TCS run works, I got your download and run through the
> instructions and I got it going on the installs that I have built.
> 
> The blob edit works as well.
> 
> One hickup I had was that it got stuck on some of the tests.  I did two
> things then,
> 
> initially I didn't have a .tcs_config file so I created one from your
> sample output as:
> 
>       Example of the contents of .tcs_config:
>        ----------------------------------------
>        svr 6.0.0                 (set version run)
>        sb bp_fb                (set boiler plate)
>        se qa_fb               (set environment)
>        srn FIREBIRD              (set run name)


Seems that I have forgotten this in my install script 

> 
> And I did an update on the ib-test to make sure I had the latest, (I got
> my copy quite early on, and some extra stuff came down).  Then it ran
> fine.  Im not sure which was particularly needed, although there was
> some comments about version < 4.0000 or something in the old logs.
> 
> Anyway both seemed reasonable things to do.
> 
> So thanks for the script, I expect that since as you say, your getting
> no errors in a VECTOR_12HOUR run, that sometime soon I should be able to
> say the same about mine - no problems so far.
> 
> The first run actually finished with no errors, but it took only about
> 20 minutes, Does the VECTOR_12HOUR actually take 12hr, or is this a cut
> down version.?

This depends on the machine you're running the tests on :-)
Try it on a 386sx.
 
> 
> Now I'll do the super server version.
> 
> I've attached the run trace it case anyone wants to have a look.

Looks as expected. I had to set one of the charset tests to read only,
because of a missing script.
I had not time yet to look at the other RO tests.

cheers 
Frank

-- 
"Fascinating creature, phoenixes, They can carry immensely heavy loads,
 their tears have healing powers and they make highly faithful pets."  
         - J.K. Rowling

Re: [Firebird-test] TCS

[Mark O'D. <mar...@lu...>]

Hi Frank (and Pat ? and whoever).

Yep the TCS run works, I got your download and run through the 
instructions and I got it going on the installs that I have built.

The blob edit works as well.

One hickup I had was that it got stuck on some of the tests.  I did two 
things then,

initially I didn't have a .tcs_config file so I created one from your 
sample output as:


      Example of the contents of .tcs_config:
       ----------------------------------------
       svr 6.0.0                 (set version run)
       sb bp_fb                (set boiler plate)
       se qa_fb               (set environment)
       srn FIREBIRD              (set run name)

And I did an update on the ib-test to make sure I had the latest, (I got 
my copy quite early on, and some extra stuff came down).  Then it ran 
fine.  Im not sure which was particularly needed, although there was 
some comments about version < 4.0000 or something in the old logs.

Anyway both seemed reasonable things to do.

So thanks for the script, I expect that since as you say, your getting 
no errors in a VECTOR_12HOUR run, that sometime soon I should be able to 
say the same about mine - no problems so far. 

The first run actually finished with no errors, but it took only about 
20 minutes, Does the VECTOR_12HOUR actually take 12hr, or is this a cut 
down version.?

Now I'll do the super server version.

I've attached the run trace it case anyone wants to have a look.

Cheers

Mark




Sch...@t-... wrote:

> I'm not sure whether my previous reply made it to the list.
> I have placed the updated databases and an installation
> to create a running TCS on linux from scratch at
> http://home.t-online.de/home/schlottmann-goedde/download/TCS_linux.tar.gz
> 
> Most of the notes have gone into the install script, but there
> is something left in the README.
> 
> Frank
> 

Re: [Firebird-test] TCS

[Mark O'D. <mar...@lu...>]

Sch...@t-... wrote:

>> Im keen to read your notes, and if it is pretty much automated I may be able
>> to fit it in before the weekend.
> 
> I'm not sure whether my previous reply made it to the list.
> I have placed the updated databases and an installation
> to create a running TCS on linux from scratch at
> http://home.t-online.de/home/schlottmann-goedde/download/TCS_linux.tar.gz

It's funny that I had just about an hour ago gone and looked this up, 
since I'd assumed it was ftp://firebird.sourceforge.net the firest time.

I have a copy, and something that repeatably installs a linux 
SuperServer, and am just checking the diffs between my version (about a 
week or so ago) and the current changes, mostly netbsd, before I go on 
into running the tester.

Cheers & many thanks.

Mark

Re: [Firebird-test] TCS

[<Sch...@t-...>]

Sorry, I have sent this only to Mark, so here it is again 
for the list.

Mark O'Donohue schrieb:
> 
>  Hi Frank
> 
> If you still don't have any takers by the week end, I'll do it.
> 
> Im keen to read your notes, and if it is pretty much automated I may be able
> to fit it in before the weekend.

I'm not sure whether my previous reply made it to the list.
I have placed the updated databases and an installation
to create a running TCS on linux from scratch at
http://home.t-online.de/home/schlottmann-goedde/download/TCS_linux.tar.gz

Most of the notes have gone into the install script, but there
is something left in the README.

Frank

-- 
"Fascinating creature, phoenixes, They can carry immensely heavy loads,
 their tears have healing powers and they make highly faithful pets."  
         - J.K. Rowling

RE: [Firebird-test] Please try this new test

[Patrick J. P. G. <Pat...@wo...>]

Frank,

Thanks for trying those two combinations.

As I thought more and more about the current firebird build I became
convinced that it was only a coincidence that it didn't abort.  The internal
generator ID was being passed to the next routine as if it was a valid
pointer - but in this case the pointer field is used to hold that ID.

I've sent in a change that prevents this.

...pat
My family sits at dinner
and the conversation gets thinner.
My thoughts are over yonder
for on this C code I do ponder.



 -----Original Message-----
From: 	fir...@li...
[mailto:fir...@li...]  On Behalf Of Frank
Schlottmann-Goedde
Sent:	Thursday, October 19, 2000 4:14 AM
To:	fir...@li...
Subject:	Re: [Firebird-test] Please try this new test

"Patrick J. P. Griffin" schrieb:
>
> Hello!
>
> I've been contemplating a potential bug in jrd/cmp.c
>
> If you can, would you please enter the test shown below and see if you get
> the correct results.
> ...pat

Here is the result against an original Inprise IB60SS:

Running global test CF_ISQL_17 ... *** failed ****
Test CF_ISQL_17 failed on 19-OCT-2000:
        initialized by frank
        on 19-OCT-2000 with boilerplate BP_FB
-------------------
59,61 c 59,61
<            4                  1228
<            1                  3428
<
---------------------------------------------
>            4                     4
>            1                    20
>
++++++++++++++++++++++++++++++++++++++++++++++++++
107,110 c 107,111
<
<            4                  1004
4205                  6630
<            1                  1007
6216                 11840
<
---------------------------------------------
> Statement failed, SQLCODE = -902
>
> database file appears corrupt ()
> -wrong page type
> -page 0 is of wrong type (expected 9, found 1)
++++++++++++++++++++++++++++++++++++++++++++++++++

-------------------

Frank
_______________________________________________
Firebird-test mailing list
Fir...@li...
http://lists.sourceforge.net/mailman/listinfo/firebird-test

Re: [Firebird-test] Please try this new test

[<Sch...@t-...>]

"Patrick J. P. Griffin" schrieb:
> 
> Hello!
> 
> I've been contemplating a potential bug in jrd/cmp.c
> 
> If you can, would you please enter the test shown below and see if you get
> the correct results.
> ...pat

Here is the result against an original Inprise IB60SS:

Running global test CF_ISQL_17 ... *** failed ****
Test CF_ISQL_17 failed on 19-OCT-2000:
        initialized by frank
        on 19-OCT-2000 with boilerplate BP_FB
-------------------
59,61 c 59,61
<            4                  1228
<            1                  3428
<
---------------------------------------------
>            4                     4
>            1                    20
>
++++++++++++++++++++++++++++++++++++++++++++++++++
107,110 c 107,111
<
<            4                  1004                 
4205                  6630
<            1                  1007                 
6216                 11840
<
---------------------------------------------
> Statement failed, SQLCODE = -902
>
> database file appears corrupt ()
> -wrong page type
> -page 0 is of wrong type (expected 9, found 1)
++++++++++++++++++++++++++++++++++++++++++++++++++

-------------------

Frank