libdbi-devel — Discussion about development of the core library and DBD interface

[libdbi-devel] sqlite and sqlite3 drivers CHAR as TEXT or INTEGER

[J. H. F. <jo...@gm...>]

Hello,

I need a simple clarification:

drivers/sqlite/dbd_sqlite.c

  837   else if (strstr(curr_type, "CHAR") /* this is a 1-byte value */
  838 	   || strstr(curr_type, "TINYINT")
  839 	   || strstr(curr_type, "INT1")) {
  840     type = FIELD_TYPE_TINY;
  841   }

drivers/sqlite3/dbd_sqlite3.c

else if (strstr(curr_type, "CHAR") /* this is a 1-byte value */
 1093 	   || strstr(curr_type, "TINYINT")
 1094 	   || strstr(curr_type, "INT1")) {
 1095     type = FIELD_TYPE_TINY;
 1096   }


But http://sqlite.org/datatypes.html and http://sqlite.org/datatype3.html says:

"The search for these strings in the type declaration is case
insensitive, of course. If any of the above strings occur anywhere in
the type declaration, then the datatype of the column is text. Notice
that the type "VARCHAR" contains "CHAR" as a substring so it is
considered text."

"If the datatype of the column contains any of the strings "CHAR",
"CLOB", or "TEXT" then that column has TEXT affinity. Notice that the
type VARCHAR contains the string "CHAR" and is thus assigned TEXT
affinity."

I think the sqlite and sqlite3 drivers must treat CHAR as TEXT not
INTEGER (TINY).

It's right?

Thanks.

-- 
-----------------------------------------------------------
João Henrique Freitas - joaohf_at_gmail.com
Campinas-SP-Brasil
BSD051283
LPI 1
http://www.joaohfreitas.eti.br

[libdbi-devel] install pkg-config file in libdbi

[Markus H. <mar...@mh...>]

Balazs Scheidler writes:
 > Hi,
 > 
 > this patch makes the source install the pkg-config file. I should probably 
 > send this to libdbi-devel, but I'm not subscribed there and the list server
 > didn't respond to my subscription request yet.
 > 
 > diff -u -r1.24 Makefile.am
 > --- Makefile.am	17 Aug 2008 21:30:40 -0000	1.24
 > +++ Makefile.am	4 Oct 2008 17:18:40 -0000
 > @@ -11,3 +11,6 @@
 >  
 >  profile:
 >  	$(MAKE) all CFLAGS="@PROFILE@"
 > +
 > +pkgconfigdir = $(libdir)/pkgconfig
 > +pkgconfig_DATA = dbi.pc
 > 

I've applied this patch to the sources and checked the new version
into cvs.

regards,
Markus

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

[libdbi-devel] [Libdbi-drivers-devel] [PATCH] fix asprintf portability problem

[Markus H. <mar...@mh...>]

Balazs Scheidler writes:
 > 
 > Hi,
 > 
 > We encountered crashes on solaris-10/amd64 platform, the root cause
 > seems to be that the asprintf implementation in libdbi uses memcpy()
 > to duplicate the argument array.
 > 
 > This is not portable.
 > 
 > The attached patch fixes it. Please apply.
 > 

Applied & checked in.

regards,
Markus


-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

Re: [libdbi-devel] improvements of test cases

[J. H. F. <jo...@gm...>]

Hello,

I finished the conversion of all tests. You can see it:

http://www.joaohfreitas.eti.br/download/tests-libdbi-cgreen-0.9.tar.gz

The design is very simple. More about tests than code design.

Tests can be run all together or single. This is very important if you
are developing one feature or tracking a bug.

I hope this can be help us to ensure the quality of libdbi project.

Thanks.


On Mon, Jun 16, 2008 at 6:29 PM, Markus Hoenicka
<mar...@mh...> wrote:
> João Henrique Freitas writes:
>  > I did test with freebsd 7.0 and ubuntu 8.04. The version of libdbi is
>  > CVS head and no exceptions was report.
>  >
>  > Can you test too?
>  >
>  > I have expected 3 failures and 0 exceptions for all drivers.
>  >
>
> Oh I see. I thought the tests were supposed to succeed. Indeed, your
> latest version does not cause exceptions with any of the drivers that
> I can test. If that means you're on track, then please go ahead with
> the remainder of the tests.
>
> regards,
> Markus
>
> --
> Markus Hoenicka
> mar...@ca...
> (Spam-protected email: replace the quadrupeds with "mhoenicka")
> http://www.mhoenicka.de
>



-- 
-----------------------------------------------------------
João Henrique Freitas - joaohf_at_gmail.com
Campinas-SP-Brasil
BSD051283
LPI 1
http://www.joaohfreitas.eti.br

Re: [libdbi-devel] API suggestions

[Rainer G. <rge...@hq...>]

> - Error handling is very inconsistent. Especially the functions that
>   return 1 (one) on success are *very* uncommon. I'd change 
> that so that
>   *all* functions (that return a status message) return 0 (zero) on
>   success and less than zero upon failure.

While it is good to have consistent error handling, this one breaks all
exitisting applications. Even worse, code would begin to fail by a
simple upgrade of the library. Thus I conclude this is a no-go.

The alternative may be to do a new call right at init, let's say
SetConsistentErrors(ON),which would modify the errors returned. So new
apps could call this and use the consistent error codes. That would
work.

Rainer

[libdbi-devel] API suggestions

[Florian F. <sou...@no...>]

Hi,

I just finished writing the `dbi' plugin for collectd, which uses the
`dbi' library for database access.

Now I have some suggestions towards your interface, but before I get to
work I'd like to get some feedback to know if it's worth investing the
time.


Here it goes. It's in no particular order, just the way it came to mind:

- Error handling is very inconsistent. Especially the functions that
  return 1 (one) on success are *very* uncommon. I'd change that so that
  *all* functions (that return a status message) return 0 (zero) on
  success and less than zero upon failure.

- The `dbi_conn_error' function is weird to use. I'd add a function
  similar to the `strerror_r' function (with GNU extension):
    char *dbi_strerror (dbi_conn c, char *buffer, size_t buffer_size);
  The function will write the error message to the memory pointed to by
  `buffer' (not writing more than `buffer_size' bytes and assuring null-
  termination) and return the `buffer' pointer. This way, error
  reporting is easy and thread-safe (should that ever become an issue):

    status = dbi_foobar (...);
    if (status != 0)
    {
      char errbuf[1024];
      fprintf (stderr, "Function dbi_foobar failed: %s.\n",
          dbi_strerror (conn, errbuf, sizeof (errbuf)));
      return (-1);
    }

- Make `dbi_conn_set_option' return an error if it is passed an unknown
  option. That error should be well defined, so that applications may
  chose to ignore it.

- Add two functions, `dbi_result_get_string_buffer' and
  `dbi_result_get_string_buffer_idx', with the following signatures:

    int dbi_result_get_string_buffer (dbi_result res, const char *name,
        char *buffer, ssize_t *buffer_size);
    int dbi_result_get_string_buffer_idx (dbi_result res, int index,
        char *buffer, ssize_t *buffer_size);

  The functions return 0 (zero) upon success, non-zero otherwise. They
  will write at most `*buffer_size' bytes into `buffer' and will assure
  null-termination of the buffer. The total number of bytes that have
  been written to `buffer' will be stored in `*buffer_size'. If the
  field is NULL, `buffer' will not be touched and a negative value will
  be stored in `*buffer_size'. If anything other than zero is returned,
  neither memory location will be touched.

- Add code to cast data of the DBI_TYPE_INTEGER, DBI_TYPE_DECIMAL and
  DBI_TYPE_STRING types into each other type. The `dbi_result_get_*'
  should be extended to use that functionality.

  In my case I don't care what format the values are in in the database
  and I can live with losing a bit of precision. People who care about
  precision can still check for the type of each column (they need to do
  that now, too, so this doesn't make their situation any worse), but we
  could also introduce a flag that's passed to the `dbi_result_get_*' or
  a configuration option that's associated with the connection or
  something.


Any comments and suggestions are very welcome, of course :)

Regards,
-octo
-- 
Florian octo Forster
Hacker in training
GnuPG: 0x91523C3D
http://verplant.org/

[libdbi-devel] [PATCH] fix asprintf portability problem

[Balazs S. <ba...@ba...>]

Hi,

We encountered crashes on solaris-10/amd64 platform, the root cause
seems to be that the asprintf implementation in libdbi uses memcpy()
to duplicate the argument array.

This is not portable.

The attached patch fixes it. Please apply.

diff -u -r1.4 asprintf.c
--- src/asprintf.c	28 Jul 2005 19:40:48 -0000	1.4
+++ src/asprintf.c	4 Oct 2008 17:30:35 -0000
@@ -14,7 +14,7 @@
 
 #ifndef HAVE_VASPRINTF
 
-int int_vasprintf(char **result, const char *format, va_list *args)
+int vasprintf(char **result, const char *format, va_list args)
 {
         const char *p = format;
   /* Add one to make sure that it is never zero, which might cause malloc
@@ -22,13 +22,13 @@
         int total_width = strlen(format) + 1;
         va_list ap;
 
-        memcpy((char*) &ap, (char*) args, sizeof(va_list));
+        va_copy(ap, args);
 
         while (*p != '\0') {
                 if (*p++ == '%') {
                         while (strchr ("-+ #0", *p)) {
-                ++p;
-            }
+                                ++p;
+                        }
 
                         if (*p == '*') {
                                 ++p;
@@ -46,8 +46,8 @@
                                         total_width += abs(va_arg(ap, int));
                                 }
                                 else {
-                        total_width += (unsigned long) strtol(p, (char**) &p, 10);
-                }
+                                        total_width += (unsigned long) strtol(p, (char**) &p, 10);
+                                }
                         }
          
                         while (strchr ("hlL", *p)) {
@@ -89,20 +89,15 @@
         }
 
         *result = malloc(total_width);
-
+        va_end(ap);
         if (*result != NULL) {
-                return vsprintf(*result, format, *args);
+                return vsprintf(*result, format, args);
         }
         else {
                 return 0;
         }
 }
 
-int vasprintf(char **result, const char *format, va_list args)
-{
-        return int_vasprintf(result, format, &args);
-}
-
 #endif /* !HAVE_VASPRINTF */
 
 #ifndef HAVE_ASPRINTF
Index: src/dbi_main.c
===================================================================
RCS file: /cvsroot/libdbi/libdbi/src/dbi_main.c,v
retrieving revision 1.85
diff -u -r1.85 dbi_main.c
--- src/dbi_main.c	27 May 2008 10:52:12 -0000	1.85
+++ src/dbi_main.c	4 Oct 2008 17:30:36 -0000
@@ -109,7 +109,6 @@
 
 /* declarations of optional external functions */
 #ifndef HAVE_VASPRINTF
-int int_vasprintf(char **result, const char *format, va_list *args);
 int vasprintf(char **result, const char *format, va_list args);
 #endif
 #ifndef HAVE_ASPRINTF


-- 
Bazsi

[libdbi-devel] install pkg-config file in libdbi

[Balazs S. <ba...@ba...>]

Hi,

this patch makes the source install the pkg-config file. I should probably 
send this to libdbi-devel, but I'm not subscribed there and the list server
didn't respond to my subscription request yet.

diff -u -r1.24 Makefile.am
--- Makefile.am	17 Aug 2008 21:30:40 -0000	1.24
+++ Makefile.am	4 Oct 2008 17:18:40 -0000
@@ -11,3 +11,6 @@
 
 profile:
 	$(MAKE) all CFLAGS="@PROFILE@"
+
+pkgconfigdir = $(libdir)/pkgconfig
+pkgconfig_DATA = dbi.pc


-- 
Bazsi

Re: [libdbi-devel] Compiling libdbi for MS Visual Studio

[Markus H. <mar...@mh...>]

Quoting Jordi Allue Cabellos <ja...@gr...>:

>
> Hello,
>
> We are developing an application in Visual studio that need acces to  
>  PostgreSQL and SQLite. It would be nice to use your library in the   
> project. We have the compiled cygwin version but we don't have an   
> import library to use this dll from VS.
> To solve this we've tried to recompile the library with VS.
> To succesfully compile we applyed the following changes:
>
>  1.- Uint -> unsigned int
>  2.- remove unistd.h
>  3.- Add copy dll functions from __MINGW32__ section to a new   
> VisualStudio section.
>  4.- remove dirent.h
>  5.- #  define  __attribute__(x)  /*NOTHING*/ to remove gcc specific  
>  error checking.
>  6.- use Windows filesystem interface for dbi_initialize.
>  7.- #define __func__ __FUNCTION__
>
> Is there any interest to include this in libdbi?
>

Hi,

before including MSVC-specific changes I'd like to find out whether  
MinGW provides a library which is sufficient to link MSVC projects  
against. If not, the required changes can be applied to libdbi.  
However, we'd need a clean patch which provides the necessary changes  
without breaking other OSes. If you can provide such a patch, I'll be  
happy to review it and test it on other platforms.

regards,
Markus


-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

[libdbi-devel] Compiling libdbi for MS Visual Studio

[Jordi A. C. <ja...@gr...>]

Hello,

We are developing an application in Visual studio that need acces to PostgreSQL and SQLite. It would be nice to use your library in the project. We have the compiled cygwin version but we don't have an import library to use this dll from VS. 
To solve this we've tried to recompile the library with VS.
To succesfully compile we applyed the following changes:

 1.- Uint -> unsigned int
 2.- remove unistd.h
 3.- Add copy dll functions from __MINGW32__ section to a new VisualStudio section.
 4.- remove dirent.h
 5.- #  define  __attribute__(x)  /*NOTHING*/ to remove gcc specific error checking.
 6.- use Windows filesystem interface for dbi_initialize.
 7.- #define __func__ __FUNCTION__

Is there any interest to include this in libdbi?

Regards,
Jordi Allué

[libdbi-devel] Patch for libdbi for win32

[Markus H. <mar...@mh...>]

Phil Longstaff writes:
 > The attached patch against the CVS version of libdbi changes 'errno' to
 > 'err_no' in some function prototypes.  On win32, when building with
 > mingw, 'errno' seems to be a #define for (*_errno()), so using 'errno'
 > in a function prototype doesn't work as expected.
 > 

Thanks for reporting and fixing this problem. For the sake of
consistency I've also used "err_no" in dbd_internal_error_handler() in
dbd_helper.c.

regards,
Markus

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

[libdbi-devel] Patch for libdbi for win32

[Phil L. <plo...@ro...>]

The attached patch against the CVS version of libdbi changes 'errno' to
'err_no' in some function prototypes.  On win32, when building with
mingw, 'errno' seems to be a #define for (*_errno()), so using 'errno'
in a function prototype doesn't work as expected.

Phil

[libdbi-devel] [ libdbi-Bugs-2048361 ] dbi_conn_quote_string* functions do not escape quotes

[SourceForge.net <no...@so...>]

Bugs item #2048361, was opened at 2008-08-12 20:38
Message generated for change (Comment added) made by mhoenicka
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=379806&aid=2048361&group_id=23824

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
>Status: Closed
Priority: 5
Private: No
Submitted By: Alexander Amelkin (spirit-rc)
Assigned to: Nobody/Anonymous (nobody)
Summary: dbi_conn_quote_string* functions do not escape quotes

Initial Comment:
This code:
----------------------
char *username="buggyrequest\" and something bad \"";
char *username2;

...

dbi_conn_quote_string_copy(conn, username, &username2);
printf("Username: %s\n");
----------------------

Results in the following printout:

----------------------
Username: "buggyrequest" and something bad ""
----------------------

Thus, the resulting username2 string can't actually be used in a call like

dbi_conn_queryf(conn, "select * from users where username=%s", username2);

and will lead to an injection of a malicious SQL code if the original username is supplied by the user.


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

>Comment By: Markus Hoenicka (mhoenicka)
Date: 2008-08-13 12:22

Message:
Logged In: YES 
user_id=85809
Originator: NO

Thanks for moving the bug report to libdbi-drivers.

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

Comment By: Alexander Amelkin (spirit-rc)
Date: 2008-08-13 10:41

Message:
Logged In: YES 
user_id=412567
Originator: YES

Bug #2049266 has been created for this issue in libdbi-drivers.

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

Comment By: Markus Hoenicka (mhoenicka)
Date: 2008-08-13 08:47

Message:
Logged In: YES 
user_id=85809
Originator: NO

Quoting is handled by each driver individually, as the database engines
support different escaping and quoting rules. In order to follow up on this
we need to know which drivers are affected. Could you please file a similar
bug report at libdbi-drivers (not libdbi) mentioning the affected drivers?
I'll then close this item here.

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

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=379806&aid=2048361&group_id=23824

[libdbi-devel] [ libdbi-Bugs-2048361 ] dbi_conn_quote_string* functions do not escape quotes

[SourceForge.net <no...@so...>]

Bugs item #2048361, was opened at 2008-08-12 22:38
Message generated for change (Comment added) made by spirit-rc
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=379806&aid=2048361&group_id=23824

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Priority: 5
Private: No
Submitted By: Alexander Amelkin (spirit-rc)
Assigned to: Nobody/Anonymous (nobody)
Summary: dbi_conn_quote_string* functions do not escape quotes

Initial Comment:
This code:
----------------------
char *username="buggyrequest\" and something bad \"";
char *username2;

...

dbi_conn_quote_string_copy(conn, username, &username2);
printf("Username: %s\n");
----------------------

Results in the following printout:

----------------------
Username: "buggyrequest" and something bad ""
----------------------

Thus, the resulting username2 string can't actually be used in a call like

dbi_conn_queryf(conn, "select * from users where username=%s", username2);

and will lead to an injection of a malicious SQL code if the original username is supplied by the user.


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

>Comment By: Alexander Amelkin (spirit-rc)
Date: 2008-08-13 12:41

Message:
Logged In: YES 
user_id=412567
Originator: YES

Bug #2049266 has been created for this issue in libdbi-drivers.

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

Comment By: Markus Hoenicka (mhoenicka)
Date: 2008-08-13 10:47

Message:
Logged In: YES 
user_id=85809
Originator: NO

Quoting is handled by each driver individually, as the database engines
support different escaping and quoting rules. In order to follow up on this
we need to know which drivers are affected. Could you please file a similar
bug report at libdbi-drivers (not libdbi) mentioning the affected drivers?
I'll then close this item here.

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

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=379806&aid=2048361&group_id=23824

[libdbi-devel] [ libdbi-Bugs-2048361 ] dbi_conn_quote_string* functions do not escape quotes

[SourceForge.net <no...@so...>]

Bugs item #2048361, was opened at 2008-08-12 20:38
Message generated for change (Comment added) made by mhoenicka
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=379806&aid=2048361&group_id=23824

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Priority: 5
Private: No
Submitted By: Alexander Amelkin (spirit-rc)
Assigned to: Nobody/Anonymous (nobody)
Summary: dbi_conn_quote_string* functions do not escape quotes

Initial Comment:
This code:
----------------------
char *username="buggyrequest\" and something bad \"";
char *username2;

...

dbi_conn_quote_string_copy(conn, username, &username2);
printf("Username: %s\n");
----------------------

Results in the following printout:

----------------------
Username: "buggyrequest" and something bad ""
----------------------

Thus, the resulting username2 string can't actually be used in a call like

dbi_conn_queryf(conn, "select * from users where username=%s", username2);

and will lead to an injection of a malicious SQL code if the original username is supplied by the user.


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

>Comment By: Markus Hoenicka (mhoenicka)
Date: 2008-08-13 08:47

Message:
Logged In: YES 
user_id=85809
Originator: NO

Quoting is handled by each driver individually, as the database engines
support different escaping and quoting rules. In order to follow up on this
we need to know which drivers are affected. Could you please file a similar
bug report at libdbi-drivers (not libdbi) mentioning the affected drivers?
I'll then close this item here.

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

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=379806&aid=2048361&group_id=23824

[libdbi-devel] [ libdbi-Bugs-2048361 ] dbi_conn_quote_string* functions do not escape quotes

[SourceForge.net <no...@so...>]

Bugs item #2048361, was opened at 2008-08-12 22:38
Message generated for change (Tracker Item Submitted) made by Item Submitter
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=379806&aid=2048361&group_id=23824

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Priority: 5
Private: No
Submitted By: Alexander Amelkin (spirit-rc)
Assigned to: Nobody/Anonymous (nobody)
Summary: dbi_conn_quote_string* functions do not escape quotes

Initial Comment:
This code:
----------------------
char *username="buggyrequest\" and something bad \"";
char *username2;

...

dbi_conn_quote_string_copy(conn, username, &username2);
printf("Username: %s\n");
----------------------

Results in the following printout:

----------------------
Username: "buggyrequest" and something bad ""
----------------------

Thus, the resulting username2 string can't actually be used in a call like

dbi_conn_queryf(conn, "select * from users where username=%s", username2);

and will lead to an injection of a malicious SQL code if the original username is supplied by the user.


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

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=379806&aid=2048361&group_id=23824

Re: [libdbi-devel] improvements of test cases

[J. H. F. <jo...@gm...>]

Hello,

I have some news...

I finished many tests of API functions. You can see at
http://joaohff.pbwiki.com/f/tests-libdbi-cgreen-0.5.tar.gz

Now, Cgreen is a shared library and to test you need download
(http://joaohff.pbwiki.com/f/cgreen-1.0-alpha3.tar.gz) and install
(./configure; make; make install) before run the tests. Expect 5 or 6
fails with Mysql, Postgresql, sqlite and sqlite3.

Not tested with FreeTDS, Ingres, mSQL, Oracle.

The firebird has passed in many tests (Run in single mode, -s option).
But I not understand why "DROP DATABASE" didn't work. Why?

I think all tests of dbi_test.c are contemplated in test_dbi_cgreen.c.
My goal are cover all function described in documentation.


PS:
I decided embraces the development of the Cgreen project.


On Mon, Jun 16, 2008 at 8:38 PM, João Henrique Freitas <jo...@gm...> wrote:
> Ok,
>
> I am starting the rest of work.
>
> Thanks
>
> On Mon, Jun 16, 2008 at 5:29 PM, Markus Hoenicka
> <mar...@mh...> wrote:
>> João Henrique Freitas writes:
>>  > I did test with freebsd 7.0 and ubuntu 8.04. The version of libdbi is
>>  > CVS head and no exceptions was report.
>>  >
>>  > Can you test too?
>>  >
>>  > I have expected 3 failures and 0 exceptions for all drivers.
>>  >
>>
>> Oh I see. I thought the tests were supposed to succeed. Indeed, your
>> latest version does not cause exceptions with any of the drivers that
>> I can test. If that means you're on track, then please go ahead with
>> the remainder of the tests.
>>
>> regards,
>> Markus
>>
>> --
>> Markus Hoenicka
>> mar...@ca...
>> (Spam-protected email: replace the quadrupeds with "mhoenicka")
>> http://www.mhoenicka.de
>>
>
>
>
> --
> -----------------------------------------------------------
> João Henrique Freitas - joaohf_at_gmail.com
> Campinas-SP-Brasil
> BSD051283
> LPI 1
> http://joaohf.pbwiki.com
> http://www.livejournal.com/users/joaohf/
>



-- 
-----------------------------------------------------------
João Henrique Freitas - joaohf_at_gmail.com
Campinas-SP-Brasil
BSD051283
LPI 1
http://joaohf.pbwiki.com
http://www.livejournal.com/users/joaohf/

Re: [libdbi-devel] improvements of test cases

[J. H. F. <jo...@gm...>]

Ok,

I am starting the rest of work.

Thanks

On Mon, Jun 16, 2008 at 5:29 PM, Markus Hoenicka
<mar...@mh...> wrote:
> João Henrique Freitas writes:
>  > I did test with freebsd 7.0 and ubuntu 8.04. The version of libdbi is
>  > CVS head and no exceptions was report.
>  >
>  > Can you test too?
>  >
>  > I have expected 3 failures and 0 exceptions for all drivers.
>  >
>
> Oh I see. I thought the tests were supposed to succeed. Indeed, your
> latest version does not cause exceptions with any of the drivers that
> I can test. If that means you're on track, then please go ahead with
> the remainder of the tests.
>
> regards,
> Markus
>
> --
> Markus Hoenicka
> mar...@ca...
> (Spam-protected email: replace the quadrupeds with "mhoenicka")
> http://www.mhoenicka.de
>



-- 
-----------------------------------------------------------
João Henrique Freitas - joaohf_at_gmail.com
Campinas-SP-Brasil
BSD051283
LPI 1
http://joaohf.pbwiki.com
http://www.livejournal.com/users/joaohf/

Re: [libdbi-devel] improvements of test cases

[Markus H. <mar...@mh...>]

João Henrique Freitas writes:
 > I did test with freebsd 7.0 and ubuntu 8.04. The version of libdbi is
 > CVS head and no exceptions was report.
 > 
 > Can you test too?
 > 
 > I have expected 3 failures and 0 exceptions for all drivers.
 > 

Oh I see. I thought the tests were supposed to succeed. Indeed, your
latest version does not cause exceptions with any of the drivers that
I can test. If that means you're on track, then please go ahead with
the remainder of the tests.

regards,
Markus

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

Re: [libdbi-devel] improvements of test cases

[J. H. F. <jo...@gm...>]

Hello.

> I'm afraid there's more work to be done here. Do you have access to
> other platforms to make sure we're not looking at platform-specific
> problems?

I did test with freebsd 7.0 and ubuntu 8.04. The version of libdbi is
CVS head and no exceptions was report.

Can you test too?

I have expected 3 failures and 0 exceptions for all drivers.

If all thinks go well, I can start the migration of all tests cases
and organize them.

http://joaohff.pbwiki.com/f/tests-libdbi-cgreen-0.3.tar.gz

Thanks
-- 
-----------------------------------------------------------
João Henrique Freitas - joaohf_at_gmail.com
Campinas-SP-Brasil
BSD051283
LPI 1
http://joaohf.pbwiki.com
http://www.livejournal.com/users/joaohf/

Re: [libdbi-devel] improvements of test cases

[J. H. F. <jo...@gm...>]

Thanks

> mysql indeed improved slightly, but at this time none of the drivers
> work according to your test suite, although I can confirm that all of
> them do work.
>
> mysql:   6 passes, 6 failures, 0 exceptions.
> pgsql:   9 passes, 3 failures, 0 exceptions.
> sqlite:  7 passes, 3 failures, 1 exception.
> sqlite3: 7 passes, 3 failures, 1 exception.
> firebird:could not connect
>
> The firebird problem is probably due to a bug in the old test program
> which you probably copied as well. I've fixed that bug in cvs.
>
> I'm afraid there's more work to be done here. Do you have access to
> other platforms to make sure we're not looking at platform-specific
> problems?
>

I'll make tests in a virtual FreeBSD and some others linux distros.


>
> regards,
> Markus
>
>
>
> João Henrique Freitas writes:
>  > Hello,
>  >
>  > Now it works.
>  >
>  > http://joaohff.pbwiki.com/f/tests-libdbi-cgreen.tar.gz
>  >
>
> --
> Markus Hoenicka
> mar...@ca...
> (Spam-protected email: replace the quadrupeds with "mhoenicka")
> http://www.mhoenicka.de
>



-- 
-----------------------------------------------------------
João Henrique Freitas - joaohf_at_gmail.com
Campinas-SP-Brasil
BSD051283
LPI 1
http://joaohf.pbwiki.com
http://www.livejournal.com/users/joaohf/

Re: [libdbi-devel] improvements of test cases

[Markus H. <mar...@mh...>]

Hi,

mysql indeed improved slightly, but at this time none of the drivers
work according to your test suite, although I can confirm that all of
them do work.

mysql:   6 passes, 6 failures, 0 exceptions.
pgsql:   9 passes, 3 failures, 0 exceptions.
sqlite:  7 passes, 3 failures, 1 exception.
sqlite3: 7 passes, 3 failures, 1 exception.
firebird:could not connect

The firebird problem is probably due to a bug in the old test program
which you probably copied as well. I've fixed that bug in cvs.

I'm afraid there's more work to be done here. Do you have access to
other platforms to make sure we're not looking at platform-specific
problems?


regards,
Markus



João Henrique Freitas writes:
 > Hello,
 > 
 > Now it works.
 > 
 > http://joaohff.pbwiki.com/f/tests-libdbi-cgreen.tar.gz
 > 

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

Re: [libdbi-devel] improvements of test cases

[J. H. F. <jo...@gm...>]

Hello,

Now it works.

http://joaohff.pbwiki.com/f/tests-libdbi-cgreen.tar.gz

I fix the problem about segment faults and mysql.

Thanks

On Tue, Jun 3, 2008 at 6:58 PM, João Henrique Freitas <jo...@gm...> wrote:
> Hello
>
> On Tue, Jun 3, 2008 at 4:50 PM, Markus Hoenicka
> <mar...@mh...> wrote:
>> Hi,
>>
>> I ran a few tests with the files you provided. The stuff now builds
>> without problems, although the tests/cgreen subdirectory still needs
>> to be integrated into the build system so we don't have to make it
>> separately.
>
> Ok
>
>>
>> Unfortunately the tests bomb out on my system regardless of the driver
>> I use. The old test system runs ok with all drivers that I can test
>> (mysql, pgsql, sqlite, sqlite3, firebird), but your test system
>> claims that three drivers are broken, while two pass without
>> completing a single test:
>>
>> mysql: 0 passes, 0 failures, 7 exceptions (this means the test
>>       *succeeds* ??)
>> pgsql: 10 passes, 3 failures, 0 exceptions
>> sqlite: 7 passes, 3 failures, 1 exception
>> sqlite3: 7 passes, 3 failures, 1 exception
>> firebird: 0 passes, 0 failures, 7 exceptions
>>
>
> One clarifications is necessary:
>  - For every tests, cgreen setup and teardown aenvironment to run the
> test. Example:
>   to execute the test test_dbi_conn_quote_binary_copy() on file
> test_dbi_cgreen_managing_queries.c, cgreen execute the
> create_schema(), execute the test and drop_schema().
>  - Yes, all the tests need a environment to run. See functions like:
> open_driver, close_driver, create_schema() and drop_schema(). With
> this, the system can be slow for few seconds. But the tests is granted
> to be successfully and truly
>  - exceptions is a problem in the tests, and need be investigated.
>
>
> We have:
>  Libdbi test suite:
>      Before and after any test group tests run, setup and teardown is call
>                 Database infrastructure tests
>                       Before and after any tests run, setup and
> teardown is call
>                 Managing Queries
>                       Before and after any tests run, setup and
> teardown is call
>                 .... ( any other group of tests ) ...
>
> The tests can be spited by any logic (database specific functions? or
> other groups)  and single tests can be run alone too.
>
> Currently, I have a problem with mysql driver and now firebird.
> Because all the tests exit with segment fault. Why others drivers not?
>
>> I have to admit that the old test system fails to check a couple of
>> things, e.g. you'll have to compare the input and output values of the
>> retrieval tests manually to spot errors. The failures of your test
>> system may actually find some of these problems and thus report more
>> errors.
>>
>
> I will migrate the old system to cgreen. And we will have a integrated
> framework to tests many drivers in every platform at sometime :).
>
>> I'd still appreciate if we could get this to work, but I'm afraid
>> we're not there yet.
>>
>
> Working on it...
>
> Thanks
>
>> regards,
>> Markus
>>
>> João Henrique Freitas writes:
>>  > Hello,
>>  >
>>  > I am working in the improvements of libdbi tests and need of your
>>  > opinion about the files attached.
>>
>> --
>> Markus Hoenicka
>> mar...@ca...
>> (Spam-protected email: replace the quadrupeds with "mhoenicka")
>> http://www.mhoenicka.de
>>
>
>
>
> --
> -----------------------------------------------------------
> João Henrique Freitas - joaohf_at_gmail.com
> Campinas-SP-Brasil
> BSD051283
> LPI 1
> http://joaohf.pbwiki.com
> http://www.livejournal.com/users/joaohf/
>



-- 
-----------------------------------------------------------
João Henrique Freitas - joaohf_at_gmail.com
Campinas-SP-Brasil
BSD051283
LPI 1
http://joaohf.pbwiki.com
http://www.livejournal.com/users/joaohf/

Re: [libdbi-devel] improvements of test cases

[J. H. F. <jo...@gm...>]

Hello

On Tue, Jun 3, 2008 at 4:50 PM, Markus Hoenicka
<mar...@mh...> wrote:
> Hi,
>
> I ran a few tests with the files you provided. The stuff now builds
> without problems, although the tests/cgreen subdirectory still needs
> to be integrated into the build system so we don't have to make it
> separately.

Ok

>
> Unfortunately the tests bomb out on my system regardless of the driver
> I use. The old test system runs ok with all drivers that I can test
> (mysql, pgsql, sqlite, sqlite3, firebird), but your test system
> claims that three drivers are broken, while two pass without
> completing a single test:
>
> mysql: 0 passes, 0 failures, 7 exceptions (this means the test
>       *succeeds* ??)
> pgsql: 10 passes, 3 failures, 0 exceptions
> sqlite: 7 passes, 3 failures, 1 exception
> sqlite3: 7 passes, 3 failures, 1 exception
> firebird: 0 passes, 0 failures, 7 exceptions
>

One clarifications is necessary:
 - For every tests, cgreen setup and teardown aenvironment to run the
test. Example:
   to execute the test test_dbi_conn_quote_binary_copy() on file
test_dbi_cgreen_managing_queries.c, cgreen execute the
create_schema(), execute the test and drop_schema().
 - Yes, all the tests need a environment to run. See functions like:
open_driver, close_driver, create_schema() and drop_schema(). With
this, the system can be slow for few seconds. But the tests is granted
to be successfully and truly
 - exceptions is a problem in the tests, and need be investigated.


We have:
 Libdbi test suite:
      Before and after any test group tests run, setup and teardown is call
                 Database infrastructure tests
                       Before and after any tests run, setup and
teardown is call
                 Managing Queries
                       Before and after any tests run, setup and
teardown is call
                 .... ( any other group of tests ) ...

The tests can be spited by any logic (database specific functions? or
other groups)  and single tests can be run alone too.

Currently, I have a problem with mysql driver and now firebird.
Because all the tests exit with segment fault. Why others drivers not?

> I have to admit that the old test system fails to check a couple of
> things, e.g. you'll have to compare the input and output values of the
> retrieval tests manually to spot errors. The failures of your test
> system may actually find some of these problems and thus report more
> errors.
>

I will migrate the old system to cgreen. And we will have a integrated
framework to tests many drivers in every platform at sometime :).

> I'd still appreciate if we could get this to work, but I'm afraid
> we're not there yet.
>

Working on it...

Thanks

> regards,
> Markus
>
> João Henrique Freitas writes:
>  > Hello,
>  >
>  > I am working in the improvements of libdbi tests and need of your
>  > opinion about the files attached.
>
> --
> Markus Hoenicka
> mar...@ca...
> (Spam-protected email: replace the quadrupeds with "mhoenicka")
> http://www.mhoenicka.de
>



-- 
-----------------------------------------------------------
João Henrique Freitas - joaohf_at_gmail.com
Campinas-SP-Brasil
BSD051283
LPI 1
http://joaohf.pbwiki.com
http://www.livejournal.com/users/joaohf/

[libdbi-devel] improvements of test cases

[Markus H. <mar...@mh...>]

Hi,

I ran a few tests with the files you provided. The stuff now builds
without problems, although the tests/cgreen subdirectory still needs
to be integrated into the build system so we don't have to make it
separately.

Unfortunately the tests bomb out on my system regardless of the driver
I use. The old test system runs ok with all drivers that I can test
(mysql, pgsql, sqlite, sqlite3, firebird), but your test system
claims that three drivers are broken, while two pass without
completing a single test:

mysql: 0 passes, 0 failures, 7 exceptions (this means the test
       *succeeds* ??)
pgsql: 10 passes, 3 failures, 0 exceptions
sqlite: 7 passes, 3 failures, 1 exception
sqlite3: 7 passes, 3 failures, 1 exception
firebird: 0 passes, 0 failures, 7 exceptions

I have to admit that the old test system fails to check a couple of
things, e.g. you'll have to compare the input and output values of the
retrieval tests manually to spot errors. The failures of your test
system may actually find some of these problems and thus report more
errors.

I'd still appreciate if we could get this to work, but I'm afraid
we're not there yet.

regards,
Markus

João Henrique Freitas writes:
 > Hello,
 > 
 > I am working in the improvements of libdbi tests and need of your
 > opinion about the files attached.

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de