python-sybase-misc

[Python-sybase-misc] new SQLAlchemy dialect; Python 3?

[Michael B. <mi...@zz...>]

Hi there -

I just wanted to give the developers a heads up that SQLAlchemy 0.6 will
feature a new Sybase ASE dialect that defaults to using Python-Sybase as
its backend.  I finally got a test ASE server up over the weekend and
managed to get a rudimental Sybase conversation going on with this driver.

Was also wondering about the development schedule of the driver, I see the
last release was in 2008.  Are there any plans to add Python 3 support to
this driver ?    When installing via Distribute, if the setup.py file
specifies a "2to3" flag then Distribute will automatically run the source
code through the 2to3 tool before installing.   I've now created Python 3
versions of three different libraries (Mako, Beaker, SQLAlchemy) in this
way and its pretty straightforward.

thanks for your efforts on this driver !

- mike

Re: [Python-sybase-misc] new SQLAlchemy dialect; Python 3?

[Sébastien S. <sa...@us...>]

Hi Mic,

great to hear about that SQLAlchemy will support Sybase.

You should note that python-sybase is unfortunately not perfectly DBAPI
compliant because of some limitations of sybase which make it hard to
perfectly follow the specification (problems with nested requests
notably). If you encounter such problem, and can reproduce it with a
unitary test, I would be happy to help correct it in python-sybase.

Also please note that the trunk has quite a lot of bug corrections
compared to the last release, so I would suggest to use it rather than
the latest official release (I should take some time to make a new release).

The development of python-sybase is rather slow at the moment; I work on
it for the needs of a legacy project at my work, but Sybase is not the
priority at the moment and so I mostly do maintenance and small bug
corrections. I didn't have time to make a proper release for a long
time, though I hope to have more free time for that this summer.

At some stage we will have to get the legacy project on which I work to
support sybase in the latest version. When that will happen I will be
able to dedicate a full month or 2 to improve python-sybase, mostly
concerning performances (and unicode support). I should be able to also
check Python 3 compatibility  at the same time. But in the meantime, I
can only ensure small bugs corrections and answering to emails.

regards

--
Sébastien Sablé



Michael Bayer a écrit :
> Hi there -
>
> I just wanted to give the developers a heads up that SQLAlchemy 0.6 will
> feature a new Sybase ASE dialect that defaults to using Python-Sybase as
> its backend.  I finally got a test ASE server up over the weekend and
> managed to get a rudimental Sybase conversation going on with this driver.
>
> Was also wondering about the development schedule of the driver, I see the
> last release was in 2008.  Are there any plans to add Python 3 support to
> this driver ?    When installing via Distribute, if the setup.py file
> specifies a "2to3" flag then Distribute will automatically run the source
> code through the 2to3 tool before installing.   I've now created Python 3
> versions of three different libraries (Mako, Beaker, SQLAlchemy) in this
> way and its pretty straightforward.
>
> thanks for your efforts on this driver !
>
> - mike
>
> ------------------------------------------------------------------------------
> Download Intel® Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Python-sybase-misc mailing list
> Pyt...@li...
> https://lists.sourceforge.net/lists/listinfo/python-sybase-misc
>   

Re: [Python-sybase-misc] new SQLAlchemy dialect; Python 3?

[Skip M. <sk...@po...>]

Thanks for the reply, Sébastien. As I only have another couple days left at
my current job, I am unlikely to do much more with it. It still builds and
seems to run okay on Python 2. I get an early core dump from the Python 3
code. Valgrind complains about ctx_alloc I think. I don't have that output
in front of me at the moment. I'll pass along an error message tomorrow.

I did a checkout from the Subversion repo, so hopefully I am using the
latest code. I do notice that it still says "0.40" though.

Skip


On Tue, Apr 20, 2010 at 9:12 AM, Sébastien Sablé <
sa...@us...> wrote:

> Hi Mic,
>
> great to hear about that SQLAlchemy will support Sybase.
>
> You should note that python-sybase is unfortunately not perfectly DBAPI
> compliant because of some limitations of sybase which make it hard to
> perfectly follow the specification (problems with nested requests
> notably). If you encounter such problem, and can reproduce it with a
> unitary test, I would be happy to help correct it in python-sybase.
>
> Also please note that the trunk has quite a lot of bug corrections
> compared to the last release, so I would suggest to use it rather than
> the latest official release (I should take some time to make a new
> release).
>
> The development of python-sybase is rather slow at the moment; I work on
> it for the needs of a legacy project at my work, but Sybase is not the
> priority at the moment and so I mostly do maintenance and small bug
> corrections. I didn't have time to make a proper release for a long
> time, though I hope to have more free time for that this summer.
>
> At some stage we will have to get the legacy project on which I work to
> support sybase in the latest version. When that will happen I will be
> able to dedicate a full month or 2 to improve python-sybase, mostly
> concerning performances (and unicode support). I should be able to also
> check Python 3 compatibility  at the same time. But in the meantime, I
> can only ensure small bugs corrections and answering to emails.
>
> regards
>
> --
> Sébastien Sablé
>
>
>
> Michael Bayer a écrit :
> > Hi there -
> >
> > I just wanted to give the developers a heads up that SQLAlchemy 0.6 will
> > feature a new Sybase ASE dialect that defaults to using Python-Sybase as
> > its backend.  I finally got a test ASE server up over the weekend and
> > managed to get a rudimental Sybase conversation going on with this
> driver.
> >
> > Was also wondering about the development schedule of the driver, I see
> the
> > last release was in 2008.  Are there any plans to add Python 3 support to
> > this driver ?    When installing via Distribute, if the setup.py file
> > specifies a "2to3" flag then Distribute will automatically run the source
> > code through the 2to3 tool before installing.   I've now created Python 3
> > versions of three different libraries (Mako, Beaker, SQLAlchemy) in this
> > way and its pretty straightforward.
> >
> > thanks for your efforts on this driver !
> >
> > - mike
> >
> > ------------------------------------------------------------
> ------------------
> > Download Intel&#174; Parallel Studio Eval
> > Try the new software tools for yourself. Speed compiling, find bugs
> > proactively, and fine-tune applications for parallel performance.
> > See why Intel Parallel Studio got high marks during beta.
> > http://p.sf.net/sfu/intel-sw-dev
> > _______________________________________________
> > Python-sybase-misc mailing list
> > Pyt...@li...
> > https://lists.sourceforge.net/lists/listinfo/python-sybase-misc
> >
>
>
> ------------------------------------------------------------
> ------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Python-sybase-misc mailing list
> Pyt...@li...
> https://lists.sourceforge.net/lists/listinfo/python-sybase-misc
>

Re: [Python-sybase-misc] new SQLAlchemy dialect; Python 3?

[Skip M. <sk...@po...>]

On Wed, Oct 5, 2016 at 4:56 PM, Skip Montanaro <sk...@po...> wrote:
>
> Valgrind complains about ctx_alloc I think. I don't have that output in front of me at the moment. I'll pass along an error message tomorrow.


Here's one example. All seem to be similar, reading off the end of a
malloc'd chunk of memory. I just ran valgrind over the Py2 version and
got no such read errors, so I suspect my problems are caused by issues
related to structure sizes. Just to make sure it wasn't related to
configuring my Python build using --with-pydebug, I rebuilt from
scratch  without it, and get the same errors. I might well have messed
something up in the conversion of the module init code. I'll take a
look at whatever explicit allocation the code does.

Skip

==28701== Invalid read of size 4
==28701==    at 0xCF223A9: com__err_load_errors (in
/opt/local/encap/sybase15_7-15.7.110+1/sybase-15.7/OCS-15_0/lib/libsybcomn_r64.so)
==28701==    by 0xCF231A4: com_err_init (in
/opt/local/encap/sybase15_7-15.7.110+1/sybase-15.7/OCS-15_0/lib/libsybcomn_r64.so)
==28701==    by 0xCCE37F4: cs__errinit (in
/opt/local/encap/sybase15_7-15.7.110+1/sybase-15.7/OCS-15_0/lib/libsybcs_r64.so)
==28701==    by 0xCCDFFEC: cs_ctx_alloc (in
/opt/local/encap/sybase15_7-15.7.110+1/sybase-15.7/OCS-15_0/lib/libsybcs_r64.so)
==28701==    by 0xC3DBD42: ctx_alloc (ctx.c:1425)
==28701==    by 0xC3E3A59: sybasect_cs_ctx_alloc (sybasect.c:177)
==28701==    by 0x4D7ECF: PyCFunction_Call (methodobject.c:109)
==28701==    by 0x5DB7CB: call_function (ceval.c:4723)
==28701==    by 0x5D3982: PyEval_EvalFrameEx (ceval.c:3277)
==28701==    by 0x5D86FD: _PyEval_EvalCodeWithName (ceval.c:4036)
==28701==    by 0x5D8864: PyEval_EvalCodeEx (ceval.c:4057)
==28701==    by 0x5BF14B: PyEval_EvalCode (ceval.c:777)
==28701==  Address 0x70afcc4 is 4 bytes inside a block of size 6 alloc'd
==28701==    at 0x4C2ABED: malloc (in
/usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==28701==    by 0xCF221DD: com__err_load_errors (in
/opt/local/encap/sybase15_7-15.7.110+1/sybase-15.7/OCS-15_0/lib/libsybcomn_r64.so)
==28701==    by 0xCF231A4: com_err_init (in
/opt/local/encap/sybase15_7-15.7.110+1/sybase-15.7/OCS-15_0/lib/libsybcomn_r64.so)
==28701==    by 0xCCE37F4: cs__errinit (in
/opt/local/encap/sybase15_7-15.7.110+1/sybase-15.7/OCS-15_0/lib/libsybcs_r64.so)
==28701==    by 0xCCDFFEC: cs_ctx_alloc (in
/opt/local/encap/sybase15_7-15.7.110+1/sybase-15.7/OCS-15_0/lib/libsybcs_r64.so)
==28701==    by 0xC3DBD42: ctx_alloc (ctx.c:1425)
==28701==    by 0xC3E3A59: sybasect_cs_ctx_alloc (sybasect.c:177)
==28701==    by 0x4D7ECF: PyCFunction_Call (methodobject.c:109)
==28701==    by 0x5DB7CB: call_function (ceval.c:4723)
==28701==    by 0x5D3982: PyEval_EvalFrameEx (ceval.c:3277)
==28701==    by 0x5D86FD: _PyEval_EvalCodeWithName (ceval.c:4036)
==28701==    by 0x5D8864: PyEval_EvalCodeEx (ceval.c:4057)