Couple of things pushed to the repos
Yanick Champoux
Just to keep peeps informed, today I was able to push a few things
to the repos:
1. a 00versions.t test that, first thing, give us the OCI client as the
Oracle server versions:
$ perl -I blib/arch/ -I blib/lib/ t/00versions.t
1..2
# OCI client library version: 11.2.0.1
ok 1
# database version: Oracle Database 10g Express Edition Release
10.2.0.1.0 - Product
ok 2
This is going to help a little bit to triage the cpantester reports.
2. I've realize that the tests were gagging diag() messages behind the
scene. I've removed that, and plan on moving the extra verbiage to
note()s, so that everything is more in line with common test usage.
3. I've discovered that our Makefile.PL supports arguments. I've
documented them, and now they are printed via Pod::Usage.
4. I've tried to reproduce
https://rt.cpan.org/Ticket/Display.html?id=13865, but with the test I've
built
(https://github.com/yanick/DBD-Oracle/commit/6a44e054059a4ae86505aecc9bbbe4a89c08a3f9),
everything seems to be fine. Can a second pair of eyes look at it to
see if I'm missing anything?
5. I've tried to reproduce
https://rt.cpan.org/Ticket/Display.html?id=69350 (the only bug that is
preventing the release of 1.30), but in my case, it's all working. Darn. :-/
Joy,
`/anick
Martin J. Evans
> Hi all,
>
> Just to keep peeps informed, today I was able to push a few things to
> the repos:
>
> 1. a 00versions.t test that, first thing, give us the OCI client as the
> Oracle server versions:
>
> $ perl -I blib/arch/ -I blib/lib/ t/00versions.t
> 1..2
> # OCI client library version: 11.2.0.1
> ok 1
> # database version: Oracle Database 10g Express Edition Release
> 10.2.0.1.0 - Product
> ok 2
Not that I am complaining but I believe this is output during perl
Makefile.PL when running sqlplus too.
>
> This is going to help a little bit to triage the cpantester reports.
>
> 2. I've realize that the tests were gagging diag() messages behind the
> scene. I've removed that, and plan on moving the extra verbiage to
> note()s, so that everything is more in line with common test usage.
>
> 3. I've discovered that our Makefile.PL supports arguments. I've
> documented them, and now they are printed via Pod::Usage.
>
> 4. I've tried to reproduce
> https://rt.cpan.org/Ticket/Display.html?id=13865, but with the test I've
> built
> (https://github.com/yanick/DBD-Oracle/commit/6a44e054059a4ae86505aecc9bbbe4a89c08a3f9),
> everything seems to be fine. Can a second pair of eyes look at it to see
> if I'm missing anything?
>
> 5. I've tried to reproduce
> https://rt.cpan.org/Ticket/Display.html?id=69350 (the only bug that is
> preventing the release of 1.30), but in my case, it's all working. Darn.
> :-/
I believe this fails for me but I'm seriously strapped for time to look
in to it right now. I could perhaps make a server available which
exhibits the problem.
Martin
--
Martin J. Evans
Wetherby, UK
Yanick Champoux
>> 1. a 00versions.t test that, first thing, give us the OCI client as the
>> Oracle server versions:
>>
> Makefile.PL when running sqlplus too.
It is. But Makefile.PL's output is a little.. shall we say...
intimidating in its prolixity? And you need to run it to get its output.
Which is not a big deal, but not as simple as to ask people to do:
prove -b t/00versions.t t/30lob.t
to have all version numbers *and* the test outputs you desire in a
single swoop. :-)
>> 5. I've tried to reproduce
>> https://rt.cpan.org/Ticket/Display.html?id=69350 (the only bug that is
>> preventing the release of 1.30), but in my case, it's all working. Darn.
>> :-/
>
> I believe this fails for me
\o/
> but I'm seriously strapped for time to look
> in to it right now. I could perhaps make a server available which
> exhibits the problem.
Okay, I'll see if I have a brilliant mind to throw at the problem.
Which architecture/version # are you using, btw?
Joy,
`/anick
Martin J. Evans
> On 11-07-28 02:46 PM, Martin J. Evans wrote:
>>> 1. a 00versions.t test that, first thing, give us the OCI client as the
>>> Oracle server versions:
>>>
>> Not that I am complaining but I believe this is output during perl
>> Makefile.PL when running sqlplus too.
>
> It is. But Makefile.PL's output is a little.. shall we say...
> intimidating in its prolixity? And you need to run it to get its output.
> Which is not a big deal, but not as simple as to ask people to do:
>
> prove -b t/00versions.t t/30lob.t
But you need to run perl Makefile.Pl before you run prove on t/30lib.t
so I still don't get how it is "not as simple".
> to have all version numbers *and* the test outputs you desire in a
> single swoop. :-)
given, that is a different issue to the one I was pointing out. It is my
experience though that generally test output without Makefile.PL output
is misleading at best. Perhaps you might find this also in the course of
time.
>>> 5. I've tried to reproduce
>>> https://rt.cpan.org/Ticket/Display.html?id=69350 (the only bug that is
>>> preventing the release of 1.30), but in my case, it's all working. Darn.
>>> :-/
>>
>> I believe this fails for me
>
> \o/
>
> > but I'm seriously strapped for time to look
>> in to it right now. I could perhaps make a server available which
>> exhibits the problem.
>
> Okay, I'll see if I have a brilliant mind to throw at the problem. Which
> architecture/version # are you using, btw?
>
> Joy,
> `/anick
>
>
>
Linux ubuntu client running 11.2 instant client.
Linux centos Oracle Server running 11.1.
Yanick Champoux
> On 29/07/2011 01:18, Yanick Champoux wrote:
>> On 11-07-28 02:46 PM, Martin J. Evans wrote:
>>>> 1. a 00versions.t test that, first thing, give us the OCI client as
>>>> the
>>>> Oracle server versions:
>>>>
>>> Not that I am complaining but I believe this is output during perl
>>> Makefile.PL when running sqlplus too.
>>
>> It is. But Makefile.PL's output is a little.. shall we say...
>> intimidating in its prolixity? And you need to run it to get its output.
>> Which is not a big deal, but not as simple as to ask people to do:
>>
>> prove -b t/00versions.t t/30lob.t
>
> But you need to run perl Makefile.Pl before you run prove on t/30lib.t
> so I still don't get how it is "not as simple".
You don't necessary run Makefile.PL. You might want to run the
tests without recompiling the whole shebang. Basically, it's just
decoupling diagnostic and compilation, which are two things that, while
we often do them in tandem, can legitimately be done on their own. As I
said, it's not a huge thing, but it can be useful. Trust me. :-)
Joy,
`/anick
--
The best compliment you could give Pythian for our service is a referral.