Обсуждение: Remove "unsupported version" warning for Greenplum

Hello,

we tested pgAdminIII 1.22 internally and found that it works sufficient
with Greenplum. A few minor problems are identified and we plan to work
on them as well, but nothing serious. We however know that there are a
number more serious problems with HAWQ (SQL on Hadoop), we plan to work
on this as well.

As a first step, attached is a patch which removes the "unsupported
version" warning for Greenplum. The patch copies a bit infrastructure
and prepares us to support later versions of Greenplum when we bump our
version number after merging with later PostgreSQL versions. One
remaining question is if the supported versions should stay in
pgConn.cpp or move to pgAdmin3.h.


Thank you,

--
                Andreas 'ads' Scherbaum
German PostgreSQL User Group
European PostgreSQL User Group - Board of Directors
Volunteer Regional Contact, Germany - PostgreSQL Project

Hi

On Mon, Jan 18, 2016 at 1:44 AM, Andreas 'ads' Scherbaum
<adsmail@wars-nicht.de> wrote:
>
> Hello,
>
> we tested pgAdminIII 1.22 internally and found that it works sufficient with
> Greenplum. A few minor problems are identified and we plan to work on them
> as well, but nothing serious. We however know that there are a number more
> serious problems with HAWQ (SQL on Hadoop), we plan to work on this as well.
>
> As a first step, attached is a patch which removes the "unsupported version"
> warning for Greenplum. The patch copies a bit infrastructure and prepares us
> to support later versions of Greenplum when we bump our version number after
> merging with later PostgreSQL versions. One remaining question is if the
> supported versions should stay in pgConn.cpp or move to pgAdmin3.h.

Your patch won't apply again. I have no idea why - I'm trying to do it
on my Mac, which is a *nix under the hood (they don't use Mac line
endings any more - that was the old Mac OS 9 and earlier from a decade
or so ago iirc). How are you creating them? The normal way is to do
something like:

git pull --rebase
git diff > foo.diff

My git config for the repo is basically what you get on a fresh
installation of git, with a git clone, e.g.

[core]
        repositoryformatversion = 0
        filemode = true
        bare = false
        logallrefupdates = true
        ignorecase = true
        precomposeunicode = true
[remote "origin"]
        url = ssh://git@git.postgresql.org/pgadmin3.git
        fetch = +refs/heads/*:refs/remotes/origin/*
[branch "master"]
        remote = origin
        merge = refs/heads/master

Anyway - patch issues aside, isn't your patch essentially the same as
the one I was looking for feedback on just before release?


http://www.postgresql.org/message-id/flat/CACjysiUQLnddN6n=NoB_0YTmnLb9SxaSn1P9SCWNbu3f09k7Gg@mail.gmail.com#CACjysiUQLnddN6n=NoB_0YTmnLb9SxaSn1P9SCWNbu3f09k7Gg@mail.gmail.com


--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake

EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

On 19.01.2016 16:03, Dave Page wrote:
> On Mon, Jan 18, 2016 at 1:44 AM, Andreas 'ads' Scherbaum
> <adsmail@wars-nicht.de> wrote:
>>
>> Hello,
>>
>> we tested pgAdminIII 1.22 internally and found that it works sufficient with
>> Greenplum. A few minor problems are identified and we plan to work on them
>> as well, but nothing serious. We however know that there are a number more
>> serious problems with HAWQ (SQL on Hadoop), we plan to work on this as well.
>>
>> As a first step, attached is a patch which removes the "unsupported version"
>> warning for Greenplum. The patch copies a bit infrastructure and prepares us
>> to support later versions of Greenplum when we bump our version number after
>> merging with later PostgreSQL versions. One remaining question is if the
>> supported versions should stay in pgConn.cpp or move to pgAdmin3.h.
>
> Your patch won't apply again. I have no idea why - I'm trying to do it
> on my Mac, which is a *nix under the hood (they don't use Mac line
> endings any more - that was the old Mac OS 9 and earlier from a decade
> or so ago iirc). How are you creating them? The normal way is to do
> something like:

That is strange. Can someone else on the list please check if the patch
applies (and then roll it back)? I did a fresh clone (again) and it
applies clean.



> Anyway - patch issues aside, isn't your patch essentially the same as
> the one I was looking for feedback on just before release?
>
>
http://www.postgresql.org/message-id/flat/CACjysiUQLnddN6n=NoB_0YTmnLb9SxaSn1P9SCWNbu3f09k7Gg@mail.gmail.com#CACjysiUQLnddN6n=NoB_0YTmnLb9SxaSn1P9SCWNbu3f09k7Gg@mail.gmail.com

I clearly have missed this one. And I see that you prefer the version in
the header file.

Also I plan to add extra support for HAWQ (SQL on Hadoop, also based on
the Greenplum engine), but this has a few issues.

Ok, let me please rework this patch, and do another check for the git
config.


Regards,

--
                Andreas 'ads' Scherbaum
German PostgreSQL User Group
European PostgreSQL User Group - Board of Directors
Volunteer Regional Contact, Germany - PostgreSQL Project

On 19.01.2016 16:03, Dave Page wrote:
>
> Your patch won't apply again. I have no idea why - I'm trying to do it
> on my Mac, which is a *nix under the hood (they don't use Mac line
> endings any more - that was the old Mac OS 9 and earlier from a decade
> or so ago iirc). How are you creating them? The normal way is to do
> something like:

And how does the attached work? Fresh clone again, only difference is a
warning (not an error) for whitespaces removed.


Regards

--
                Andreas 'ads' Scherbaum
German PostgreSQL User Group
European PostgreSQL User Group - Board of Directors
Volunteer Regional Contact, Germany - PostgreSQL Project

On Tue, Jan 19, 2016 at 10:54 PM, Andreas 'ads' Scherbaum
<adsmail@wars-nicht.de> wrote:
> On 19.01.2016 16:03, Dave Page wrote:
>>
>>
>> Your patch won't apply again. I have no idea why - I'm trying to do it
>> on my Mac, which is a *nix under the hood (they don't use Mac line
>> endings any more - that was the old Mac OS 9 and earlier from a decade
>> or so ago iirc). How are you creating them? The normal way is to do
>> something like:
>
>
> And how does the attached work? Fresh clone again, only difference is a
> warning (not an error) for whitespaces removed.

Still doesn't apply. I tried on the following systems:

Mac OS X 10.11.1 - git version 2.5.4 (Apple Git-61)
Windows 7 Enterprise SP1 - git version 1.8.1.msysgit.1
CentOS release 6.7 (Final) - git version 1.7.1

I'm fairly convinced at this stage that there's something funky on
your system. Perhaps we should take a look next week when we're both
in Brussels?

--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake

EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

On 21.01.2016 10:31, Dave Page wrote:
> On Tue, Jan 19, 2016 at 10:54 PM, Andreas 'ads' Scherbaum
> <adsmail@wars-nicht.de> wrote:
>> On 19.01.2016 16:03, Dave Page wrote:
>>>
>>>
>>> Your patch won't apply again. I have no idea why - I'm trying to do it
>>> on my Mac, which is a *nix under the hood (they don't use Mac line
>>> endings any more - that was the old Mac OS 9 and earlier from a decade
>>> or so ago iirc). How are you creating them? The normal way is to do
>>> something like:
>>
>>
>> And how does the attached work? Fresh clone again, only difference is a
>> warning (not an error) for whitespaces removed.
>
> Still doesn't apply. I tried on the following systems:
>
> Mac OS X 10.11.1 - git version 2.5.4 (Apple Git-61)
> Windows 7 Enterprise SP1 - git version 1.8.1.msysgit.1
> CentOS release 6.7 (Final) - git version 1.7.1
>
> I'm fairly convinced at this stage that there's something funky on
> your system. Perhaps we should take a look next week when we're both
> in Brussels?
>

After debugging back and forth with Magnus, it looks like that Google
Mail is fooling you. Your downloaded file has a different line ending,
and your file is 3269 bytes, where the original file is 3210 bytes.
That's 59 additional line breaks.

4fa0990a1020e425fe95b99ea9f186de  gp-warning2.diff

The file you download from the archive:

http://www.postgresql.org/message-id/569EBEA2.5040903@wars-nicht.de

is also correct.

--
                Andreas 'ads' Scherbaum
German PostgreSQL User Group
European PostgreSQL User Group - Board of Directors
Volunteer Regional Contact, Germany - PostgreSQL Project

On Thu, Jan 21, 2016 at 10:54 PM, Magnus Hagander <magnus@hagander.net> wrote:
>
>
> On Thu, Jan 21, 2016 at 11:35 PM, Andreas 'ads' Scherbaum
> <adsmail@wars-nicht.de> wrote:
>>
>> On 21.01.2016 10:31, Dave Page wrote:
>>>
>>> On Tue, Jan 19, 2016 at 10:54 PM, Andreas 'ads' Scherbaum
>>> <adsmail@wars-nicht.de> wrote:
>>>>
>>>> On 19.01.2016 16:03, Dave Page wrote:
>>>>>
>>>>>
>>>>>
>>>>> Your patch won't apply again. I have no idea why - I'm trying to do it
>>>>> on my Mac, which is a *nix under the hood (they don't use Mac line
>>>>> endings any more - that was the old Mac OS 9 and earlier from a decade
>>>>> or so ago iirc). How are you creating them? The normal way is to do
>>>>> something like:
>>>>
>>>>
>>>>
>>>> And how does the attached work? Fresh clone again, only difference is a
>>>> warning (not an error) for whitespaces removed.
>>>
>>>
>>> Still doesn't apply. I tried on the following systems:
>>>
>>> Mac OS X 10.11.1 - git version 2.5.4 (Apple Git-61)
>>> Windows 7 Enterprise SP1 - git version 1.8.1.msysgit.1
>>> CentOS release 6.7 (Final) - git version 1.7.1
>>>
>>> I'm fairly convinced at this stage that there's something funky on
>>> your system. Perhaps we should take a look next week when we're both
>>> in Brussels?
>>>
>>
>> After debugging back and forth with Magnus, it looks like that Google Mail
>> is fooling you. Your downloaded file has a different line ending, and your
>> file is 3269 bytes, where the original file is 3210 bytes. That's 59
>> additional line breaks.
>>
>> 4fa0990a1020e425fe95b99ea9f186de  gp-warning2.diff
>>
>> The file you download from the archive:
>>
>> http://www.postgresql.org/message-id/569EBEA2.5040903@wars-nicht.de
>>
>> is also correct.

Well that's weird. But why is it only happening with your patches? I
apply patches from others constantly without issues.

> It's also worth noticing that "patch" has no problem with either of the two
> files, but "git apply" does. At least on my system.

Yeah, patch does work for me. I'm so used to using 'git apply' these
days that I didn't think of trying that.

So what am I applying? Does my patch work for you Ads, or do you want
to update yours to move the version numbers?

--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake

EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

On Fri, Jan 22, 2016 at 9:57 AM, Magnus Hagander <magnus@hagander.net> wrote:
>
>
> On Fri, Jan 22, 2016 at 10:43 AM, Dave Page <dpage@pgadmin.org> wrote:
>>
>> On Thu, Jan 21, 2016 at 10:54 PM, Magnus Hagander <magnus@hagander.net>
>> wrote:
>> >
>> >
>> > On Thu, Jan 21, 2016 at 11:35 PM, Andreas 'ads' Scherbaum
>> > <adsmail@wars-nicht.de> wrote:
>> >>
>> >> On 21.01.2016 10:31, Dave Page wrote:
>> >>>
>> >>> On Tue, Jan 19, 2016 at 10:54 PM, Andreas 'ads' Scherbaum
>> >>> <adsmail@wars-nicht.de> wrote:
>> >>>>
>> >>>> On 19.01.2016 16:03, Dave Page wrote:
>> >>>>>
>> >>>>>
>> >>>>>
>> >>>>> Your patch won't apply again. I have no idea why - I'm trying to do
>> >>>>> it
>> >>>>> on my Mac, which is a *nix under the hood (they don't use Mac line
>> >>>>> endings any more - that was the old Mac OS 9 and earlier from a
>> >>>>> decade
>> >>>>> or so ago iirc). How are you creating them? The normal way is to do
>> >>>>> something like:
>> >>>>
>> >>>>
>> >>>>
>> >>>> And how does the attached work? Fresh clone again, only difference is
>> >>>> a
>> >>>> warning (not an error) for whitespaces removed.
>> >>>
>> >>>
>> >>> Still doesn't apply. I tried on the following systems:
>> >>>
>> >>> Mac OS X 10.11.1 - git version 2.5.4 (Apple Git-61)
>> >>> Windows 7 Enterprise SP1 - git version 1.8.1.msysgit.1
>> >>> CentOS release 6.7 (Final) - git version 1.7.1
>> >>>
>> >>> I'm fairly convinced at this stage that there's something funky on
>> >>> your system. Perhaps we should take a look next week when we're both
>> >>> in Brussels?
>> >>>
>> >>
>> >> After debugging back and forth with Magnus, it looks like that Google
>> >> Mail
>> >> is fooling you. Your downloaded file has a different line ending, and
>> >> your
>> >> file is 3269 bytes, where the original file is 3210 bytes. That's 59
>> >> additional line breaks.
>> >>
>> >> 4fa0990a1020e425fe95b99ea9f186de  gp-warning2.diff
>> >>
>> >> The file you download from the archive:
>> >>
>> >> http://www.postgresql.org/message-id/569EBEA2.5040903@wars-nicht.de
>> >>
>> >> is also correct.
>>
>> Well that's weird. But why is it only happening with your patches? I
>> apply patches from others constantly without issues.
>
>
> What MIME types do you get those in typically? Could be that gmail is
> reacting to it being text/x-patch and not application/x-patch or something?

text/x-patch or application/octet-stream seem to be fine from others.
The one thing I noticed is the majority of patches I've received from
others tend to be base64 encoded, whilst Ads' is not.

I'd be interested to try a patch that's been zipped before attachment.

--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake

EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

On 22.01.2016 11:04, Dave Page wrote:
> On Fri, Jan 22, 2016 at 9:57 AM, Magnus Hagander <magnus@hagander.net> wrote:
>>
>>
>> On Fri, Jan 22, 2016 at 10:43 AM, Dave Page <dpage@pgadmin.org> wrote:
>>>
>>> On Thu, Jan 21, 2016 at 10:54 PM, Magnus Hagander <magnus@hagander.net>
>>> wrote:
>>>>
>>>>
>>>> On Thu, Jan 21, 2016 at 11:35 PM, Andreas 'ads' Scherbaum
>>>> <adsmail@wars-nicht.de> wrote:
>>>>>
>>>>> On 21.01.2016 10:31, Dave Page wrote:
>>>>>>
>>>>>> On Tue, Jan 19, 2016 at 10:54 PM, Andreas 'ads' Scherbaum
>>>>>> <adsmail@wars-nicht.de> wrote:
>>>>>>>
>>>>>>> On 19.01.2016 16:03, Dave Page wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> Your patch won't apply again. I have no idea why - I'm trying to do
>>>>>>>> it
>>>>>>>> on my Mac, which is a *nix under the hood (they don't use Mac line
>>>>>>>> endings any more - that was the old Mac OS 9 and earlier from a
>>>>>>>> decade
>>>>>>>> or so ago iirc). How are you creating them? The normal way is to do
>>>>>>>> something like:
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> And how does the attached work? Fresh clone again, only difference is
>>>>>>> a
>>>>>>> warning (not an error) for whitespaces removed.
>>>>>>
>>>>>>
>>>>>> Still doesn't apply. I tried on the following systems:
>>>>>>
>>>>>> Mac OS X 10.11.1 - git version 2.5.4 (Apple Git-61)
>>>>>> Windows 7 Enterprise SP1 - git version 1.8.1.msysgit.1
>>>>>> CentOS release 6.7 (Final) - git version 1.7.1
>>>>>>
>>>>>> I'm fairly convinced at this stage that there's something funky on
>>>>>> your system. Perhaps we should take a look next week when we're both
>>>>>> in Brussels?
>>>>>>
>>>>>
>>>>> After debugging back and forth with Magnus, it looks like that Google
>>>>> Mail
>>>>> is fooling you. Your downloaded file has a different line ending, and
>>>>> your
>>>>> file is 3269 bytes, where the original file is 3210 bytes. That's 59
>>>>> additional line breaks.
>>>>>
>>>>> 4fa0990a1020e425fe95b99ea9f186de  gp-warning2.diff
>>>>>
>>>>> The file you download from the archive:
>>>>>
>>>>> http://www.postgresql.org/message-id/569EBEA2.5040903@wars-nicht.de
>>>>>
>>>>> is also correct.
>>>
>>> Well that's weird. But why is it only happening with your patches? I
>>> apply patches from others constantly without issues.
>>
>>
>> What MIME types do you get those in typically? Could be that gmail is
>> reacting to it being text/x-patch and not application/x-patch or something?
>
> text/x-patch or application/octet-stream seem to be fine from others.
> The one thing I noticed is the majority of patches I've received from
> others tend to be base64 encoded, whilst Ads' is not.
>
> I'd be interested to try a patch that's been zipped before attachment.

Yes, I decided to gzip them next time, before sending them.

Honestly I have no idea why only my patches. There is nothing unusual in
my Thunderbird, pretty much a default installation. And downloading my
patch from Thunderbird or the PG Archive results in the correct file. So
it must be a Google thing.

Ok, now that this problem is solved, I will move the versions into the
header files and send you a new version soon.


Thanks,

--
                Andreas 'ads' Scherbaum
German PostgreSQL User Group
European PostgreSQL User Group - Board of Directors
Volunteer Regional Contact, Germany - PostgreSQL Project

On 18.01.2016 02:44, Andreas 'ads' Scherbaum wrote:
>
> we tested pgAdminIII 1.22 internally and found that it works sufficient
> with Greenplum. A few minor problems are identified and we plan to work
> on them as well, but nothing serious. We however know that there are a
> number more serious problems with HAWQ (SQL on Hadoop), we plan to work
> on this as well.
>
> As a first step, attached is a patch which removes the "unsupported
> version" warning for Greenplum. The patch copies a bit infrastructure
> and prepares us to support later versions of Greenplum when we bump our
> version number after merging with later PostgreSQL versions. One
> remaining question is if the supported versions should stay in
> pgConn.cpp or move to pgAdmin3.h.

Next attempt (this time packed diff).

The attached patch adds recognition for Greenplum HAWQ (SQL on Hadoop),
where I know that it has problems with some catalog tables. An error
message is still shown for HAWQ. The error message during connect for
GPDB is removed. Version information is moved to header file.


Regards,

--
                Andreas 'ads' Scherbaum
German PostgreSQL User Group
European PostgreSQL User Group - Board of Directors
Volunteer Regional Contact, Germany - PostgreSQL Project

On Thu, Feb 4, 2016 at 12:22 PM, Andreas 'ads' Scherbaum
<adsmail@wars-nicht.de> wrote:
> On 18.01.2016 02:44, Andreas 'ads' Scherbaum wrote:
>>
>>
>> we tested pgAdminIII 1.22 internally and found that it works sufficient
>> with Greenplum. A few minor problems are identified and we plan to work
>> on them as well, but nothing serious. We however know that there are a
>> number more serious problems with HAWQ (SQL on Hadoop), we plan to work
>> on this as well.
>>
>> As a first step, attached is a patch which removes the "unsupported
>> version" warning for Greenplum. The patch copies a bit infrastructure
>> and prepares us to support later versions of Greenplum when we bump our
>> version number after merging with later PostgreSQL versions. One
>> remaining question is if the supported versions should stay in
>> pgConn.cpp or move to pgAdmin3.h.
>
>
> Next attempt (this time packed diff).

And this time it works perfectly :-)

> The attached patch adds recognition for Greenplum HAWQ (SQL on Hadoop),
> where I know that it has problems with some catalog tables. An error message
> is still shown for HAWQ. The error message during connect for GPDB is
> removed. Version information is moved to header file.

Thanks - applied to the 1.22 and master branches.

--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake

EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

On 08.02.2016 11:23, Dave Page wrote:
> On Thu, Feb 4, 2016 at 12:22 PM, Andreas 'ads' Scherbaum
> <adsmail@wars-nicht.de> wrote:
>
>> The attached patch adds recognition for Greenplum HAWQ (SQL on Hadoop),
>> where I know that it has problems with some catalog tables. An error message
>> is still shown for HAWQ. The error message during connect for GPDB is
>> removed. Version information is moved to header file.
>
> Thanks - applied to the 1.22 and master branches.

Thanks for submitting the patch.

Just that you know (because I've seen your commit message): it will barf
when connecting to HAWQ - because there are catalog errors. The patch
removes the warning for Greenplum Database only. I'm not going to fix
this in III, but keep an eye on it for 4.

--
                Andreas 'ads' Scherbaum
German PostgreSQL User Group
European PostgreSQL User Group - Board of Directors
Volunteer Regional Contact, Germany - PostgreSQL Project