phppgadmin-devel — For Developers Only

[ppa-dev] Patches to roll to 9.3

[Karl O. P. <ko...@me...>]

Hi,

I've a new branch, rollto93, that rolls ppa to 9.3
in terms of the structure of the programs in the
filesystem (/help and class/database/ stuff) without
actually changing any functionality.

It's based off the revhelp branch, excepting the
last commit.  The 9.3 help system uses the newly
introduced static id targets in the 9.3 pg source.
(Which will probably not be rebuilt and appear
on-line until some hours have passed.)  So,
the help system links will work for 9.3 and
will show less rot in the future.

https://github.com/kpinc/phppgadmin/commits/rollto93

Regards,


Karl <ko...@me...>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

[ppa-dev] Enhance selenium reset

[Karl O. P. <ko...@me...>]

Hi,

Just pushed another fix/enhancement to my branch which
resets the selenium tests.  Fixes fail of reset if
tests abort when the altered db exists.

https://github.com/kpinc/phppgadmin/commits/selenium-enabled

Karl <ko...@me...>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

[ppa-dev] Fwd: Re: [HACKERS] Make targets of doc links used by phpPgAdmin static [peter_e@gmx.net]

[Karl O. P. <ko...@me...>]

We now have static links for the help system.  Yay!


Karl <ko...@me...>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

[ppa-dev] Display portions of the interface that will fail?

[Karl O. P. <ko...@me...>]

Re: irc below

I've got it in my head that the interface should only display
an action when the user has the ability to carry out the action.
We use the $data->has* tests to avoid displaying parts of the
interface that don't apply to older pg releases, why shouldn't
the interface also pay attention to permissions and anything
else that can used to tell that the operation is not permitted
-- and if not permitted then not display the interface?

So, I test for insert permission before displaying the
import tab for tables, permission and insert-ness on
views likewise.  I plan to carry this through the
display interface in general.  E.g. do not display
a delete button if the user has no delete permission.

Now, the issue with views that arises is: what about rules?
I've avoided thinking about rules entirely; they are
depreciated in favor of triggers and we won't be worse
off then we are now in terms of functionality if rules
are ignored.  On the other hand perhaps there's an easy
way to tell if there's an instead rule on insert; my casual
glance hasn't found one.   So, if there's an easy way
to test against rules, fine we can do that.  Otherwise
I'm planning to ignore them when it comes to determining
whether an operation will succeed or fail.

Thoughts?

[16:04:20]<karlpinc> Humm.  I'd not paid attention to the foreign key 
part.  I  
                     always make a view for that sort of thing and put 
a        
                     insert trigger on it.  Sadly the stock ppa does 
not        
                     import into a view.  However I've some patches 
that do     
                     so.  I'm re-working them now for re-submission to 
ppa --   
                     mostly to add ppa self-tests to test the changes 
but I've  
                     tweaked things in a way that breaks backward       
       
                     compatibility  with older pg releases so I need a 
teeny    
                     fix for
          <karlpinc> that too.  I'm in the process of writing these 
patches     
                     now.  I've got the self tests done but when I went 
to do   
                     backward compatibity I started shaving yaks.  I'm 
back on  
                     track now.

[17:22:50]<xzi11a> that seems odd to me. i wouldnt think an admin 
interface     
                   would care about how the insert against the view is 
carried  
                   out 


Karl <ko...@me...>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

[ppa-dev] Fixes to the help system

[Karl O. P. <ko...@me...>]

Hi,

In preparation for supporting the next PG release
I've cleaned up the help system.  There are a
series of patches found at:

https://github.com/kpinc/phppgadmin/commits/revhelp

These patches:

Reverse the include() hierarchy in the ppa help
system, the stuff in help/.  This is similar to what
was done long ago in classes/database/.

Fixes all the broken help links.  Probably.  I did
a lot of testing but did not click every link
of every supported PG release.
This should be a lot better than what we've got
now even if there are some errors left.

The last patch adds some comments that, hopefully,
specify the link targets to be used in the PG 9.3 docs.

It was very annoying to have to re-discover the
automatically generated ids within the PG html
pages as they changed between PG releases.  To solve
this problem for the future I have submitted a patch
to PG which will make static the ids used as link
targets by the PPA help system.  This patch needs
a reviewer.  Please consider reviewing it so that
these changes get into PG and we greatly reduce
the link-rot in the PPA help system.  The patch
can be found in the PG commitfest system and
that system must be used to review the work.

https://commitfest.postgresql.org/action/patch_view?id=1107

The comments in the last patch in the ppa revhelp
branch document the static links introduced in the PG
patch.

I'd like to build on the revhelp patch series moving forward
with supporting some of the PG 9.3 (and 9.1 and 9.2) view
features so I anxiously await resolution on the
patches I'm submitting.

Regards,

Karl <ko...@me...>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

Re: [ppa-dev] Views, materalized views, and tables

[Karl O. P. <ko...@me...>]

On 05/01/2013 09:44:04 AM, Robert Treat wrote:
> Honestly I think i would object to this. Tables/Views(updatable or
> materialized  are different things, and they do have some different
> properties. While I do think we should be adding in support for
> updatable views and materialized views,  I think this probably
> violates the principal of least surprise.

I would argue the opposite.  Every one of my users has difficulty
distinguishing between views and tables, which is sort of the
point.  Of course a dba needs to know the difference, but
a dba can tell he difference.

> I think we should probably take a look at how psql handles some of
> this stuff; for example I think \dv will list all views, materialized
> or not.

The counter argument is to look at the way pg handles this stuff.
The system catalogs keep views and tables and materialized views
all in pg_class.

> We need to make sure there is some way to differentiate 
> those,
> because the view page should be smart enough to recognize what it is
> dealing with, and then present the proper options (ie. no "refresh"
> option on a regular view, only on a materialized view)

Right.  Just as some views are, say, deletable (and some tables
are not deletable because the user does not have permission)
sometimes a "delete" button should appear in the interface and
sometimes not.  Likewise, sometimes a "refresh" button will
appear when it's a materialized view.

I agree that tables, views, materialized views must be distinguishable,
and easily.  Something I forgot to mention was that I was also thinking
about putting checkboxes across the top of the "relations" tab page;
one checkbox each for table, view, materialized view.  
And a refresh button to make the checkboxes work (although this
could be js).  This would
let the user see and interact with whatever combination of table-like
entities is of interest.  That way you get fine-grained control when
you want it and an overview by default.
> 
> No matter how this works out, we're going to need to draw up some
> lines around this for older versions; a first step would be to set up
> 9.3 support in the database class hierarchy, and to maybe map out 
> some
> of the feature flags we'll be needing for the interface.

Yes, there will have to be a 9.2 class created as part of this process.


> 
> 
> Robert Treat
> 
> 
> On Wed, May 1, 2013 at 9:38 AM, Karl O. Pinc <ko...@me...> wrote:
> > Hi,
> >
> > Since pg-dev now has materialized views and views
> > support insert/update/delete
> > views are, in some sense, no longer much different from
> > tables.  ("Simple" views automatically
> > insert/update/delete in pg-dev, more complex ones require
> > triggers.)
> >
> > My plan is to
> > get rid of separate tabs for tables and views
> > and put views, materialized views and tables
> > under a single tab called "relations".  They will be
> > distinguished by an icon in the left hand
> > column (with mouseover alt= text of
> > view/materialized view/table).
> > We already have icons for tables and views
> > and I've made another (all 3 attached) that combines
> > the table and view icon to represent materialized
> > views.  Buttons will only appear supporting
> > a specific data-involving operation, e.g. delete, when the
> > user has permissions to perform the operation
> > and when the underlying relation supports
> > the operation.
> >
> > Of course views, materialized views, and tables
> > will continue to have different pages for
> > creation/modification.
> >
> > Objections?  Discussion?  I'd like to get to this
> > today.
> >
> >
> > Karl <ko...@me...>
> > Free Software:  "You don't pay back, you pay forward."
> >                  -- Robert A. Heinlein
> >
> > P.S.  Speaking of icons, should we be keeping icon "source"
> > as gimp .xcf files (or something) somewhere?  The materialized
> > view icon was done with layers, but this goes away when
> > exported as a png.
> >
> >
> > 
> ------------------------------------------------------------------------------
> > Introducing AppDynamics Lite, a free troubleshooting tool for
> Java/.NET
> > Get 100% visibility into your production application - at no cost.
> > Code-level diagnostics for performance bottlenecks with <2% 
> overhead
> > Download for free and get started troubleshooting in minutes.
> > http://p.sf.net/sfu/appdyn_d2d_ap1
> > _______________________________________________
> > phpPgAdmin-devel mailing list
> > php...@li...
> > https://lists.sourceforge.net/lists/listinfo/phppgadmin-devel
> >
> 
> ------------------------------------------------------------------------------
> Introducing AppDynamics Lite, a free troubleshooting tool for
> Java/.NET
> Get 100% visibility into your production application - at no cost.
> Code-level diagnostics for performance bottlenecks with <2% overhead
> Download for free and get started troubleshooting in minutes.
> http://p.sf.net/sfu/appdyn_d2d_ap1
> _______________________________________________
> phpPgAdmin-devel mailing list
> php...@li...
> https://lists.sourceforge.net/lists/listinfo/phppgadmin-devel
> 
> 




Karl <ko...@me...>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

Re: [ppa-dev] Views, materalized views, and tables

[Robert T. <ro...@xz...>]

Honestly I think i would object to this. Tables/Views(updatable or
materialized  are different things, and they do have some different
properties. While I do think we should be adding in support for
updatable views and materialized views,  I think this probably
violates the principal of least surprise.

I think we should probably take a look at how psql handles some of
this stuff; for example I think \dv will list all views, materialized
or not. We need to make sure there is some way to differentiate those,
because the view page should be smart enough to recognize what it is
dealing with, and then present the proper options (ie. no "refresh"
option on a regular view, only on a materialized view)

No matter how this works out, we're going to need to draw up some
lines around this for older versions; a first step would be to set up
9.3 support in the database class hierarchy, and to maybe map out some
of the feature flags we'll be needing for the interface.


Robert Treat


On Wed, May 1, 2013 at 9:38 AM, Karl O. Pinc <ko...@me...> wrote:
> Hi,
>
> Since pg-dev now has materialized views and views
> support insert/update/delete
> views are, in some sense, no longer much different from
> tables.  ("Simple" views automatically
> insert/update/delete in pg-dev, more complex ones require
> triggers.)
>
> My plan is to
> get rid of separate tabs for tables and views
> and put views, materialized views and tables
> under a single tab called "relations".  They will be
> distinguished by an icon in the left hand
> column (with mouseover alt= text of
> view/materialized view/table).
> We already have icons for tables and views
> and I've made another (all 3 attached) that combines
> the table and view icon to represent materialized
> views.  Buttons will only appear supporting
> a specific data-involving operation, e.g. delete, when the
> user has permissions to perform the operation
> and when the underlying relation supports
> the operation.
>
> Of course views, materialized views, and tables
> will continue to have different pages for
> creation/modification.
>
> Objections?  Discussion?  I'd like to get to this
> today.
>
>
> Karl <ko...@me...>
> Free Software:  "You don't pay back, you pay forward."
>                  -- Robert A. Heinlein
>
> P.S.  Speaking of icons, should we be keeping icon "source"
> as gimp .xcf files (or something) somewhere?  The materialized
> view icon was done with layers, but this goes away when
> exported as a png.
>
>
> ------------------------------------------------------------------------------
> Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET
> Get 100% visibility into your production application - at no cost.
> Code-level diagnostics for performance bottlenecks with <2% overhead
> Download for free and get started troubleshooting in minutes.
> http://p.sf.net/sfu/appdyn_d2d_ap1
> _______________________________________________
> phpPgAdmin-devel mailing list
> php...@li...
> https://lists.sourceforge.net/lists/listinfo/phppgadmin-devel
>

[ppa-dev] Views, materalized views, and tables

[Karl O. P. <ko...@me...>]

Hi,

Since pg-dev now has materialized views and views
support insert/update/delete 
views are, in some sense, no longer much different from 
tables.  ("Simple" views automatically
insert/update/delete in pg-dev, more complex ones require
triggers.)

My plan is to 
get rid of separate tabs for tables and views
and put views, materialized views and tables
under a single tab called "relations".  They will be
distinguished by an icon in the left hand
column (with mouseover alt= text of
view/materialized view/table).  
We already have icons for tables and views
and I've made another (all 3 attached) that combines
the table and view icon to represent materialized
views.  Buttons will only appear supporting
a specific data-involving operation, e.g. delete, when the
user has permissions to perform the operation
and when the underlying relation supports
the operation.

Of course views, materialized views, and tables
will continue to have different pages for
creation/modification.

Objections?  Discussion?  I'd like to get to this
today.


Karl <ko...@me...>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

P.S.  Speaking of icons, should we be keeping icon "source"
as gimp .xcf files (or something) somewhere?  The materialized
view icon was done with layers, but this goes away when
exported as a png.

[ppa-dev] Patches for ppa self-tests

[Karl O. P. <ko...@me...>]

Hi,

I've a number of patches for inclusion into ppa.
I thought, rather than attach them here, I'd
link to the branches on github and describe them.

If this isn't a good process then please let me
know what works better.


https://github.com/kpinc/phppgadmin/commits/selenium-login

1 patch.
Deletes cookies on selenium login so that previous
failed or paused tests do not affect the execution of
subsequent manually executed tests.  This allows
the user more control over manual test execution.


https://github.com/kpinc/phppgadmin/commits/selenium-enabled

3 patches.
Internationalizes the selenium link on the ppa intro page.
Makes the link "hot" only when selenium has been configured.
Adds a "reset" link to clean cruft from failed self-test runs.


https://github.com/kpinc/phppgadmin/commits/selenium-doc

4 patches.
Changes to tests/selenium/README to make it more better.
The last patch documents the "reset" link above.


https://github.com/kpinc/phppgadmin/commits/gitignore

1 patch.
Frob .gitignore to ignore emacs backup files.

Regards,

Karl <ko...@me...>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

Re: [ppa-dev] Fwd: Patch for ALTER TABLE CLUSTER ON

[Karl O. P. <ko...@me...>]

On 04/25/2013 08:12:20 PM, Robert Treat wrote:
> Wow, copy/paste fail? Well, the one you pointed to is probably
> helpful, although I originally was going to link to
> https://github.com/phppgadmin/phppgadmin/
> commit/6cf0df2174631b8f0351de9fb60c636d1a45ccfe

Thanks.


Karl <ko...@me...>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

Re: [ppa-dev] Fwd: Patch for ALTER TABLE CLUSTER ON

[Robert T. <ro...@xz...>]

Wow, copy/paste fail? Well, the one you pointed to is probably
helpful, although I originally was going to link to
https://github.com/phppgadmin/phppgadmin/commit/6cf0df2174631b8f0351de9fb60c636d1a45ccfe

Robert Treat

On Thu, Apr 25, 2013 at 12:30 PM, Karl O. Pinc <ko...@me...> wrote:
> Hi Robert,
>
> I don't understand your message below.  I don't see a reference
> to a commit and the patch which goes with the referenced
> email does not seem to do anything regards tests.
>
> Are you talking about commit
> 943531194410352da0847d2e766067d08f53ae8d
> ?
>
>
>
> On 04/23/2013 10:54:48 AM, Robert Treat wrote:
>> I did take a look at it then and thought we should probably add some
>> tests
>> into the system when we commit it. In the mean time I got side
>> tracked
>> by
>> the release. If you want to take a swing at adding tests, here is a
>> basic
>> commit which shows tests being added, it might be of help. Otherwise
>> it's
>> back on my todo list, so it will be looked at before to much longer.
>>
>> Sorry for not responding previously.
>>
>>
>> Robert Treat
>> conjecture: xzilla.net
>> consulting: omniti.com
>>
>>
>> On Sun, Apr 21, 2013 at 12:06 PM, phb07 <ph...@ap...> wrote:
>>
>> >  Hi all,
>> >
>> > Has someone had a look on the patch I proposed some months ago ?
>> >
>> > Thanks by advance.
>> > Philippe.
>> >
>> > -------- Message original --------  Sujet: Patch for ALTER TABLE
>> CLUSTER
>> > ON  Date : Thu, 27 Dec 2012 12:33:09 +0100  De : phb07
>> > <ph...@ap...> <ph...@ap...>  Pour :
>> > php...@li...
>> >
>> >
>> > Hi all,
>> >
>> > Under my Christmas Tree, I have found the attached patch ;-).
>> >
>> > As far as I can understand, it adds a capability to the alter table
>> > functionality, namely to specify, change or remove a clustered
>> index
>> > definition.
>> > It does the equivalent of the "ALTER TABLE CLUSTER ON ..." and
>> "ALTER
>> > TABLE SET WITHOUT CLUSTER" sql statements.
>> >
>> > The patch should apply on the HEAD.
>> > It only works for postgres version >= 8.0 as 7.4 doesn't support
>> the
>> SET
>> > WITHOUT CLUSTER clause.
>> >
>> > Unfortunately, Christmas Father has not tested this patch with
>> pg7.4.
>> > He has not included additional Selenium regression test too (he is
>> not
>> > yet familiar with this nice tool ;-)).
>> >
>> > Let me know what you think about this patch (as a first patch
>> supplied,
>> > it may need some refinements). I will forward to its author...
>> >
>> > With my best wishes,
>> > Philippe Beaudoin (beaud76)
>> >
>> >
>> >
>> >
>> >
>> >
>> >
>> >
>> ------------------------------------------------------------------------------
>> > Precog is a next-generation analytics platform capable of advanced
>> > analytics on semi-structured data. The platform includes APIs for
>> building
>> > apps and a phenomenal toolset for data science. Developers can use
>> > our toolset for easy data analysis & visualization. Get a free
>> account!
>> > http://www2.precog.com/precogplatform/slashdotnewsletter
>> > _______________________________________________
>> > phpPgAdmin-devel mailing list
>> > php...@li...
>> > https://lists.sourceforge.net/lists/listinfo/phppgadmin-devel
>> >
>> >
>>
>
> ------quoted attachment------
>> ------------------------------------------------------------------------------
>> Try New Relic Now & We'll Send You this Cool Shirt
>> New Relic is the only SaaS-based application performance monitoring
>> service
>> that delivers powerful full stack analytics. Optimize and monitor
>> your
>> browser, app, & servers with just a few lines of code. Try New Relic
>> and get this awesome Nerd Life shirt!
>> http://p.sf.net/sfu/newrelic_d2d_apr
>
> ------quoted attachment------
>> _______________________________________________
>> phpPgAdmin-devel mailing list
>> php...@li...
>> https://lists.sourceforge.net/lists/listinfo/phppgadmin-devel
>>
>
>
>
>
> Karl <ko...@me...>
> Free Software:  "You don't pay back, you pay forward."
>                  -- Robert A. Heinlein
>

Re: [ppa-dev] langcheck and Chinese translation

[Hans C. <ha...@gm...>]

On Tue, Apr 23, 2013 at 1:26 PM, Hans Chen <ha...@gm...> wrote:
> On Mon, Apr 22, 2013 at 5:42 PM, Jehan-Guillaume (ioguix) de Rorthais
> <io...@fr...> wrote:
>> On 22/04/2013 13:10, Hans Chen wrote:
>>> Hi All
>>
>> Hello,
>
> Thank you to check it.
>
>>
>>> I have posted pull request to github, would anyone check it?
>>
>> your pull request add some non related commits, especialy the merge one:
>>
>> https://github.com/cozhan/phppgadmin/commit/83011cddb6a083a779dcad3c2e674d5672c52350
>
> Yes, my mistake. I will fix it.

I have removed the merge one. But still curious how can I keep
updating from phppgadmin and no merge commit to pull request?

>
>>
>>> two major modifications:
>>>
>>> 1. complete the Chinese utf8 translation.
>>
>> could you squash/fixup the related commits together by any chance (using
>> git rebase -i) ?
>
> OK, I will squash all translation revisions.

Done, fixup all translations.

>
>> Moreover:
>>
>> 3. what is this commit:
>>
>> https://github.com/cozhan/phppgadmin/commit/8ead03b41d7e9e46eec2cd15bb1f7119e40fe3c7
>>
>> Could you tell me why this is a bug ?
>
> There is one space between check box and "full" / "analyze" on the
> Vacuum page, but not for "freeze".
> It is actually not a "bug", just like a typo.

Because of you have pushed this commit, should I remove it from pull
request? Sorry I am not very familiar with pull request in github. May
I manually choose which commits should be in pull request?

Best Regards,

Hans Chen

Re: [ppa-dev] Fwd: Patch for ALTER TABLE CLUSTER ON

[Karl O. P. <ko...@me...>]

Hi Robert,

I don't understand your message below.  I don't see a reference
to a commit and the patch which goes with the referenced
email does not seem to do anything regards tests.

Are you talking about commit
943531194410352da0847d2e766067d08f53ae8d
?



On 04/23/2013 10:54:48 AM, Robert Treat wrote:
> I did take a look at it then and thought we should probably add some
> tests
> into the system when we commit it. In the mean time I got side 
> tracked
> by
> the release. If you want to take a swing at adding tests, here is a
> basic
> commit which shows tests being added, it might be of help. Otherwise
> it's
> back on my todo list, so it will be looked at before to much longer.
> 
> Sorry for not responding previously.
> 
> 
> Robert Treat
> conjecture: xzilla.net
> consulting: omniti.com
> 
> 
> On Sun, Apr 21, 2013 at 12:06 PM, phb07 <ph...@ap...> wrote:
> 
> >  Hi all,
> >
> > Has someone had a look on the patch I proposed some months ago ?
> >
> > Thanks by advance.
> > Philippe.
> >
> > -------- Message original --------  Sujet: Patch for ALTER TABLE
> CLUSTER
> > ON  Date : Thu, 27 Dec 2012 12:33:09 +0100  De : phb07
> > <ph...@ap...> <ph...@ap...>  Pour :
> > php...@li...
> >
> >
> > Hi all,
> >
> > Under my Christmas Tree, I have found the attached patch ;-).
> >
> > As far as I can understand, it adds a capability to the alter table
> > functionality, namely to specify, change or remove a clustered 
> index
> > definition.
> > It does the equivalent of the "ALTER TABLE CLUSTER ON ..." and
> "ALTER
> > TABLE SET WITHOUT CLUSTER" sql statements.
> >
> > The patch should apply on the HEAD.
> > It only works for postgres version >= 8.0 as 7.4 doesn't support 
> the
> SET
> > WITHOUT CLUSTER clause.
> >
> > Unfortunately, Christmas Father has not tested this patch with
> pg7.4.
> > He has not included additional Selenium regression test too (he is
> not
> > yet familiar with this nice tool ;-)).
> >
> > Let me know what you think about this patch (as a first patch
> supplied,
> > it may need some refinements). I will forward to its author...
> >
> > With my best wishes,
> > Philippe Beaudoin (beaud76)
> >
> >
> >
> >
> >
> >
> >
> > 
> ------------------------------------------------------------------------------
> > Precog is a next-generation analytics platform capable of advanced
> > analytics on semi-structured data. The platform includes APIs for
> building
> > apps and a phenomenal toolset for data science. Developers can use
> > our toolset for easy data analysis & visualization. Get a free
> account!
> > http://www2.precog.com/precogplatform/slashdotnewsletter
> > _______________________________________________
> > phpPgAdmin-devel mailing list
> > php...@li...
> > https://lists.sourceforge.net/lists/listinfo/phppgadmin-devel
> >
> >
> 

------quoted attachment------
> ------------------------------------------------------------------------------
> Try New Relic Now & We'll Send You this Cool Shirt
> New Relic is the only SaaS-based application performance monitoring
> service 
> that delivers powerful full stack analytics. Optimize and monitor 
> your
> browser, app, & servers with just a few lines of code. Try New Relic
> and get this awesome Nerd Life shirt!
> http://p.sf.net/sfu/newrelic_d2d_apr

------quoted attachment------
> _______________________________________________
> phpPgAdmin-devel mailing list
> php...@li...
> https://lists.sourceforge.net/lists/listinfo/phppgadmin-devel
> 




Karl <ko...@me...>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

Re: [ppa-dev] default db on login page

[Robert T. <ro...@xz...>]

On Sat, Apr 20, 2013 at 9:26 AM, Jehan-Guillaume (ioguix) de Rorthais
<io...@fr...> wrote:
> On 19/04/2013 23:46, Hans Chen wrote:
>> On Fri, Apr 19, 2013 at 4:52 PM, Robert Treat <ro...@xz...> wrote:
>>>
>>> On Fri, Apr 19, 2013 at 10:14 AM, Karl O. Pinc <ko...@me...> wrote:
>>>> On 04/18/2013 10:09:30 PM, Robert Treat wrote:
>>>>> On Thu, Apr 18, 2013 at 2:26 PM, Karl O. Pinc <ko...@me...> wrote:
>>>>>> On 04/18/2013 10:47:35 AM, Hans Chen wrote:
>>>>
>>>>>>> * Add support for default db to connect to on login page (we had
>>>>>>> several
>>>>>>> users complaining about that lately)
>>>>
>>>>>>   I'd think the default
>>>>>> would go in the config file as another config parameter,
>>>>>> but if you've a different idea it might be good to discuss
>>>>>> it ahead of time.
>>>>>>
>>>>>
>>>>> Actually, a better feature would be one I had started dabbling with,
>>>>> which I called "libpq_login". The basic idea was that you could set
>>>>> libpq_login to true in the config, and instead of displaying the
>>>>> normal sets of servers, instead we would show a login box where you
>>>>> could enter user, host, database, port type information, or
>>>>> alternatively a url formatted libpq connect string.
>>>>
>>>> I don't see this as a substitute for having a default server
>>>> and thereby having the "home page" start by asking for username
>>>> and password.  The whole point, to my way of thinking, is that
>>>> the user not have to think about the server at all.  I don't know,
>>>> but I'd guess that a lot of installations have just a single server.
>>>>
>>>
>>> That's fair. If there is demand for the default server login, it's
>>> probably simpler to implement.
>>>
>>
>> I am little curious what is the complain from users. Seems current way
>> is not bad if only single server.
>
> Some user would like to be able to connect directly on a given database,
> not to the defaultdb given in the conf file.
>
>>>>> This would allow
>>>>> people to be able to login without having to be able to get access to
>>>>> the command line to configure every server.
>>>>
>>>> By this you mean without having to edit the ppa config file?
>>>>
>>>> I agree it would be nice to not be limited to the servers
>>>> listed in the config file.  There's a temptation to want
>>>> to be able to put some limits on what servers could be
>>>> reached so as to avoid an "open relay" sort of situation.
>>>> But I think it'd be best to avoid over-design at this point.
>>>>
>>>
>>> Yeah... I've mostly seen this from people running in the cloud, or
>>> doing heroku type stuff, where the connection credentials seem to be
>>> more fluid.
>>
>> A connection string like : pg://user:password@server:port/database is
>> cool. May I start from it?
>
> It already exists since PostgreSQL 9.1. See :
>
> http://www.postgresql.org/docs/current/interactive/libpq-connect.html#LIBPQ-CONNSTRING
>
> So if Robert is adding a form to connect to a database using a libpq
> connection string, then this syntax will be automatically supported.
>
> I see no point to adding this with some parsing pain inside PPA though.
>

Right, I'm basically talking about building on top of what Postgres
already supports. That said, if we feel that we want people to be able
to use a non-formatted string, and be able to input user name,
password, host, etc... for less than 9.1 servers, I could see adding
an empty connection form and then parsing it and connecting with it.
The hard part of this really isn't the credentials capturing, its
rewiring the backend to accept those credentials rather than using
info from the config file.

Also, to be clear, this is a totally different use case / feature than
what Karl is talking about, which also seems like a valid use case.


Robert Treat
play: xzilla.net
work: omniti.com

Re: [ppa-dev] Fwd: Patch for ALTER TABLE CLUSTER ON

[Robert T. <ro...@xz...>]

I did take a look at it then and thought we should probably add some tests
into the system when we commit it. In the mean time I got side tracked by
the release. If you want to take a swing at adding tests, here is a basic
commit which shows tests being added, it might be of help. Otherwise it's
back on my todo list, so it will be looked at before to much longer.

Sorry for not responding previously.


Robert Treat
conjecture: xzilla.net
consulting: omniti.com


On Sun, Apr 21, 2013 at 12:06 PM, phb07 <ph...@ap...> wrote:

>  Hi all,
>
> Has someone had a look on the patch I proposed some months ago ?
>
> Thanks by advance.
> Philippe.
>
> -------- Message original --------  Sujet: Patch for ALTER TABLE CLUSTER
> ON  Date : Thu, 27 Dec 2012 12:33:09 +0100  De : phb07
> <ph...@ap...> <ph...@ap...>  Pour :
> php...@li...
>
>
> Hi all,
>
> Under my Christmas Tree, I have found the attached patch ;-).
>
> As far as I can understand, it adds a capability to the alter table
> functionality, namely to specify, change or remove a clustered index
> definition.
> It does the equivalent of the "ALTER TABLE CLUSTER ON ..." and "ALTER
> TABLE SET WITHOUT CLUSTER" sql statements.
>
> The patch should apply on the HEAD.
> It only works for postgres version >= 8.0 as 7.4 doesn't support the SET
> WITHOUT CLUSTER clause.
>
> Unfortunately, Christmas Father has not tested this patch with pg7.4.
> He has not included additional Selenium regression test too (he is not
> yet familiar with this nice tool ;-)).
>
> Let me know what you think about this patch (as a first patch supplied,
> it may need some refinements). I will forward to its author...
>
> With my best wishes,
> Philippe Beaudoin (beaud76)
>
>
>
>
>
>
>
> ------------------------------------------------------------------------------
> Precog is a next-generation analytics platform capable of advanced
> analytics on semi-structured data. The platform includes APIs for building
> apps and a phenomenal toolset for data science. Developers can use
> our toolset for easy data analysis & visualization. Get a free account!
> http://www2.precog.com/precogplatform/slashdotnewsletter
> _______________________________________________
> phpPgAdmin-devel mailing list
> php...@li...
> https://lists.sourceforge.net/lists/listinfo/phppgadmin-devel
>
>

Re: [ppa-dev] alter column order

[Robert T. <ro...@xz...>]

On Mon, Apr 22, 2013 at 8:49 AM, Jehan-Guillaume (ioguix) de Rorthais
<io...@fr...> wrote:
> On 20/04/2013 17:17, Karl O. Pinc wrote:
>> On 04/20/2013 09:34:04 AM, Hans Chen wrote:
>>> On Sat, Apr 20, 2013 at 4:11 PM, Karl O. Pinc <ko...@me...> wrote:
>>>> On 04/20/2013 08:04:04 AM, Jehan-Guillaume (ioguix) de Rorthais
>>> wrote:
>>>>> On 20/04/2013 00:03, Hans Chen wrote:
<snip>,
>>> Yes, actually there are also some thoughts from pg community. See
>>> here: http://wiki.postgresql.org/wiki/Alter_column_position
>>>
>>> I think recreate table or recreate column would have trouble if the
>>> old column is with constraint.
>>
>> Not only constraints but also indexes, clustering, views, and triggers
>> come to mind.
>>
>>>
>>> I will tested both recreate and my way, and will also spend several
>>> days to investigate postgresql source code.
>>
>> Yes.
>> The right way would be to work first with pg and get changing
>> column ordering supported there and then add that support to ppa.
>
> +1000
>

Yeah, I looked into this a number of years back, and it's somewhat
complicated; right now attnum is used within the postgres catalogs and
source code to represent both physical and logical position of a row.
If you were going to try and implement this, I think you need to do it
in parts; first modify the postgres catalogs and code to have a
separate idea of both the physical and logical column position, then
you can implement the column position changes, then if you want bonus
points, implement automatic physical reordering for optimized page
layout. There are some good threads on hackers about this, though they
are quite a bit older, but the state of that code hasn't changed much.
Before you code anything though, you'll want to get some kind of buy
in on the hackers list (which I'll admit is bike-sheddingly
dangerous).

Robert Treat
play: xzilla.net
work: omniti.com

Re: [ppa-dev] langcheck and Chinese translation

[Jehan-Guillaume (i. de R. <io...@fr...>]

On 23/04/2013 13:26, Hans Chen wrote:
>> > 3. what is this commit:
>> >
>> > https://github.com/cozhan/phppgadmin/commit/8ead03b41d7e9e46eec2cd15bb1f7119e40fe3c7
>> >
>> > Could you tell me why this is a bug ?
> There is one space between check box and "full" / "analyze" on the
> Vacuum page, but not for "freeze".
> It is actually not a "bug", just like a typo.

I committed/pushed your cosmetic fix in master and REL_5-1 branches with
some edition to the commit messages.

Thank you,

/ioguix

Re: [ppa-dev] langcheck and Chinese translation

[Hans C. <ha...@gm...>]

On Mon, Apr 22, 2013 at 5:42 PM, Jehan-Guillaume (ioguix) de Rorthais
<io...@fr...> wrote:
> On 22/04/2013 13:10, Hans Chen wrote:
>> Hi All
>
> Hello,

Thank you to check it.

>
>> I have posted pull request to github, would anyone check it?
>
> your pull request add some non related commits, especialy the merge one:
>
> https://github.com/cozhan/phppgadmin/commit/83011cddb6a083a779dcad3c2e674d5672c52350

Yes, my mistake. I will fix it.

>
>> two major modifications:
>>
>> 1. complete the Chinese utf8 translation.
>
> could you squash/fixup the related commits together by any chance (using
> git rebase -i) ?

OK, I will squash all translation revisions.

>
>> 2. modify langcheck to give correct instruciton if the language
>> parameter is with .php
>
> I need some more time to understand this (I just don't have time right now).

Thanks. FYR, type command under ./lang/: php langcheck english.php.
Currently it will say : Error: File not found.
The patch make it to display correct instruction:
Usage: langcheck <language>
<language> is the filename without the .php extension

>
> Moreover:
>
> 3. what is this commit:
>
> https://github.com/cozhan/phppgadmin/commit/8ead03b41d7e9e46eec2cd15bb1f7119e40fe3c7
>
> Could you tell me why this is a bug ?

There is one space between check box and "full" / "analyze" on the
Vacuum page, but not for "freeze".
It is actually not a "bug", just like a typo.

>
> thank you !
>
>> Best Regards,
>>
>> Hans Chen
>

Re: [ppa-dev] langcheck and Chinese translation

[Jehan-Guillaume (i. de R. <io...@fr...>]

On 22/04/2013 13:10, Hans Chen wrote:
> Hi All

Hello,

> I have posted pull request to github, would anyone check it?

your pull request add some non related commits, especialy the merge one:

https://github.com/cozhan/phppgadmin/commit/83011cddb6a083a779dcad3c2e674d5672c52350

> two major modifications:
> 
> 1. complete the Chinese utf8 translation.

could you squash/fixup the related commits together by any chance (using
git rebase -i) ?

> 2. modify langcheck to give correct instruciton if the language
> parameter is with .php

I need some more time to understand this (I just don't have time right now).

Moreover:

3. what is this commit:

https://github.com/cozhan/phppgadmin/commit/8ead03b41d7e9e46eec2cd15bb1f7119e40fe3c7

Could you tell me why this is a bug ?

thank you !

> Best Regards,
> 
> Hans Chen

Re: [ppa-dev] alter column order

[Jehan-Guillaume (i. de R. <io...@fr...>]

On 20/04/2013 17:17, Karl O. Pinc wrote:
> On 04/20/2013 09:34:04 AM, Hans Chen wrote:
>> On Sat, Apr 20, 2013 at 4:11 PM, Karl O. Pinc <ko...@me...> wrote:
>>> On 04/20/2013 08:04:04 AM, Jehan-Guillaume (ioguix) de Rorthais
>> wrote:
>>>> On 20/04/2013 00:03, Hans Chen wrote:
>>>>> Hi
>>>>>
>>>>> I am considering to add function to change position of columns 
>> in
>>>>> table.
> 
>>> Anyhow you could re-order table columns by saving the
>>> table's data, re-creating the
>>> table, and transferring the data into it.

This is a very heavy process for most PostgreSQL clusters I know.
Moreover, your use case (AFAIU, messing with data for ergonomic
improvement on one user interface side) does not help to sell me this
feature neither :(

This just do not deserve so much trouble on the database side.

>> Yes, actually there are also some thoughts from pg community. See
>> here: http://wiki.postgresql.org/wiki/Alter_column_position
>>
>> I think recreate table or recreate column would have trouble if the
>> old column is with constraint.
> 
> Not only constraints but also indexes, clustering, views, and triggers
> come to mind.
> 
>>
>> I will tested both recreate and my way, and will also spend several
>> days to investigate postgresql source code.
> 
> Yes.
> The right way would be to work first with pg and get changing
> column ordering supported there and then add that support to ppa.

+1000

> Karl <ko...@me...>
> Free Software:  "You don't pay back, you pay forward."
>                  -- Robert A. Heinlein

[ppa-dev] langcheck and Chinese translation

[Hans C. <ha...@gm...>]

Hi All

I have posted pull request to github, would anyone check it?

two major modifications:

1. complete the Chinese utf8 translation.
2. modify langcheck to give correct instruciton if the language
parameter is with .php

Best Regards,

Hans Chen

[ppa-dev] Fwd: Patch for ALTER TABLE CLUSTER ON

[phb07 <ph...@ap...>]

Hi all,

Has someone had a look on the patch I proposed some months ago ?

Thanks by advance.
Philippe.

-------- Message original --------
Sujet: 	Patch for ALTER TABLE CLUSTER ON
Date : 	Thu, 27 Dec 2012 12:33:09 +0100
De : 	phb07 <ph...@ap...>
Pour : 	php...@li...



Hi all,

Under my Christmas Tree, I have found the attached patch ;-).

As far as I can understand, it adds a capability to the alter table
functionality, namely to specify, change or remove a clustered index
definition.
It does the equivalent of the "ALTER TABLE CLUSTER ON ..." and "ALTER
TABLE SET WITHOUT CLUSTER" sql statements.

The patch should apply on the HEAD.
It only works for postgres version >= 8.0 as 7.4 doesn't support the SET
WITHOUT CLUSTER clause.

Unfortunately, Christmas Father has not tested this patch with pg7.4.
He has not included additional Selenium regression test too (he is not
yet familiar with this nice tool ;-)).

Let me know what you think about this patch (as a first patch supplied,
it may need some refinements). I will forward to its author...

With my best wishes,
Philippe Beaudoin (beaud76)

Re: [ppa-dev] alter column order

[Karl O. P. <ko...@me...>]

On 04/20/2013 09:34:04 AM, Hans Chen wrote:
> On Sat, Apr 20, 2013 at 4:11 PM, Karl O. Pinc <ko...@me...> wrote:
> > On 04/20/2013 08:04:04 AM, Jehan-Guillaume (ioguix) de Rorthais
> wrote:
> >> On 20/04/2013 00:03, Hans Chen wrote:
> >> > Hi
> >> >
> >> > I am considering to add function to change position of columns 
> in
> >> > table.

> > Anyhow you could re-order table columns by saving the
> > table's data, re-creating the
> > table, and transferring the data into it.
> 
> Yes, actually there are also some thoughts from pg community. See
> here: http://wiki.postgresql.org/wiki/Alter_column_position
> 
> I think recreate table or recreate column would have trouble if the
> old column is with constraint.

Not only constraints but also indexes, clustering, views, and triggers
come to mind.

> 
> I will tested both recreate and my way, and will also spend several
> days to investigate postgresql source code.

Yes.
The right way would be to work first with pg and get changing
column ordering supported there and then add that support to ppa.



Karl <ko...@me...>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

Re: [ppa-dev] alter column order

[Hans C. <ha...@gm...>]

On Sat, Apr 20, 2013 at 3:04 PM, Jehan-Guillaume (ioguix) de Rorthais
<io...@fr...> wrote:
> On 20/04/2013 00:03, Hans Chen wrote:
>> Hi
>>
>> I am considering to add function to change position of columns in
>> table.
>
> I am not sure to understand clearly what you want to achieve. PostgreSQL
> doesn't supports "changing" columns position.
>
>> Have tested to modify value of attnum in pg_attribute is a
>> possible solution.
>
> Then  you probably corrupt your database in some ways in guess. You
> should never issue write queries to the catalog.

I agree that is correct. I found there are some differences between
different version pg catalog structure.

To write it is dangerous although I have no other good way to do it.

>
>> I may add two buttons (up and down) to table view.
>>
>> let me know what is your suggestions.
>
> If you really want to move your columns position, then create a view.
>
> BTW, why do you want to move your table columns in the first place ?

I have to offer a better view for somebody need to fill in data by
ppg. Some times a new added column should be in front or middle
logically.

>
>>
>> Best Regards,
>>
>> Hans Chen

Re: [ppa-dev] alter column order

[Hans C. <ha...@gm...>]

On Sat, Apr 20, 2013 at 4:11 PM, Karl O. Pinc <ko...@me...> wrote:
> On 04/20/2013 08:04:04 AM, Jehan-Guillaume (ioguix) de Rorthais wrote:
>> On 20/04/2013 00:03, Hans Chen wrote:
>> > Hi
>> >
>> > I am considering to add function to change position of columns in
>> > table.
>>
>> I am not sure to understand clearly what you want to achieve.
>> PostgreSQL
>> doesn't supports "changing" columns position.
>
> Some thoughts....
>
> No, it's not supported by pg but that does not mean it
> might not be useful, and I've had related thoughts along
> similar lines.
>
> (The "browse" feature is reasonably
> useless for anything other than a general overview
> of what's in the table, for example, since it does not sort
> the rows.  So, it'd be nice to be able to allow each individual user
> a way to configure how they want each table sorted.
> Of course clustering is also an option.  But I think that's
> mixing performance tuning with user interface.)
>
> Anyhow you could re-order table columns by saving the
> table's data, re-creating the
> table, and transferring the data into it.

Yes, actually there are also some thoughts from pg community. See
here: http://wiki.postgresql.org/wiki/Alter_column_position

I think recreate table or recreate column would have trouble if the
old column is with constraint.

I will tested both recreate and my way, and will also spend several
days to investigate postgresql source code.

>
> Karl <ko...@me...>
> Free Software:  "You don't pay back, you pay forward."
>                  -- Robert A. Heinlein
>