headers
Joe Stump | 1 Aug 2005 05:11
Gravatar

Re: XML_Atom?

I'm about 60% done with an entire XML_Feed abstraction setup (for  
reading feeds). I'm interested in whoever is working on  
XML_Feed_Writer. Maybe we could combine our efforts into a single  
package (reading and writing feeds).

I've abstracted Entry, Info (feed info) and Enclosures thus far. The  
idea is that no matter if you're parsing (or creating) an Atom/RSS/ 
etc. feed all of the core components will look the same and act the  
same.

If anyone is interested let me know. I should have basic parsing done  
for RSS and Atom in the next few days.

--Joe

On Jul 31, 2005, at 8:00 PM, Firman Wandayandi wrote:

> Hi Joe,
>
> On 7/31/05, Joe Stump <joe <at> joestump.net> wrote:
>
>> Is there an Atom parser in PEAR? I searched and didn't find anything
>> - just wondering if I'm missing something. If there isn't anything I
>> suppose I'll start hacking around.
>>
>>
>
> I was proposed the XML_Atom package about several months ago, but we
> agreed to merge XML_RSS package with it, the package will be named
> XML_Syndication, which is provide for build (write) and parse Atom and
(Continue reading)

Permalink | Reply | 0 comments |
headers
Koistya `Navin | 1 Aug 2005 05:55
Picon

RE: XML_Atom?


Hello,

There is a package at PEPr called "XML_Feed_Writer", maybe the better name
for parser package would be "XML_Feed_Parser", huh?

Sincerely,
Koistya `Navin
http://navin.biz

> -----Original Message-----
> From: Joe Stump [mailto:joe <at> joestump.net]
> Sent: Sunday, July 31, 2005 8:59 PM
> To: James Stewart
> Cc: PEAR Dev List
> Subject: Re: [PEAR-DEV] XML_Atom?
> 
> Ha! I just put in about 2 hours last night starting to create
> XML_Feeds with factory pattern for RSS and Atom. Send your code over
> and I'll see what overlaps. I'm definitely interested in this project.
> 
> --Joe
> 
> On Jul 31, 2005, at 6:51 AM, James Stewart wrote:
> 
> > On Jul 31, 2005, at 12:04 AM, Joe Stump wrote:
> >
> >> Is there an Atom parser in PEAR? I searched and didn't find
> >> anything - just wondering if I'm missing something. If there isn't
> >> anything I suppose I'll start hacking around.
(Continue reading)

Permalink | Reply | 2 comments |
headers
Joe Stump | 1 Aug 2005 06:31
Gravatar

Re: XML_Atom?

No, I think everything should be XML_Feed - it should be an all  
encompassing package for reading and writing feeds, which is why I  
sent my email about XML_Feed_Writer. We should combine forces to  
create XML_Feed IMO.

I'll be contacting those who proposed XML_Feed_Writer and talk to  
them about combining forces. As a side note, our efforts will most  
likely deprecate XML_RSS. I've found that parser doesn't support  
enclosures, among other features I want to include in XML_Feed.

At any rate, I want to make it clear I'd like to work WITH  
XML_Feed_Writer's devs and create a great feed reading/writing  
package for PEAR. If you are an XML_Feed_Writer dev please email me  
offlist and we'll hook up and compare notes.

Good times!

--Joe

On Jul 31, 2005, at 8:55 PM, Koistya `Navin wrote:

>
>
>
> Hello,
>
> There is a package at PEPr called "XML_Feed_Writer", maybe the  
> better name
> for parser package would be "XML_Feed_Parser", huh?
>
(Continue reading)

Permalink | Reply | 1 comment |
headers
PEPr | 1 Aug 2005 10:00
Picon

[PEPr] Proposal finished Images::Image_Canvas


PEPr has automatically finished the proposal on Images::Image_Canvas.

Sum of Votes: 7 (0 conditional)
Result:       This proposal was accepted

Further details on the status of the proposal and the votes can be found here:
http://pear.php.net/pepr/pepr-proposal-show.php?id=212

If you are the person who initiated the proposal,
please read the 'Proposal Finished' instructions at
http://pear.php.net/manual/en/newmaint.proposal.step3.php#newmaint.proposal.step3.stage4

-- 
Sent by PEPr, the automatic proposal system at http://pear.php.net

--

-- 
PEAR Development Mailing List (http://pear.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
Permalink | Reply | 0 comments |
headers
Markus Wolff | 1 Aug 2005 12:02
Picon

Fun with DataObject, Transactions, PostgreSQL and Sequences

Hi there,

I'm using DataObject (1.7.15) with PostgreSQL (8.0.1) now for the first 
time and I've encountered some strange phenomenon.

I have two DataObjects where I first look into the database to see if a 
record with the given parameters already exists. If it doesn't, it is to 
be created. The database is in virgin state, so there is no data and 
also the sequences for the tables do not exist yet. All operations are 
encapsulated within one transaction. So far, so good.

The first table is called "language". Everything goes as planned:

<SNIP>
Language: QUERY: SELECT * FROM language WHERE language.name = 'German'
[...I have added a bit of my own debug info, so the following deviates a 
bit from the standard DO debug output...]
[db_error: message="DB Error: no such table" code=-18 mode=return 
level=notice prefix="" info="SELECT NEXTVAL('language_seq') 
[nativecode=ERROR:  relation "language_seq" does not exist]"]
Creating sequence...SeqName format is: %s_seq
FINAL SEQNAME: language_seq
NEXTVAL: 1
Language: QUERY: INSERT INTO language (language_id , name ) VALUES ( 1 , 
'German' )
Language: query: QUERY DONE IN  0.033642053604126 seconds
</SNIP>

As expected, an error is thrown on the first call to NEXTVAL, as the 
sequence doesn't exist yet. The sequence is then created and the INSERT
(Continue reading)

Permalink | Reply | 2 comments |
headers
Alexey Borzov | 1 Aug 2005 12:11
Picon
Favicon

Re: Fun with DataObject, Transactions, PostgreSQL and Sequences

Hi,

Markus Wolff wrote:
> LOG:  statement: CREATE SEQUENCE pagesize_seq
> ERROR:  current transaction is aborted, commands ignored until end of 
> transaction block
> </SNIP>
> 
> I have no clue why the "begin;" is coming *after* the first call to 
> NEXTVAL, as $do->query('BEGIN'); is actually the first thing I do after 
> creating the first DataObject.
> 
> Anyway, cross-checking with psql shows that it seems to be very likely 
> that Postgres has some problems with using DDL statements (such as 
> CREATE SEQUENCE) within transactions. Leave out the transaction and all 
> is good, all is well and everyone lived happily everafter.

In fact PostgreSQL doesn't have problems with DDL in transactions. It has 
problems with errors in transactions (as the log above suggests): if there is an 
error then the whole transaction is effectively rolled back.

Since you are using 8.0 you may consider using SAVEPOINT before calling 
nextval() and ROLLBACK TO SAVEPOINT in case of non-existant sequence, in that 
case only the part of the transaction will be rolled back.

--

-- 
PEAR Development Mailing List (http://pear.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
(Continue reading)

Permalink | Reply | 1 comment |
headers
Alexey Borzov | 1 Aug 2005 12:47
Picon
Favicon

A question to __autoload() fanboys

Hi,

Since the flame [1] about the evil QuickForm developers not willing to fix the 
bugs in other peoples's code [2] somehow died down I'd like to ask a new 
technical question.

I was going through QuickForm thinking on whether to actually fix the "bug" 
#3197 and found a following code fragment in 
HTML_QuickForm_Renderer_ITStatic::_getTplBlock():

         if (is_a($this->_tpl, 'html_template_sigma')) {
             ... code for HTML_Template_Sigma ...
         } else {
             ... code for HTML_Template_IT ...
         }

OK, I *can* fix the case of 'html_template_sigma' here, but the problem is the 
following: if a person uses HTML_Template_IT package he most probably does not 
have HTML_Template_Sigma installed (and vice versa), thus trying to autoload the 
class will fail! How should I fix this problem?

Scott, what will you suggest *this time*?

[1] http://marc.theaimsgroup.com/?l=pear-dev&m=112239979300570&w=2
[2] http://pear.php.net/bugs/bug.php?id=3197

--

-- 
PEAR Development Mailing List (http://pear.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
(Continue reading)

Permalink | Reply | 6 comments |
headers
Christian Stocker | 1 Aug 2005 13:31
Picon

Re: A question to __autoload() fanboys


On 1.8.2005 12:47 Uhr, Alexey Borzov wrote:
> Hi,
> 
> Since the flame [1] about the evil QuickForm developers not willing to
> fix the bugs in other peoples's code [2] somehow died down I'd like to
> ask a new technical question.
> 
> I was going through QuickForm thinking on whether to actually fix the
> "bug" #3197 and found a following code fragment in
> HTML_QuickForm_Renderer_ITStatic::_getTplBlock():
> 
>         if (is_a($this->_tpl, 'html_template_sigma')) {
>             ... code for HTML_Template_Sigma ...
>         } else {
>             ... code for HTML_Template_IT ...
>         }
> 
> OK, I *can* fix the case of 'html_template_sigma' here, but the problem
> is the following: if a person uses HTML_Template_IT package he most
> probably does not have HTML_Template_Sigma installed (and vice versa),
> thus trying to autoload the class will fail! How should I fix this problem?
> 
> Scott, what will you suggest *this time*?

It won't find the class (or the corresponding file), 'cause it's
actually not there... One can handle that in __autoload easily.

While asking for html_quickform_group when it's expected to be
HTML_QuickForm_group (from Scott's perspective) is a whole different
(Continue reading)

Permalink | Reply | 1 comment |
headers
Michael Wallner | 1 Aug 2005 14:31
Picon
Gravatar

Re: A question to __autoload() fanboys

Hi Alexey Borzov, you wrote:

> OK, I *can* fix the case of 'html_template_sigma' here, but the problem
> is the following: if a person uses HTML_Template_IT package he most
> probably does not have HTML_Template_Sigma installed (and vice versa),
> thus trying to autoload the class will fail! How should I fix this problem?

So?
is_a() returns false, fine isn't it?
Nothing to fix.

Regards,
--
Michael - < mike( <at> )php.net >
Permalink | Reply | 2 comments |
headers
Alexey Borzov | 1 Aug 2005 14:36
Picon
Favicon

Re: Re: A question to __autoload() fanboys

Hi,

Michael Wallner wrote:
>>OK, I *can* fix the case of 'html_template_sigma' here, but the problem
>>is the following: if a person uses HTML_Template_IT package he most
>>probably does not have HTML_Template_Sigma installed (and vice versa),
>>thus trying to autoload the class will fail! How should I fix this problem?
> 
> 
> So?
> is_a() returns false, fine isn't it?
> Nothing to fix.

Yes, that was exactly my point.

--

-- 
PEAR Development Mailing List (http://pear.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
Permalink | Reply | 1 comment |

Gmane