Tagging missing pieces...

[Elias Torres <el...@torrez.us>]

Hi Guys,

I'm out for the day (classes, meetings, etc) so I won't be able to
develop but I want to give you a heads up before you deploy to BSC that
I found a couple of missing things in the tagging implementation. Sorry
for the bad timing.

Basically, I found that in one of the VM models (I believe URLModel)
where we compute URLs $url.entries.rss(), ... we are taking in
$model.categoryPath, but we are not taking tags, so any URLs like <link
rel="self"/> in feeds will not contain the tags from the request. Plus I
noticed that I did not check the deprecated VM files that could need
access to tags.

One more thing, I also need to add getPopularTags() to the PageModel
like I did for SiteModel().

I could get to these changes tomorrow, but Allen, maybe you could help
me if you guys are releasing today.

BTW, are we branching 3.1 today?

-Elias

Re: Tagging missing pieces...

[Allen Gilliland <al...@sun.com>]

Elias,

I just realized that we still don't have a way to form feed urls by tags 
and the URLModel method for getting a url to a tag page only accepts a 
single tag, not a list.  We need to get those added for 3.1.

-- Allen


Elias Torres wrote:
> 
> Allen Gilliland wrote:
>> I actually have one to add to the list.  Comments below ...
>>
>> Elias Torres wrote:
>>> Hi Guys,
>>>
>>> I'm out for the day (classes, meetings, etc) so I won't be able to
>>> develop but I want to give you a heads up before you deploy to BSC that
>>> I found a couple of missing things in the tagging implementation. Sorry
>>> for the bad timing.
>>>
>>> Basically, I found that in one of the VM models (I believe URLModel)
>>> where we compute URLs $url.entries.rss(), ... we are taking in
>>> $model.categoryPath, but we are not taking tags, so any URLs like <link
>>> rel="self"/> in feeds will not contain the tags from the request. Plus I
>>> noticed that I did not check the deprecated VM files that could need
>>> access to tags.
>>
>> Yes, we can add a new method for getting a url to rss/atom feeds by tag
>> in the URLModel, that's easy to do.  However, we are not going to make
>> any changes to the deprecated models and macros to add new
>> functionality.  Those should only get changed to ensure that they
>> continue working, but not for any other reason.
>>
>>
> 
> good.
> 
>>> One more thing, I also need to add getPopularTags() to the PageModel
>>> like I did for SiteModel().
>>
>> Nope, you already have it in WebsiteData, so it's available via
>> $model.weblog.getPopularTags()
> 
> how quickly do I forget.
> 
>>
>> However, one item we haven't figured out which should probably be
>> addressed is what to do about tag urls which specify tags which don't
>> exist.  i.e. /weblog/tags/lkweehlkwer
>>
>> The way things are setup right now that is a valid page which will
>> simply not show any entries, and I don't think that's what we want to do
>> for a couple reasons.  1) It's a bit inefficient and silly to have *all*
>> tag urls be assumed correct and 2) It's horrible for the cache since
>> that means a rogue spider or malicious user could completely empty the
>> cache of valid content by filling it with pages/feeds from bogus tag urls.
>>
>> I think what we should probably do is add some tag checking to the
>> validation sections of both the page and feed servlets to actually check
>> that if tags were specified that they actually exist.  if none of the
>> tags specified exist then we can offer a 404.
> 
> Sounds like the right thing to do.
> 
>>
>>> I could get to these changes tomorrow, but Allen, maybe you could help
>>> me if you guys are releasing today.
>>
>> I will take a look at what I can today and see if I can get these done.
> 
> Let me know if you made any progress on it.
> 
>> -- Allen
>>
>>
>>> BTW, are we branching 3.1 today?
>>>
>>> -Elias

Re: Tagging missing pieces...

[Allen Gilliland <al...@sun.com>]

Elias,

do you have any time to look at these items?

-- Allen


Allen Gilliland wrote:
> 
> 
> Elias Torres wrote:
>>
>> Allen Gilliland wrote:
>>> I actually have one to add to the list.  Comments below ...
>>>
>>> Elias Torres wrote:
>>>> Hi Guys,
>>>>
>>>> I'm out for the day (classes, meetings, etc) so I won't be able to
>>>> develop but I want to give you a heads up before you deploy to BSC that
>>>> I found a couple of missing things in the tagging implementation. Sorry
>>>> for the bad timing.
>>>>
>>>> Basically, I found that in one of the VM models (I believe URLModel)
>>>> where we compute URLs $url.entries.rss(), ... we are taking in
>>>> $model.categoryPath, but we are not taking tags, so any URLs like <link
>>>> rel="self"/> in feeds will not contain the tags from the request. 
>>>> Plus I
>>>> noticed that I did not check the deprecated VM files that could need
>>>> access to tags.
>>>
>>> Yes, we can add a new method for getting a url to rss/atom feeds by tag
>>> in the URLModel, that's easy to do.  However, we are not going to make
>>> any changes to the deprecated models and macros to add new
>>> functionality.  Those should only get changed to ensure that they
>>> continue working, but not for any other reason.
>>>
>>>
>>
>> good.
> 
> 
> I didn't get this done yesterday because I ran into issues doing an 
> array in velocity, so maybe someone with better velocity skills can 
> tackle this.  I had added these 2 methods to the URLModel ...
> 
> public String rss(String[] tags, boolean excerpts) {
>             return URLUtilities.getWeblogFeedURL(weblog, locale, 
> "entries", "rss", null, Arrays.asList(tags), excerpts, true);
>         }
> 
> public String atom(String[] tags, boolean excerpts) {
>             return URLUtilities.getWeblogFeedURL(weblog, locale, 
> "entries", "atom", null, Arrays.asList(tags), excerpts, true);
>         }
> 
> 
>>
>>>> One more thing, I also need to add getPopularTags() to the PageModel
>>>> like I did for SiteModel().
>>>
>>> Nope, you already have it in WebsiteData, so it's available via
>>> $model.weblog.getPopularTags()
>>
>> how quickly do I forget.
>>
>>>
>>> However, one item we haven't figured out which should probably be
>>> addressed is what to do about tag urls which specify tags which don't
>>> exist.  i.e. /weblog/tags/lkweehlkwer
>>>
>>> The way things are setup right now that is a valid page which will
>>> simply not show any entries, and I don't think that's what we want to do
>>> for a couple reasons.  1) It's a bit inefficient and silly to have *all*
>>> tag urls be assumed correct and 2) It's horrible for the cache since
>>> that means a rogue spider or malicious user could completely empty the
>>> cache of valid content by filling it with pages/feeds from bogus tag 
>>> urls.
>>>
>>> I think what we should probably do is add some tag checking to the
>>> validation sections of both the page and feed servlets to actually check
>>> that if tags were specified that they actually exist.  if none of the
>>> tags specified exist then we can offer a 404.
>>
>> Sounds like the right thing to do.
> 
> 
> I think this is something that we need to go ahead an do as soon as we 
> can, because the more I think about it the more I am sure that this 
> needs to be in the release.
> 
> Unfortunately it doesn't seem like we ever actually lookup the TagData 
> objects during the request cycle, so maybe we need a new 
> tagsExist(weblog, tags) method in the manager and then call it from page 
> & feed servlets to do validation?  Would that be the most efficient way 
> to validate the tags entered by the user?
> 
> 
>>
>>>
>>>> I could get to these changes tomorrow, but Allen, maybe you could help
>>>> me if you guys are releasing today.
>>>
>>> I will take a look at what I can today and see if I can get these done.
>>
>> Let me know if you made any progress on it.
> 
> 
> Unfortunately no, didn't get anything done on these items yesterday.
> 
> -- Allen
> 
> 
>>
>>> -- Allen
>>>
>>>
>>>> BTW, are we branching 3.1 today?
>>>>
>>>> -Elias

Re: Tagging missing pieces...

[Allen Gilliland <al...@sun.com>]

Elias Torres wrote:
> 
> Allen Gilliland wrote:
>> I actually have one to add to the list.  Comments below ...
>>
>> Elias Torres wrote:
>>> Hi Guys,
>>>
>>> I'm out for the day (classes, meetings, etc) so I won't be able to
>>> develop but I want to give you a heads up before you deploy to BSC that
>>> I found a couple of missing things in the tagging implementation. Sorry
>>> for the bad timing.
>>>
>>> Basically, I found that in one of the VM models (I believe URLModel)
>>> where we compute URLs $url.entries.rss(), ... we are taking in
>>> $model.categoryPath, but we are not taking tags, so any URLs like <link
>>> rel="self"/> in feeds will not contain the tags from the request. Plus I
>>> noticed that I did not check the deprecated VM files that could need
>>> access to tags.
>>
>> Yes, we can add a new method for getting a url to rss/atom feeds by tag
>> in the URLModel, that's easy to do.  However, we are not going to make
>> any changes to the deprecated models and macros to add new
>> functionality.  Those should only get changed to ensure that they
>> continue working, but not for any other reason.
>>
>>
> 
> good.


I didn't get this done yesterday because I ran into issues doing an 
array in velocity, so maybe someone with better velocity skills can 
tackle this.  I had added these 2 methods to the URLModel ...

public String rss(String[] tags, boolean excerpts) {
             return URLUtilities.getWeblogFeedURL(weblog, locale, 
"entries", "rss", null, Arrays.asList(tags), excerpts, true);
         }

public String atom(String[] tags, boolean excerpts) {
             return URLUtilities.getWeblogFeedURL(weblog, locale, 
"entries", "atom", null, Arrays.asList(tags), excerpts, true);
         }


> 
>>> One more thing, I also need to add getPopularTags() to the PageModel
>>> like I did for SiteModel().
>>
>> Nope, you already have it in WebsiteData, so it's available via
>> $model.weblog.getPopularTags()
> 
> how quickly do I forget.
> 
>>
>> However, one item we haven't figured out which should probably be
>> addressed is what to do about tag urls which specify tags which don't
>> exist.  i.e. /weblog/tags/lkweehlkwer
>>
>> The way things are setup right now that is a valid page which will
>> simply not show any entries, and I don't think that's what we want to do
>> for a couple reasons.  1) It's a bit inefficient and silly to have *all*
>> tag urls be assumed correct and 2) It's horrible for the cache since
>> that means a rogue spider or malicious user could completely empty the
>> cache of valid content by filling it with pages/feeds from bogus tag urls.
>>
>> I think what we should probably do is add some tag checking to the
>> validation sections of both the page and feed servlets to actually check
>> that if tags were specified that they actually exist.  if none of the
>> tags specified exist then we can offer a 404.
> 
> Sounds like the right thing to do.


I think this is something that we need to go ahead an do as soon as we 
can, because the more I think about it the more I am sure that this 
needs to be in the release.

Unfortunately it doesn't seem like we ever actually lookup the TagData 
objects during the request cycle, so maybe we need a new 
tagsExist(weblog, tags) method in the manager and then call it from page 
& feed servlets to do validation?  Would that be the most efficient way 
to validate the tags entered by the user?


> 
>>
>>> I could get to these changes tomorrow, but Allen, maybe you could help
>>> me if you guys are releasing today.
>>
>> I will take a look at what I can today and see if I can get these done.
> 
> Let me know if you made any progress on it.


Unfortunately no, didn't get anything done on these items yesterday.

-- Allen


> 
>> -- Allen
>>
>>
>>> BTW, are we branching 3.1 today?
>>>
>>> -Elias

Re: Tagging missing pieces...

[Elias Torres <el...@torrez.us>]

Allen Gilliland wrote:
> I actually have one to add to the list.  Comments below ...
> 
> Elias Torres wrote:
>> Hi Guys,
>>
>> I'm out for the day (classes, meetings, etc) so I won't be able to
>> develop but I want to give you a heads up before you deploy to BSC that
>> I found a couple of missing things in the tagging implementation. Sorry
>> for the bad timing.
>>
>> Basically, I found that in one of the VM models (I believe URLModel)
>> where we compute URLs $url.entries.rss(), ... we are taking in
>> $model.categoryPath, but we are not taking tags, so any URLs like <link
>> rel="self"/> in feeds will not contain the tags from the request. Plus I
>> noticed that I did not check the deprecated VM files that could need
>> access to tags.
> 
> 
> Yes, we can add a new method for getting a url to rss/atom feeds by tag
> in the URLModel, that's easy to do.  However, we are not going to make
> any changes to the deprecated models and macros to add new
> functionality.  Those should only get changed to ensure that they
> continue working, but not for any other reason.
> 
> 

good.

>>
>> One more thing, I also need to add getPopularTags() to the PageModel
>> like I did for SiteModel().
> 
> 
> Nope, you already have it in WebsiteData, so it's available via
> $model.weblog.getPopularTags()

how quickly do I forget.

> 
> 
> However, one item we haven't figured out which should probably be
> addressed is what to do about tag urls which specify tags which don't
> exist.  i.e. /weblog/tags/lkweehlkwer
> 
> The way things are setup right now that is a valid page which will
> simply not show any entries, and I don't think that's what we want to do
> for a couple reasons.  1) It's a bit inefficient and silly to have *all*
> tag urls be assumed correct and 2) It's horrible for the cache since
> that means a rogue spider or malicious user could completely empty the
> cache of valid content by filling it with pages/feeds from bogus tag urls.
> 
> I think what we should probably do is add some tag checking to the
> validation sections of both the page and feed servlets to actually check
> that if tags were specified that they actually exist.  if none of the
> tags specified exist then we can offer a 404.

Sounds like the right thing to do.

> 
> 
>>
>> I could get to these changes tomorrow, but Allen, maybe you could help
>> me if you guys are releasing today.
> 
> 
> I will take a look at what I can today and see if I can get these done.

Let me know if you made any progress on it.

> 
> -- Allen
> 
> 
>>
>> BTW, are we branching 3.1 today?
>>
>> -Elias
> 

Re: Tagging missing pieces...

[Allen Gilliland <al...@sun.com>]

I actually have one to add to the list.  Comments below ...

Elias Torres wrote:
> Hi Guys,
> 
> I'm out for the day (classes, meetings, etc) so I won't be able to
> develop but I want to give you a heads up before you deploy to BSC that
> I found a couple of missing things in the tagging implementation. Sorry
> for the bad timing.
> 
> Basically, I found that in one of the VM models (I believe URLModel)
> where we compute URLs $url.entries.rss(), ... we are taking in
> $model.categoryPath, but we are not taking tags, so any URLs like <link
> rel="self"/> in feeds will not contain the tags from the request. Plus I
> noticed that I did not check the deprecated VM files that could need
> access to tags.


Yes, we can add a new method for getting a url to rss/atom feeds by tag 
in the URLModel, that's easy to do.  However, we are not going to make 
any changes to the deprecated models and macros to add new 
functionality.  Those should only get changed to ensure that they 
continue working, but not for any other reason.


> 
> One more thing, I also need to add getPopularTags() to the PageModel
> like I did for SiteModel().


Nope, you already have it in WebsiteData, so it's available via 
$model.weblog.getPopularTags()


However, one item we haven't figured out which should probably be 
addressed is what to do about tag urls which specify tags which don't 
exist.  i.e. /weblog/tags/lkweehlkwer

The way things are setup right now that is a valid page which will 
simply not show any entries, and I don't think that's what we want to do 
for a couple reasons.  1) It's a bit inefficient and silly to have *all* 
tag urls be assumed correct and 2) It's horrible for the cache since 
that means a rogue spider or malicious user could completely empty the 
cache of valid content by filling it with pages/feeds from bogus tag urls.

I think what we should probably do is add some tag checking to the 
validation sections of both the page and feed servlets to actually check 
that if tags were specified that they actually exist.  if none of the 
tags specified exist then we can offer a 404.


> 
> I could get to these changes tomorrow, but Allen, maybe you could help
> me if you guys are releasing today.


I will take a look at what I can today and see if I can get these done.

-- Allen


> 
> BTW, are we branching 3.1 today?
> 
> -Elias