################################################################################# # # Product: Reptile # Refresh-URL: http://bugzilla.openprivacy.org/dump-bugs.cgi?product=Reptile # Add-URL: http://bugzilla.openprivacy.org/enter_bug.cgi?product=Reptile # Base-URL: http://bugzilla.openprivacy.org/ # Author: Kevin A. Burton ( burton@apache.org | burton@openprivacy.org ) # Last-modified: Wed Feb 27 02:13:11 2002 # Version: $Id: TODO.bugzilla,v 1.51 2002/02/27 10:33:04 burton Exp $ # Protocol-version: 1.0.2 # Commentary: # # Do not modify any of the above headers when using under an offline application # such as Emacs bugzilla-mode. # # This is a static dump of the bug database. Since this is a public database it # may be modified by other users. You should update often. # # More info available at http://relativity.yi.org/bugoff. # ################################################################################# --------------------------------------------------------------------------------- ID: 1 Short Description: Need to cache username Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=1 Product: Reptile Version: 0.6.0 Platform: Other Operating System: other Status: NEW Severity: blocker Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: When we login we ALWAYS have to enter a new username. We should cache this value. (cookie?) --------------------------------------------------------------------------------- ID: 9 Short Description: The Reptile sequence parsing system needs to be rewritten. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=9 Product: Reptile Version: 0.0.1 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: - only normalize SequenceDispatchEntries when necessary. - only parse sequence-dispatch.xml when necessary - keep everything in a SequenceManager. --------------------------------------------------------------------------------- ID: 25 Short Description: Fully support dublin core metainfo in RSS feeds. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=25 Product: Reptile Version: 0.0.2 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: http://www.oreillynet.com/meerkat/index.php?&p=760 AKA display the following... like meerkat... [Creator: Steve Anglin, Subject: Java, Publisher: O'Reilly and Associates, Date: 2001-08-02 10:42:42, Format: text/html, Language: en-us, Rights: Copyright 2001, O'Reilly and Associates] This should be fairly easy but only show it for the item when we are editing a channel. ------- Additional Comments From burton@relativity.yi.org 2002-02-27 01:50 ------- needs to be done in meta-update --------------------------------------------------------------------------------- ID: 27 Short Description: Rewrite the ContenTransformer system into a single mapping (with metainfo) (ContentTransformerMap?) Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=27 Product: Reptile Version: 0.0.2 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: I need to do a lot of research into the best way to implement this. It would be nice if we could use this new mechanism to parse out content-transformers.xml once on startup and everytime the file changes. We should also support a singleton and an instance object which includes the metainfo. This should also allow us to map "preferred" input formats for specific output formats. --------------------------------------------------------------------------------- ID: 30 Short Description: Display session timeout ETA in browser. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=30 Product: Reptile Version: 0.0.2 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: It is VERY annoying to look at a Reptile in an browser and not know how stale of an instance this is. IE there should be some way for me to look at a Reptile window and tell it my session has expired or not. It should be easy to put a little control in the top left which would count down, in minutes, the remaining time in my session. IE Session: 5 minutes to timeout Write some javascript which will count this down within the browser and timeout when necessary. Display an ERROR in red when the session has timed out. --------------------------------------------------------------------------------- ID: 35 Short Description: Should be able to Create, Update, Delete layers Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=35 Product: Reptile Version: 0.0.2 Platform: Other Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: nigel@nuix.com.au Component: default URL: http://localhost:8050/reptile/servlet/reptile;jsessionid=urv28hj5k1?reptile.sequence.name=urn:settings Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: Which layers are present is an essential user customisation feature. Thus a user should be able to 1. create layers, 2. delete them (with a warning that this will unsubscribe them from all the chanels within that layer) and 3. alter the name of a layer. Ideally they should be able to move channels between layers, but this might be functionality relating to management of channels (add, delete etc). ------- Additional Comments From nigel@nuix.com.au 2001-09-12 23:06 ------- Sorry, yes I do realise you can add layers, and can add channels to layers, but deleting/renaming them still remains an essential customisation feature ------- Additional Comments From burton@relativity.yi.org 2001-09-13 01:04 ------- Until this make it into an official build... you can rename a 'layer' by editing your 'subscriptions.xml' file in your reptile installation (usually REPTILE/webapps/reptile/xml/users/default/subscriptions.xml and just change the layer attribute to the new name. --------------------------------------------------------------------------------- ID: 36 Short Description: Need a new feature 'mark for later reading' Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=36 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: Often I come across an article that while interesting, I don't have the time to read at the current moment. We need a new feature so that we can mark an article for reading at a later date. This would also need a page so that we can view all information (description, etc) of the article in question. --------------------------------------------------------------------------------- ID: 41 Short Description: Weblog missing functionality. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=41 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: - can't edit an existing item - can't delete an existing item. --------------------------------------------------------------------------------- ID: 42 Short Description: OCS streams don't have correct information. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=42 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: We need to have a way to produce our *own* OCS based on the 5 other OCS streams we know about. The problem is that most of the OCS feeds we know about don't have correct information. We need to write a system which can pull own the OCS feeds and also pull own their content so that we can produce a better OCS feed with correct information. - also include images. ------- Additional Comments From burton@relativity.yi.org 2001-12-18 00:26 ------- At the very minimum we should try to cache this information on disk ahead of time. We also need to write a Content detection mechanism which could determine the content type of a given XML feed. Basically we need to assume that everyone is going to lie to us. --------------------------------------------------------------------------------- ID: 48 Short Description: Include page/content generation time at the bottom of page.xml Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=48 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: Include statistics of page generation time at the bottom of each page. This will be usefull in the long term as more people can assess the performace of Reptile. --------------------------------------------------------------------------------- ID: 52 Short Description: Need to include a 'quickbar' at the top of new URLs Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=52 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: I need to have the following in the 'quick bar' when I jump to new URLs - search box - weblog publish link. The only problem with this is that we need to use a form for POST data and not a link. Is it possible to do this via post? SUPPLEMENTAL: // created on Mon Sep 24 2001 02:53 PM If we include the description the URL is going to be too long and cause problems. END_SUPPLEMENTAL ------- Additional Comments From burton@relativity.yi.org 2001-09-24 15:24 ------- Actually. If we do this correctly all we will need is a URL. When we migrate everything over to using RSS 1.0 and channel history, we can pull the description out of the history and we won't have to pass it. ------- Additional Comments From burton@relativity.yi.org 2001-09-26 11:13 ------- This bar also needs support for marking the current url 'read' --------------------------------------------------------------------------------- ID: 53 Short Description: Reptile should be able to cache remote HTML locally. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=53 Product: Reptile Version: 0.6.0 Platform: PC Operating System: Linux Status: NEW Severity: blocker Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: We should be able to write a module so that we can bring down remote content locally so that we can view the local file (with all CSS, Javascript, images mirrored correctly) instead of the remote one. This will allow us to: - view content offline - add support for something like lucene for indexing the content. - get the content even if there is a network outage like on the 911 tragedy. ------- Additional Comments From burton@relativity.yi.org 2001-09-24 15:54 ------- This shouldn't be too bad if we use something like jakarta regexp to rewrite and mirror certain content. --------------------------------------------------------------------------------- ID: 54 Short Description: Only update the panther cache if content is 'valid' Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=54 Product: Reptile Version: 0.6.0 Platform: PC Operating System: Linux Status: ASSIGNED Severity: blocker Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: Every once and a while, we will do a cache update, pull down invalid content, and then seriously blow away that control. The problem is that at time interval T1 everything works fine. At time interval T2 the cache is stale so Reptile does an update. At time interval T3 the cache is updated but we get invalid XML. This can happen in a NUMBER of situations including: - the content producer created invalid XML - the HTTP channel was redirected or modified. This can happen in the case of 802.11 roaming where they do captured portal style authentication. Also if the user uses a proxy server this can happen. We need a way to have cache updates become tranactional. If any hooks fail we should abort the content update. These hooks need to be run BEFORE we save the updated content to disk. ------- Additional Comments From burton@relativity.yi.org 2002-02-15 17:56 ------- OK... This NEEDS to be fixed before 0.6.0 Basically when I am sitting at Starbucks using Mobilestar, if I am not logged in I pull down content which is text/plain and still valid but is HTML. I don't know if there is any way around this. Try to setup a URL connection to an external site and see if Mobilestar is using HTTP correctly. --------------------------------------------------------------------------------- ID: 56 Short Description: Should be possible to localize and version control the DB. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=56 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: We could install the the database as HOME/.reptile/db_rev_0.0.1 There are some problems here. Basically I would have to copy the DB (with schema applied) from the reptile install into REPTILE_HOME if it doesn't exist. This way we can update reptile from a previous version and if the DB is compatible, we can use it. ------- Additional Comments From burton@relativity.yi.org 2001-11-28 12:22 ------- This could be done by a standardized XML dump. Torque could provide this functionality --------------------------------------------------------------------------------- ID: 57 Short Description: Local weblog needs to be implicitly available via OCS. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=57 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: Make sure that the local weblog is implicitly available via the OCS search feature. This is required because we need to be able to search for other user's content via JXTA or Freenet. ------- Additional Comments From burton@relativity.yi.org 2001-09-28 18:17 ------- Working now. We were requiring that content be updated by the proxy before running the handler. --------------------------------------------------------------------------------- ID: 59 Short Description: Reptile needs a plugin architecture. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=59 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: We need a way to have Reptile support plugins. This would be a system where we could plugin a significant amount of flexibility without having to break/modify the core. - deployment descriptors - support the ability to run tasks - ability to register SOAP endpoints. - ability to modify the UI so that we can install our own XSLT. - should support their own databases (Hypersonic) This will be documented under /docs/proposals/PLUGIN_ARCHITECTURE --------------------------------------------------------------------------------- ID: 65 Short Description: Search results page should include item metainfo Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=65 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: Each item should include: - favicon - channel title - layer it is in. This should be with an enable-feature mechanism so that RSS can enable certain features in the presentation. --------------------------------------------------------------------------------- ID: 66 Short Description: Rewrite the way Favicons are discovered. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=66 Product: Reptile Version: 0.5.1 Platform: PC Operating System: Linux Status: ASSIGNED Severity: blocker Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: The current favicon icon impl is broken. 1. It requires at least one operation to pull down the favicon (the update is a side effect). 2. It is WAY too slow. Instead.. Rename Favicon.java FaviconExtension.java This should then only use the DB to pull out Favicons. The MetaUpdate extension should figure out if this favicon exists and if it does it should upate the DB. ------- Additional Comments From burton@relativity.yi.org 2001-12-01 22:25 ------- We also need to be smart enough to make sure this is a binary file. Sometimes we can get back text or even html ------- Additional Comments From burton@relativity.yi.org 2001-12-20 16:24 ------- WE shoud also update nonstandard Favicons. AKA an article with the following: We should be able to use this to update the Favicon table. Note that this depends on the HTCache integration --------------------------------------------------------------------------------- ID: 69 Short Description: Add support for Lucene Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=69 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: After we add support for the HTCache, we should be able to index (with lucene) the cached content. We should then provide another search index for this (text search) --------------------------------------------------------------------------------- ID: 71 Short Description: Register new users functionality Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=71 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: Need the ability to register new "users" with the system. This should be a DB table of USER which contains additional metainfo. Each one of these users is just used for preference storage within Reptile. These will be backed with multiple nyms in the future. --------------------------------------------------------------------------------- ID: 76 Short Description: Create a SEARCH_FORMS table for keeping track of RSS forms. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=76 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: This way we can run arbitrary searches on any site. --------------------------------------------------------------------------------- ID: 78 Short Description: HTParser integration. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=78 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: We need to support the HTParser system for trying to guess the description of an HTML file. This should be run from the MetaUpdate system after we fetch an HTML document and it is in the HTCache. ------- Additional Comments From burton@relativity.yi.org 2001-12-01 20:52 ------- - This should support finding the description for RSS items that don't have descriptions. This is really annoying when we want to view a channel and there is no description for --------------------------------------------------------------------------------- ID: 79 Short Description: Support Reputation Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=79 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: - Initial support for reputation within Reptile via Sierra. This depends on Fionna which is still pending. --------------------------------------------------------------------------------- ID: 80 Short Description: Consider changing the MetaUpdate system to support SAX instead of XSLT Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=80 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: ASSIGNED Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: Right now our MetaUpdate system uses XSLT, sequence and the content-transformer system. This very bloated. - requires that we init Xalan correctly - has startup overhead of forcing stylesheet compiles - content won't be in the cache so we will get cache failures - Anything non RSS 1.0 will require 3 Transforms If we rewrite this to use a SAX handler we would be able to startup MUCH faster. If we still used RSS as an input for our channels these would display slower due to cache failures. This will probably be rewritten to use the DB though. We should also do some benchmarking here to prove that this would be faster. The only drawback here is that we will have to write a new SAX parser for EVERY format. This won't be too bad because they are all very similar. The MetaUpdate system could still be used because we would offload from the SAX parser directly to the specific UpdateX class. ------- Additional Comments From burton@relativity.yi.org 2001-12-01 22:30 ------- This has the ADDED advantage (an important one) that we could index other RSS channels as well as the current channels. This way we can update them in the background so that we have ALL articles indexed and ALL channels would have updated (correct) descriptions. ------- Additional Comments From burton@relativity.yi.org 2001-12-02 13:27 ------- Perhaps we should wait until XSLTC is integrated. This way we can see if the performance increase would merit rewriting the current system. ------- Additional Comments From burton@relativity.yi.org 2001-12-19 12:51 ------- Before we do this we should be careful. The current system works fine, it is just slow. We should wait until the Sequence pipe is faster. When we are using SAX and XSLTC we can see how fast it is. Unfortunately I think we have to wait a file for this. --------------------------------------------------------------------------------- ID: 82 Short Description: Need the ability to jump to specific subscriptions. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=82 Product: Reptile Version: 0.1.0 Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: We need the ability to jump to a specific subscription *quickly*. AKA a control at the top of the page or to the write of "Category". --------------------------------------------------------------------------------- ID: 86 Short Description: Escape descriptions with style anchors Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=86 Product: Reptile Version: 0.6.0 Platform: PC Operating System: Linux Status: NEW Severity: blocker Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: This is part of RSS 0.92 so we have to support it. The following RSS feed uses this http://jakarta.radiopossibility.com/xml/rss.xml The content is escaped inline as <a href=&qt; If we have a correct anchor. Go ahead and escape it, shouldn't hurt anything and the regexp is somewhat OK. ------- Additional Comments From burton@relativity.yi.org 2001-12-17 18:02 ------- use jakarta regexp for this ------- Additional Comments From burton@relativity.yi.org 2001-12-17 18:55 ------- I have already started work on this. It is in EscapeExtension.html() ------- Additional Comments From burton@relativity.yi.org 2001-12-18 00:21 ------- It turns out that if we just disable output escaping we will get the text raw. So if <a href= would just be without closing it. Dont konw --------------------------------------------------------------------------------- ID: 88 Short Description: Removed redundant HTML elements Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=88 Product: Reptile Version: 0.6.0 Platform: PC Operating System: Linux Status: NEW Severity: major Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: We need to escape elements so that we remove STUPID HTML encoding by non-xml presentation engines. Specifically don't allow things like in a title. --------------------------------------------------------------------------------- ID: 95 Short Description: Integrate the HTCache system so that remote URLs are kept on disk. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=95 Product: Reptile Version: 0.6.0 Platform: PC Operating System: Linux Status: NEW Severity: blocker Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: We need to integrate this into Reptile. --------------------------------------------------------------------------------- ID: 100 Short Description: Hypersonic database connections represent the on-disk data. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=100 Product: Reptile Version: 0.7.0 Platform: PC Operating System: Linux Status: ASSIGNED Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: The DB connection at jdbc:hsqldb:http://localhost:8050/reptile/servlet/hypersonic Is loaded from DISK and not from memory. See if we can get this to represented on disk content. Maybe the servlet should load from memory? ------- Additional Comments From burton@relativity.yi.org 2001-12-23 00:10 ------- It doesn't look like this is possible. I changed the HypersonicServlet's 'database' to . and it created a NEW db. We need to figure out how to fetch the one that is used and has been initialized on disk. --------------------------------------------------------------------------------- ID: 103 Short Description: We need to initialize all OCS channels. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=103 Product: Reptile Version: unspecified Platform: PC Operating System: Linux Status: ASSIGNED Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: When Reptile is first installed, we need to initialze all RSS channels. Right now our OCS providers aren't giving us complete or accurate information. Specifically we don't have: The image for the channel. The title and description for the channel. This should only happen ONCE. We need to have an INITIAILZED column on the CHANNEL table. If INITIALIZED is 0 then we should go ahead and update the DB. This is a major problem for us and it needs to be done RIGHT. - We can't create too many threads - We can't update too often as each channel will take from 1-5 seconds to parse. - We should take a break between each update so that the users CPU isn't at MAX for long periods of time (AKA yield). ------- Additional Comments From burton@relativity.yi.org 2001-12-23 13:33 ------- The only problem here is that we need to estimate the performance overhead here. Specifically if we have too much overhead there is NO way that we could process this many channels. In order to do this we are going to have to spend a LOT of time optimizing this. It could be possible to get this down to a 30 minute initialization time but this would require only 1/2 second of parsing time per channel. To be realistic we will have to get this down to a few hunder miliseconds. ------- Additional Comments From burton@relativity.yi.org 2001-12-23 13:34 ------- And this may require a dedicated feed initialization system. One that has a ChannelInitializer for each channel. ------- Additional Comments From burton@relativity.yi.org 2001-12-23 13:39 ------- Also, we should document this system, how it uses a threadpool, how it keeps from using too many resources, etc. --------------------------------------------------------------------------------- ID: 104 Short Description: Need to support RSS 0.92 and CDF Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=104 Product: Reptile Version: unspecified Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: Right quite a number of feeds aren't supported because we don't have stylesheets for them: http://www.scripting.com/dtd/rss-0.92.dtd http://www.moreover.com/xml/wddx_0090.dtd http://msdn.microsoft.com/workshop/delivery/channel/overview/overview.asp http://my.netscape.com/publish/formats/rss-0.9 Added up, this is probably another 1000 or so feeds. We need ot fix this before 1.0. --------------------------------------------------------------------------------- ID: 106 Short Description: Documentation: add a "design" document. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=106 Product: Reptile Version: unspecified Platform: PC Operating System: Linux Status: ASSIGNED Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: This should include the following - the Reptile distribution/bundle mechanism - The way Reptile functions under Tomcat with Hypersonic, Torque, etc. - Libraries we use and what we are using them for. - Network protocol like stack. - how we plug into decentralized networks. - Tasks system ------- Additional Comments From burton@relativity.yi.org 2002-01-02 20:49 ------- We also need to document how the database is setup. - we use torque - reptile-schema.xml - database tables. Is there a way to generate an entity relationship diagram? --------------------------------------------------------------------------------- ID: 110 Short Description: Document XML Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=110 Product: Reptile Version: unspecified Platform: PC Operating System: Linux Status: ASSIGNED Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: We need to document our XML namespace and how it is used. This is a core portion of Reptile and it needs to be under "XML" in the Table of Contents. ------- Additional Comments From burton@relativity.yi.org 2002-01-04 10:28 ------- We should also describe the used of mozilla-error.xsl which is used to provide a default UI for errors. --------------------------------------------------------------------------------- ID: 111 Short Description: Need Pipe mechanism for transferring data via SOAP, HTTP, JXTA, etc. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=111 Product: Reptile Version: unspecified Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: We need a way to get non HTTP traffic into Reptile. Specifically how do we get content from multiple network transports into the system. - BEEP - JXTA - HTTP - SOAP This should become a propoal. What should this be called? Pipe? Passage? Tube? ------- Additional Comments From burton@relativity.yi.org 2002-02-27 02:12 ------- This should be a Panther 'ProtocolAdapter' which supports regexp of the URL --------------------------------------------------------------------------------- ID: 112 Short Description: Proposal: Document a micropayment plugin system. Edit URL: http://bugzilla.openprivacy.org/show_bug.cgi?id=112 Product: Reptile Version: unspecified Platform: PC Operating System: Linux Status: NEW Severity: normal Priority: P2 Resolution: Assigned: burton@relativity.yi.org Reporter: burton@relativity.yi.org Component: default URL: Target Milestone: --- QA Contact: 0 Status Whiteboard: Keywords: Description: This seems very straight forward. Document a system that allows us to support micropayments.