Menu
▾
▴
fofredux-devel — Developers discussion, bug fixing, and coding of fofredux
You can subscribe to this list here.
| 2005 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(92) |
Dec
(142) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2006 |
Jan
(33) |
Feb
(65) |
Mar
(76) |
Apr
(172) |
May
(124) |
Jun
(45) |
Jul
(76) |
Aug
(78) |
Sep
(1) |
Oct
(20) |
Nov
(6) |
Dec
(2) |
| 2007 |
Jan
(12) |
Feb
(7) |
Mar
(17) |
Apr
(3) |
May
|
Jun
(2) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(1) |
|
From: Andrew T. (nilspace) <nil...@us...> - 2005-12-07 20:34:23
|
Hi Everyone, Since Kevin pointed out that there are some small issues, but nothing showstopping, I've gone ahead and tagged the current cvs RC-0_2. To check out do the following: cvs -d:use...@cv...:/cvsroot/fofredux co -r RC-0_2 -d fofredux_0_2 fofredux This will checkout the tag to a new directory (-d fofredux_0_2). You can then do any testing in there. Please file bug reports and mark the version and approximate importance. Also, is there any input on how we want to do documentation (besides small readme's in the folder). A Wiki is often nice, but sometimes can get confusing. Or are static pages good enough for now. An RSS feed of instructions? :) Any particular thoughts? Andy |
|
From: Kevin <ke...@dr...> - 2005-12-07 05:33:27
|
Andrew Turner (nilspace) wrote:
>Has anyone else tested out the new install.php I made. We need to make
>sure that it allows users to safely upgrade to the FoFRedux without
>losing their data - though we can also recommend people doing an OPML
>export/import.
>
>
>
When I do a clean install, I get this error.
/Creating indexes...
Can't create index in table frt_items. MySQL says: *Key column 'flag'
doesn't exist in table*/
Suggestion: What if the install page presented a menu of options before
doing anything. It would really simplify things if the user explicity
specified what should be done:
* clean install (throw error if any tables exist)
* upgrade FoF tables (alter tables to match new schema)
* migrate data from FoF tables (leaving original tables intact)
* upgrade from previous FoFr version (not necessary initially)
cheers,
Kevin
|
|
From: Andrew T. (nilspace) <nil...@us...> - 2005-12-07 04:39:27
|
Has anyone else tested out the new install.php I made. We need to make sure that it allows users to safely upgrade to the FoFRedux without losing their data - though we can also recommend people doing an OPML export/import. Once we are ready - we can tag the repository as RC-0_2, do testing, update, move the tags to the new RC, and when ready, tag it as Release-0_2 and make a tarball. Andy On 12/6/05, Kevin <ke...@dr...> wrote: > The among the outstanding bugs, I don't see a show stopper. Are we > ready to build a release? > > --- > Kevin Bennett > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log fi= les > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_id=3D7637&alloc_id=3D16865&op=3Dclick > _______________________________________________ > Fofredux-devel mailing list > Fof...@li... > https://lists.sourceforge.net/lists/listinfo/fofredux-devel > -- Andrew Turner ajt...@hi... 42.4266N x 83.4931W http://highearthorbit.com Northville, Michigan, USA Photos - http://flickr.com/photos/ajturner Travel - http://highearthorbit.com/projects/location/ |
|
From: Kevin <ke...@dr...> - 2005-12-07 02:06:41
|
The among the outstanding bugs, I don't see a show stopper. Are we ready to build a release? --- Kevin Bennett |
|
From: Kevin <ke...@dr...> - 2005-12-05 23:15:01
|
Agreed. So, I'll go ahead w/ the subdirectory for tests. > branching is not a good idea. it doesn't promote keeping them > together. Putting them in another directory is the best way to do it > then. That way, when we do branch/tag the FoFR code, the tests will > also be tagged/branch. > > Andrew > > On 12/5/05, Kevin <ke...@dr...> wrote: >> >> > Well, you could add a new project in the CVS repository - >> > fofredux-test or something similar. Anyone could check out the >> > fofredux project, and devs could subsequently checkout fofredux-test= : >> > >> > cvs -d:ext:nil...@cv...:/cvsroot/fofredux co fofredux >> > cvs -d:ext:nil...@cv...:/cvsroot/fofredux co fofredux-test >> >> The unit tests should evolve as the code does. They should be tagged >> and >> branched to follow the code that they test. Wouldn't putting the unit >> tests in their own module hinder this? >> >> Writing, running, and maintaining the unit tests should be made as eas= y >> as >> possible. Otherwise, they won't get used. That's why my first though= t >> was making it a subdirectory and just excluding it when creating >> tarballs >> for release. >> >> -- >> Kevin >> >> >> >> ------------------------------------------------------- >> This SF.net email is sponsored by: Splunk Inc. Do you grep through log >> files >> for problems? Stop! Download the new AJAX search engine that makes >> searching your log files as easy as surfing the web. DOWNLOAD SPLUNK= ! >> http://ads.osdn.com/?ad_idv37&alloc_id=16865&opclick >> _______________________________________________ >> Fofredux-devel mailing list >> Fof...@li... >> https://lists.sourceforge.net/lists/listinfo/fofredux-devel >> > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log > files > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_idv37&alloc_id=16865&op=3Dclick > _______________________________________________ > Fofredux-devel mailing list > Fof...@li... > https://lists.sourceforge.net/lists/listinfo/fofredux-devel > --=20 Kevin |
|
From: Andrew T. (nilspace) <nil...@us...> - 2005-12-05 22:15:40
|
branching is not a good idea. it doesn't promote keeping them together. Putting them in another directory is the best way to do it then. That way, when we do branch/tag the FoFR code, the tests will also be tagged/branch. Andrew On 12/5/05, Kevin <ke...@dr...> wrote: > > > Well, you could add a new project in the CVS repository - > > fofredux-test or something similar. Anyone could check out the > > fofredux project, and devs could subsequently checkout fofredux-test: > > > > cvs -d:ext:nil...@cv...:/cvsroot/fofredux co fofredux > > cvs -d:ext:nil...@cv...:/cvsroot/fofredux co fofredux-test > > The unit tests should evolve as the code does. They should be tagged and > branched to follow the code that they test. Wouldn't putting the unit > tests in their own module hinder this? > > Writing, running, and maintaining the unit tests should be made as easy a= s > possible. Otherwise, they won't get used. That's why my first thought > was making it a subdirectory and just excluding it when creating tarballs > for release. > > -- > Kevin > > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log fi= les > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_idv37&alloc_id=16865&opclick > _______________________________________________ > Fofredux-devel mailing list > Fof...@li... > https://lists.sourceforge.net/lists/listinfo/fofredux-devel > |
|
From: Kevin <ke...@dr...> - 2005-12-05 21:58:04
|
> Well, you could add a new project in the CVS repository - > fofredux-test or something similar. Anyone could check out the > fofredux project, and devs could subsequently checkout fofredux-test: > > cvs -d:ext:nil...@cv...:/cvsroot/fofredux co fofredux > cvs -d:ext:nil...@cv...:/cvsroot/fofredux co fofredux-test The unit tests should evolve as the code does. They should be tagged and branched to follow the code that they test. Wouldn't putting the unit tests in their own module hinder this? Writing, running, and maintaining the unit tests should be made as easy a= s possible. Otherwise, they won't get used. That's why my first thought was making it a subdirectory and just excluding it when creating tarballs for release. --=20 Kevin |
|
From: Andrew T. (nilspace) <nil...@us...> - 2005-12-05 16:50:29
|
Well, you could add a new project in the CVS repository - fofredux-test or something similar. Anyone could check out the fofredux project, and devs could subsequently checkout fofredux-test: cvs -d:ext:nil...@cv...:/cvsroot/fofredux co fofredux cvs -d:ext:nil...@cv...:/cvsroot/fofredux co fofredux-test On 12/5/05, Kevin <ke...@dr...> wrote: > Andrew Turner (nilspace) wrote: > > >Only concerned that we haven't done it yet and more often. ;) > > > >I think this is an excellent idea. It had crossed my mind, but got > >pushed back. Thanks for putting this together. I assume we can write > >tests for all the major functionality: adding, removing, viewing > >feeds, generated OPML, generated RSS (future feature)? > > > > > FoFr is a small project, but it can still benefit from unit testing. It > makes heavy use of functions to organize the backend code.(see also: "a > good thing") This type of code organization makes it easier to test. > > I figured each fof function will have it's own test case class. I've > already created one for fof_content_sanitize(). > > Question: Should the tests be included in code releases? I don't see > the benefit in that. I figure they will just live in the CVS repository > for developers to use. > > >I haven't seen SimpleTest before, I'll make sure to check it out. > > > > > The two choices are SimpleTest and PHPUnit. I've only ever used > SimpleTest. I gave the list a heads up in case someone wanted to pipe > up with compelling reasons to use PHPUnit instead. If no one has > objections, I'll commit what I have. > > > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log fi= les > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_id=3D7637&alloc_id=3D16865&op=3Dclick > _______________________________________________ > Fofredux-devel mailing list > Fof...@li... > https://lists.sourceforge.net/lists/listinfo/fofredux-devel > -- Andrew Turner ajt...@hi... 42.4266N x 83.4931W http://highearthorbit.com Northville, Michigan, USA Photos - http://flickr.com/photos/ajturner Travel - http://highearthorbit.com/projects/location/ |
|
From: Kevin <ke...@dr...> - 2005-12-05 14:30:44
|
Andrew Turner (nilspace) wrote: >Only concerned that we haven't done it yet and more often. ;) > >I think this is an excellent idea. It had crossed my mind, but got >pushed back. Thanks for putting this together. I assume we can write >tests for all the major functionality: adding, removing, viewing >feeds, generated OPML, generated RSS (future feature)? > > FoFr is a small project, but it can still benefit from unit testing. It makes heavy use of functions to organize the backend code.(see also: "a good thing") This type of code organization makes it easier to test. I figured each fof function will have it's own test case class. I've already created one for fof_content_sanitize(). Question: Should the tests be included in code releases? I don't see the benefit in that. I figure they will just live in the CVS repository for developers to use. >I haven't seen SimpleTest before, I'll make sure to check it out. > > The two choices are SimpleTest and PHPUnit. I've only ever used SimpleTest. I gave the list a heads up in case someone wanted to pipe up with compelling reasons to use PHPUnit instead. If no one has objections, I'll commit what I have. |
|
From: Andrew T. (nilspace) <nil...@us...> - 2005-12-05 04:06:36
|
Only concerned that we haven't done it yet and more often. ;) I think this is an excellent idea. It had crossed my mind, but got pushed back. Thanks for putting this together. I assume we can write tests for all the major functionality: adding, removing, viewing feeds, generated OPML, generated RSS (future feature)? I haven't seen SimpleTest before, I'll make sure to check it out. Andy On 12/4/05, Kevin <ke...@dr...> wrote: > > To help with debugging, I've started writing some unit tests using > SimpleTest. I plan on putting them in a subdirectory called "tests". > If anyone has any concerns reguarding this, let me know. > > --Kevin > > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log fi= les > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_id=3D7637&alloc_id=3D16865&op=3Dclick > _______________________________________________ > Fofredux-devel mailing list > Fof...@li... > https://lists.sourceforge.net/lists/listinfo/fofredux-devel > |
|
From: Kevin <ke...@dr...> - 2005-12-05 03:52:52
|
To help with debugging, I've started writing some unit tests using SimpleTest. I plan on putting them in a subdirectory called "tests". If anyone has any concerns reguarding this, let me know. --Kevin |
|
From: Kevin <ke...@dr...> - 2005-12-03 09:09:16
|
Andrew Turner (nilspace) wrote: >I've modified the install.php to handle either creating a new install, >or upgrading a current FoF db to meet the new db schema. This includes >such nice-ities as changing the nasty, icky "read" column to a more >amenable "flag" column. > >In the course of doing this, as well as fixing the title, some style, >etc., I've pretty much modified every file in the tree. Nothing major, >just minor tweaks here and there. Added default category. > >I think these changes should all be part of the 0.2 release, but I >don't want to check into HEAD and make people's merging a real >headache. I know Benjy & Kevin are both making system changes. >However, checking into a branch seems like more work than necessary, >as these really should be on head. > >Let me know if it's ok with you to checkin to HEAD. > Fine by me. I only have one change pending which shouldn't be difficult to merge. |
|
From: Andrew T. (nilspace) <nil...@us...> - 2005-12-03 02:08:44
|
I've modified the install.php to handle either creating a new install, or upgrading a current FoF db to meet the new db schema. This includes such nice-ities as changing the nasty, icky "read" column to a more amenable "flag" column. In the course of doing this, as well as fixing the title, some style, etc., I've pretty much modified every file in the tree. Nothing major, just minor tweaks here and there. Added default category. I think these changes should all be part of the 0.2 release, but I don't want to check into HEAD and make people's merging a real headache. I know Benjy & Kevin are both making system changes. However, checking into a branch seems like more work than necessary, as these really should be on head. Let me know if it's ok with you to checkin to HEAD. Andy -- Andrew Turner nil...@us... 42.4266N x 83.4931W http://highearthorbit.com Northville, Michigan, USA |
|
From: Benjamin S. <bs...@cr...> - 2005-12-02 19:08:39
|
On Fri, Dec 02, 2005 at 01:28:50PM -0500, Katie Bechtold wrote: > Hi! I'm a bit late to the party, but I'm a new developer on > fofredux. I've been a feedonfeeds user for a while, and like many > of you, I was impatient to see some development happen, so I > tinkered with my own installation while awaiting a fork of the > project, which I'm excited to see. I've migrated several bugs and > feature requests from feedonfeeds to fofredux. I'll leave it to Andy or someone else to say a more extended thanks, but the short version is: Awesome, thanks! > One of the "customizations" I hacked on to my installation of > feedonfeeds was to add my own cookies in magpierss's > _fetch_remote_file() function, because I wanted LiveJournal to > recognize me and give me my friends' friends-locked-entry feed items > in addition to the public ones. Oh, good hack -- I had exactly the same problem, and tried using variations on some of the published hacks for passing passwords to feeds, but they don't work for LJ (as discussed somewhere on LJ). On the one hand, though, I think that that patch should end up in magpie. On the other hand, there *has* to be an interface for getting those cookies into your subscriptions, so at least part of it has to be in the magpie clients. I've never looked at magpie's code, and I don't know how you implemetned it, but off the top of my head, there should be an API separation here that should look something like this: Magpie's methods should take cookies for feeds as an optional argument, and then FORF can store and pass those cookies to magpie. > I'd be interested in adding this sort of functionality in a less > kludge-y and more generalized way, but I'm still looking into whether > and how magpierss might support that. Er, right. See above. Magpie probably shouldn't know anything about storage of cookies, and should only take them as another parameter. --Benjy |
|
From: Katie B. <ka...@ho...> - 2005-12-02 18:31:17
|
Hi! I'm a bit late to the party, but I'm a new developer on fofredux. I've been a feedonfeeds user for a while, and like many of you, I was impatient to see some development happen, so I tinkered with my own installation while awaiting a fork of the project, which I'm excited to see. I've migrated several bugs and feature requests from feedonfeeds to fofredux. I understand there's going to be a freeze-test-release cycle coming up for fofredux. I hope someone's working on that issue with the views being messed up until the user adds at least one category; that one had me stymied when I first checked out fofredux. One of the "customizations" I hacked on to my installation of feedonfeeds was to add my own cookies in magpierss's _fetch_remote_file() function, because I wanted LiveJournal to recognize me and give me my friends' friends-locked-entry feed items in addition to the public ones. I'd be interested in adding this sort of functionality in a less kludge-y and more generalized way, but I'm still looking into whether and how magpierss might support that. By way of introducing myself, I write and test embedded spacecraft flight software by day. I've been working on New Horizons, a spacecraft bound for Pluto/Charon, for the past few years, and I'm really excited for its launch next month. By night, I help out with LinuxChix, read, practice yoga, study Japanese, and spend far too much time using feedonfeeds to check up on the 90+ feeds I've subscribed to. :) Lately, I've been getting into amateur radio, particularly amateur radio satellite communication. -- Katie Bechtold http://hoteldetective.org/ ke...@us... |
|
From: Kevin <ke...@dr...> - 2005-12-02 17:53:51
|
> I had this happen with FoF, but haven't checked FoFr. I had thought it > was due to the original feed perhaps accidentally posting it twice. > I'm also not sure how magpie determines if a post gets a new postdate > (for example, the author edits the post), how it knows it's not a new > article, but an update. Should it delete the old one, keep it, update > it? And what about any tags, read/unread, etc? The way the current code reads. Before an item is inserted into the database, it looks for an existing entry for the link. If it was not found, then it is inserted. The problem is, I am seeing 2 item rows with identical link values. This should not happen. I did some more refactoring of fof_update_feed in hopes to fix this. I'll let run for a day or two before comitting. > Since Benjy is still working on keyboard stuff, do we want to freeze > where we are and begin testing/bug fixing? > > Andy > > On 12/1/05, Kevin <ke...@dr...> wrote: >> >> I'm seeing a problem where sometimes a duplicate item is added to the >> database. The value for link is idential for both entries in fr_items= . >> I >> commited a changed that I thought might fix it but I had another >> occurance >> yesterday. Has anyone else seen this? I also had it happen with FoF, >> so >> It's not new bug, but it is annoying and I would like to try and final= ly >> fix it. >> >> >> -- >> Kevin >> >> >> >> ------------------------------------------------------- >> This SF.net email is sponsored by: Splunk Inc. Do you grep through log >> files >> for problems? Stop! Download the new AJAX search engine that makes >> searching your log files as easy as surfing the web. DOWNLOAD SPLUNK= ! >> http://ads.osdn.com/?ad_idv37&alloc_id=16865&opclick >> _______________________________________________ >> Fofredux-devel mailing list >> Fof...@li... >> https://lists.sourceforge.net/lists/listinfo/fofredux-devel >> > > > -- > Andrew Turner > ajt...@hi... 42.4266N x 83.4931W > http://highearthorbit.com Northville, Michigan, USA > > Photos - http://flickr.com/photos/ajturner > Travel - http://highearthorbit.com/projects/location/ > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log > files > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_idv37&alloc_id=16865&op=3Dclick > _______________________________________________ > Fofredux-devel mailing list > Fof...@li... > https://lists.sourceforge.net/lists/listinfo/fofredux-devel > --=20 Kevin |
|
From: Kevin <ke...@dr...> - 2005-12-02 17:42:24
|
That's what I thought, but that's not what current code represents. The default item sort uses timestamp. You'd think it'd be more useful to sor= t by the article pubdate / dcdate, if available. > Off the top of my head, I think one of them is supposed to be the time > at which the post was cached by FOF, and the other is supposed to be th= e > date information from the actual post. However, I don't know which one > is supposed to be which. > --=20 Kevin |
|
From: Benjamin S. <bs...@cr...> - 2005-12-02 17:02:34
|
Off the top of my head, I think one of them is supposed to be the time at which the post was cached by FOF, and the other is supposed to be the date information from the actual post. However, I don't know which one is supposed to be which. On Thu, Dec 01, 2005 at 09:00:38PM -0700, Kevin wrote: > I'm a bit confused on what the purpose and usage of the timestamp and > dcdate columns are. They seem to be used almost interchangably and I > have a feeling it shouldn't be that way.. > > Cheers, > Kevin > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log files > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click > _______________________________________________ > Fofredux-devel mailing list > Fof...@li... > https://lists.sourceforge.net/lists/listinfo/fofredux-devel |
|
From: Kevin <ke...@dr...> - 2005-12-02 03:56:30
|
I'm a bit confused on what the purpose and usage of the timestamp and
dcdate columns are. They seem to be used almost interchangably and I
have a feeling it shouldn't be that way..
Cheers,
Kevin
|
|
From: Carlos K. <ck...@gm...> - 2005-12-02 01:46:10
|
Could that problem have something to do with this bug ? http://sourceforge.net/tracker/index.php?func=3Ddetail&aid=3D1081298&group_= id=3D82800&atid=3D567249 I'll catch up with all the other messages tomorrow or at least this weekend= ! On 12/1/05, Andrew Turner <ajt...@hi...> wrote: > > I had this happen with FoF, but haven't checked FoFr. I had thought it > was due to the original feed perhaps accidentally posting it twice. > I'm also not sure how magpie determines if a post gets a new postdate > (for example, the author edits the post), how it knows it's not a new > article, but an update. Should it delete the old one, keep it, update > it? And what about any tags, read/unread, etc? > > Since Benjy is still working on keyboard stuff, do we want to freeze > where we are and begin testing/bug fixing? > > Andy > > On 12/1/05, Kevin <ke...@dr...> wrote: > > > > I'm seeing a problem where sometimes a duplicate item is added to the > > database. The value for link is idential for both entries in > fr_items. I > > commited a changed that I thought might fix it but I had another > occurance > > yesterday. Has anyone else seen this? I also had it happen with FoF, > so > > It's not new bug, but it is annoying and I would like to try and finall= y > > fix it. > > > > > > -- > > Kevin > > > > > > > > ------------------------------------------------------- > > This SF.net email is sponsored by: Splunk Inc. Do you grep through log > files > > for problems? Stop! Download the new AJAX search engine that makes > > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > > http://ads.osdn.com/?ad_idv37&alloc_id=16865&opclick > > _______________________________________________ > > Fofredux-devel mailing list > > Fof...@li... > > https://lists.sourceforge.net/lists/listinfo/fofredux-devel > > > > > -- > Andrew Turner > ajt...@hi... 42.4266N x 83.4931W > http://highearthorbit.com Northville, Michigan, USA > > Photos - http://flickr.com/photos/ajturner > Travel - http://highearthorbit.com/projects/location/ > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log > files > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_idv37&alloc_id=16865&opclick > _______________________________________________ > Fofredux-devel mailing list > Fof...@li... > https://lists.sourceforge.net/lists/listinfo/fofredux-devel > |
|
From: Benjamin S. <bs...@cr...> - 2005-12-01 22:40:47
|
On Thu, Dec 01, 2005 at 03:33:46PM -0700, Kevin wrote: > > Very good suggestions. I really like changing the link to "Today's > > Unread items". And the ajax stuff sounds intriguing, though I imagine > > you don't actually need Ajax. Or do you actually wait for a server > > response before removing the div from the DOM? (ba-dim-dim :) This > > would be very nice - as long as it degrades to non-js browsers. > > I'm playing with xajax on a side project at work. I'm impressed with how > easy it is to manipulate the DOM from the server side. So with very > little effort we could do exactly what you describe above. An event on > the client sends a request to the server marking an item as read, then the > server sends a response back to the client telling it to hide/remove the > item from the page. All without a refresh on view.php. Nice, huh? This is exactly what I already had implemented. Yes, it waited to remove the div until it got a "confirmed that this is marked as read" from the server, but it could instead be done by just building up a list without actually making the remote call (move the "read" articles down to the bottom of the page into a "read" bin or something). I have not used xajax, but here's what I have used: In my on disk hack against old FOF, I used prototype.js, which was poorly documented and I believe originated out of Rails. I would not use this again. For my last work project, I used Sajax, which I liked a lot because it made it really easy to bind ajax functions to specific php/perl/whatever functions on the server, and it generated all the ajax functions for you, and you were able to specify an arbitrary callback function. I especially liked the ability to specify the arbitrary callback function, because it gave me a lot more flexibility to do exactly what I needed to do. And making it degrade is easy, because we could just make the AJAX read functionality be hidden unless javascript is available to enable it. No matter how I do this, I 1. already have at least one version of it implemented, and 2. it's pretty easy. I'll work on it some tonight. --Benjy |
|
From: Andrew T. <ajt...@hi...> - 2005-12-01 22:38:41
|
I'm definitely a fan of Ajax in general - I just wrote an article on it for MacTech magazine. My point was just that XML isn't required, esp. in this case - though not necessarily a bad idea. I guess we could return a list of the items that were successfully marked as read in the database. We also don't have to support a server response - remove it from the DOM/view and send the mark read but don't handle any response. On 12/1/05, Kevin <ke...@dr...> wrote: > > > Very good suggestions. I really like changing the link to "Today's > > Unread items". And the ajax stuff sounds intriguing, though I imagine > > you don't actually need Ajax. Or do you actually wait for a server > > response before removing the div from the DOM? (ba-dim-dim :) This > > would be very nice - as long as it degrades to non-js browsers. > > I'm playing with xajax on a side project at work. I'm impressed with how > easy it is to manipulate the DOM from the server side. So with very > little effort we could do exactly what you describe above. An event on > the client sends a request to the server marking an item as read, then th= e > server sends a response back to the client telling it to hide/remove the > item from the page. All without a refresh on view.php. Nice, huh? > > > > -- > Kevin > > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log fi= les > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_idv37&alloc_id=16865&opclick > _______________________________________________ > Fofredux-devel mailing list > Fof...@li... > https://lists.sourceforge.net/lists/listinfo/fofredux-devel > -- Andrew Turner ajt...@hi... 42.4266N x 83.4931W http://highearthorbit.com Northville, Michigan, USA Photos - http://flickr.com/photos/ajturner Travel - http://highearthorbit.com/projects/location/ |
|
From: Kevin <ke...@dr...> - 2005-12-01 22:33:49
|
> Very good suggestions. I really like changing the link to "Today's > Unread items". And the ajax stuff sounds intriguing, though I imagine > you don't actually need Ajax. Or do you actually wait for a server > response before removing the div from the DOM? (ba-dim-dim :) This > would be very nice - as long as it degrades to non-js browsers. I'm playing with xajax on a side project at work. I'm impressed with how easy it is to manipulate the DOM from the server side. So with very little effort we could do exactly what you describe above. An event on the client sends a request to the server marking an item as read, then th= e server sends a response back to the client telling it to hide/remove the item from the page. All without a refresh on view.php. Nice, huh? --=20 Kevin |
|
From: Andrew T. <ajt...@hi...> - 2005-12-01 21:59:49
|
Very good suggestions. I really like changing the link to "Today's Unread items". And the ajax stuff sounds intriguing, though I imagine you don't actually need Ajax. Or do you actually wait for a server response before removing the div from the DOM? (ba-dim-dim :) This would be very nice - as long as it degrades to non-js browsers. Good ideas on Keyboard as well. My suggestion at this point - let's roll-out v0.2 as is. We'll make a branch for the keyboard stuff and allow adventuresome types to try it out and give us feedback on what scheme they like/use and we can twiddle the branch until it feels solid & useable by many. Once that happens, we can roll it into 0.3+? Andy On 12/1/05, Benjamin Stewart <bs...@cr...> wrote: > > I had thought kbd nav would also include selecting different feeds, > > etc. I tend to use Page Up and Page Down for scrolling through pages. > > Might I suggest space and shift+space for just jumping through pages? > shift+space rules. > > > I read this page: > > http://www.quirksmode.org/viewport/compatibility.html > > quirksmode is pretty super useful. > > > Which discusses compliant means for doing scrolling methods. Combined > > with "Access Keys", it's possible to do: > > -- > > <script type=3D"text/javascript"> > > function fr_scroll(inc) > > { > > var x,y; > .... > > } > > </script> > > <a href=3D"javascript:scroll(0,0)" accesskey=3D"h">Top</a> > > <a href=3D"javascript:fr_scroll(-200)" accesskey=3D"j">Up</a> > > <a href=3D"javascript:fr_scroll(200)" accesskey=3D"k">Down</a> > > -- > > So uh, that code could probably be adapted pretty easily, since all I > need to do is build a list of items on the page and then be able to jump > to those items. > > I'll see about using this code tonight to see how hard it is. I used > the refeed code just because it was there and did what I wanted and > allowed for a quick proof of concept. > > > I think the marking is *very* good idea - though I couldn't test it on > > a large number of feeds. How does it know to jump to the next feed > > item to be read? > > Oh, it doesn't actually do that. It just jumps to the next item. > > So I kind of made an oversight here, which has to do with how I read my > news: > > When I read, I read "today's unread items" instead of "today's items." > On my personal menu, I added a link to > /view.php?what=3Dall&when=3Dtoday&what=3Dunread -- the what=3Dunread is s= uper > useful. On top of that, I have a second hack where I can *ajax* mark > something read, which causes it to disappear from the page. > > Taken together, these two things mean that I never see "read" items when > I'm reading, which means I don't have to skip read items. In the > current implementation, read items would be visible if I weren't using > what=3Dunread, and marked items are visible no matter what. > > I'll be able to make it so that read items AND mark items get skipped, > and 'j' can be "next unread / unmarked" and something else (control-j?) > is absolute next item... But implementing the "I only want to see my > unread news" by just putting the "today unread" link in the menu would > be a nice feature to toss in there. > > I'll do it in my next commit. > > > I guess it can get the X,Y page location of the next > > check-box and then scrollTo(x,y) that location. > > I'm not scrolling to the checkbox; I'm scrolling to the top of the div. > > > In the end, I'm not sure what all the ReFeeds js stuff is including, > > A lot of crap. =3D) > > > but I think it's too heavy handed for just up/down and marking stuff. > > I think we'll be better served by writing up our own JS code that's > > smaller. The above code does scroll up/down (albeit with ctrl-j, > > ctrl-k) and I would think the mark would be straight-forward? > > As mentioned, I'll follow up on optimizing this tonight. But as > implemented, I should be able to change the implementing javascript > without changing the HTML at all -- I'll just need to remove the extra > javascript includes. > > --Benjy > -- Andrew Turner ajt...@hi... 42.4266N x 83.4931W http://highearthorbit.com Northville, Michigan, USA Photos - http://flickr.com/photos/ajturner Travel - http://highearthorbit.com/projects/location/ |
|
From: Benjamin S. <bs...@cr...> - 2005-12-01 21:25:43
|
> I had thought kbd nav would also include selecting different feeds, > etc. I tend to use Page Up and Page Down for scrolling through pages. Might I suggest space and shift+space for just jumping through pages? shift+space rules. > I read this page: > http://www.quirksmode.org/viewport/compatibility.html quirksmode is pretty super useful. > Which discusses compliant means for doing scrolling methods. Combined > with "Access Keys", it's possible to do: > -- > <script type="text/javascript"> > function fr_scroll(inc) > { > var x,y; .... > } > </script> > <a href="javascript:scroll(0,0)" accesskey="h">Top</a> > <a href="javascript:fr_scroll(-200)" accesskey="j">Up</a> > <a href="javascript:fr_scroll(200)" accesskey="k">Down</a> > -- So uh, that code could probably be adapted pretty easily, since all I need to do is build a list of items on the page and then be able to jump to those items. I'll see about using this code tonight to see how hard it is. I used the refeed code just because it was there and did what I wanted and allowed for a quick proof of concept. > I think the marking is *very* good idea - though I couldn't test it on > a large number of feeds. How does it know to jump to the next feed > item to be read? Oh, it doesn't actually do that. It just jumps to the next item. So I kind of made an oversight here, which has to do with how I read my news: When I read, I read "today's unread items" instead of "today's items." On my personal menu, I added a link to /view.php?what=all&when=today&what=unread -- the what=unread is super useful. On top of that, I have a second hack where I can *ajax* mark something read, which causes it to disappear from the page. Taken together, these two things mean that I never see "read" items when I'm reading, which means I don't have to skip read items. In the current implementation, read items would be visible if I weren't using what=unread, and marked items are visible no matter what. I'll be able to make it so that read items AND mark items get skipped, and 'j' can be "next unread / unmarked" and something else (control-j?) is absolute next item... But implementing the "I only want to see my unread news" by just putting the "today unread" link in the menu would be a nice feature to toss in there. I'll do it in my next commit. > I guess it can get the X,Y page location of the next > check-box and then scrollTo(x,y) that location. I'm not scrolling to the checkbox; I'm scrolling to the top of the div. > In the end, I'm not sure what all the ReFeeds js stuff is including, A lot of crap. =) > but I think it's too heavy handed for just up/down and marking stuff. > I think we'll be better served by writing up our own JS code that's > smaller. The above code does scroll up/down (albeit with ctrl-j, > ctrl-k) and I would think the mark would be straight-forward? As mentioned, I'll follow up on optimizing this tonight. But as implemented, I should be able to change the implementing javascript without changing the HTML at all -- I'll just need to remove the extra javascript includes. --Benjy |
5 messages has been excluded from this view by a project administrator.
