Menu
▾
▴
fofredux-devel
|
From: Evan R. <eva...@gm...> - 2006-03-16 23:31:04
|
Hi Guys, my first time jumping in here, so allow me to first say thanks for picking up this great project again. (if you guys are still active...saw that the dev mails were only up til late Feb at sf.net). In the course of doing a complete overhaul of my server this week, it was time to get a fresher feeds reader, and naturally i went with something i knew, as i had been using fof for well over a year now. After installing fofredux, and importing my feeds and their corresponding items (i don't delete anything), i noticed that updates on some of the feed= s were losing their dates. They were in fact defaulting to 01-01-1970, and i= t was all atom (see anything at blogger.com) and RSS1.0 (e.g. slashdot), whic= h is actually the Atom format. RSS 2.0 feeds are fine. Digging through the code, i realized it's not entirely our problem, as we call the magpie parsing function parse_w3cdtf for those two types of feeds, which is not functioning as desired. In fact, these dates are not w3c standard: w3c contains no : in the timezone, whereas Atom time does. The fix for this problem...at least so far on my side...is to simply replac= e this function with the native PHP strtotime() twice in init.php. I've tested on PHP 5.0.5 and 5.1.2 with success. Don't have any php 4.X.Xrunning here. I guess the real fix would be to repair the magpie function...but I'm not sure it's necessary when php provides native support. Hope that might help anyone else out there with a similar problem...and let me know if i've destroyed something in the process. Regards, Evan |
|
From: Kevin <ke...@dr...> - 2006-03-17 04:21:34
|
Evan Roth wrote: > Hi Guys, > > my first time jumping in here, so allow me to first say thanks for > picking up this great project again. (if you guys are still > active...saw that the dev mails were only up til late Feb at sf.net > <http://sf.net>). > Welcome Evan. The list has been a bit slow over the last few weeks. I expect things will pick up after the 0.3 release. (hint hint: Andrew) > In the course of doing a complete overhaul of my server this week, it > was time to get a fresher feeds reader, and naturally i went with > something i knew, as i had been using fof for well over a year now. > > After installing fofredux, and importing my feeds and their > corresponding items (i don't delete anything), i noticed that updates > on some of the feeds were losing their dates. They were in fact > defaulting to 01-01-1970, and it was all atom (see anything at > blogger.com <http://blogger.com>) and RSS1.0 (e.g. slashdot), which is > actually the Atom format. RSS 2.0 feeds are fine. > What version of fofredux are you using? I subscribe to slashdot also and don't see this behavior with a recent CVS snapshot. > Digging through the code, i realized it's not entirely our problem, as > we call the magpie parsing function parse_w3cdtf for those two types > of feeds, which is not functioning as desired. In fact, these dates > are not w3c standard: w3c contains no : in the timezone, whereas Atom > time does. > > The fix for this problem...at least so far on my side...is to simply > replace this function with the native PHP strtotime() twice in init.php. > I've tested on PHP 5.0.5 and 5.1.2 with success. Don't have any php > 4.X.X running here. > > I guess the real fix would be to repair the magpie function...but I'm > not sure it's necessary when php provides native support. > > Hope that might help anyone else out there with a similar > problem...and let me know if i've destroyed something in the process. I'm fine with converting to a php builtin function as long as it behaves the same. According to the php site, strtotime() has been around since version 3. There has to be some reason why magpierss includes it's own implementation. --Kevin |
|
From: Andrew T. <ajt...@hi...> - 2006-03-17 05:12:37
|
On 3/16/06, Kevin <ke...@dr...> wrote: > Evan Roth wrote: > > Hi Guys, and gal(s) > > > > > Welcome Evan. The list has been a bit slow over the last few weeks. I > expect things will pick up after the 0.3 release. (hint hint: Andrew) Yes yes, welcome - we always welcome new users/supporters. Oh - yeah, I haven't pushed that - my bad. I will be out of town this weekend, but I think the last thing I had done was thrown out there the question to all devs if they felt good about an 0.3 release. I think we should go ahead with it, but not necessarily push it as a big release, since we know we're making severe (and grand) UI changes and some DB changes (specifically search) I will tag the CVS (or Subversion, you said they're linked, right) as Release Candidate 0.3, we'll do testing this weekend and early next week. Plan on release by the end of next week. > > After installing fofredux, and importing my feeds and their > > corresponding items (i don't delete anything), i noticed that updates > > on some of the feeds were losing their dates. They were in fact > > defaulting to 01-01-1970, and it was all atom (see anything at > > blogger.com <http://blogger.com>) and RSS1.0 (e.g. slashdot), which is > > actually the Atom format. RSS 2.0 feeds are fine. > > > What version of fofredux are you using? I subscribe to slashdot also > and don't see this behavior with a recent CVS snapshot. > > > Digging through the code, i realized it's not entirely our problem, as > > we call the magpie parsing function parse_w3cdtf for those two types > > of feeds, which is not functioning as desired. In fact, these dates > > are not w3c standard: w3c contains no : in the timezone, whereas Atom > > time does. > > > > The fix for this problem...at least so far on my side...is to simply > > replace this function with the native PHP strtotime() twice in init.php= . > > I've tested on PHP 5.0.5 and 5.1.2 with success. Don't have any php > > 4.X.X running here. > > > > I guess the real fix would be to repair the magpie function...but I'm > > not sure it's necessary when php provides native support. > > > > Hope that might help anyone else out there with a similar > > problem...and let me know if i've destroyed something in the process. > I'm fine with converting to a php builtin function as long as it behaves > the same. According to the php site, strtotime() has been around since > version 3. There has to be some reason why magpierss includes it's own > implementation. Perhaps they didn't realize it was available? But that would be suprising. We should submit this fix to our respository, and include it in testing. Andrew -- Andrew Turner ajt...@hi... 42.4266N x 83.4931W http://highearthorbit.com Northville, Michigan, USA |
|
From: Evan R. <eva...@gm...> - 2006-03-17 10:05:30
|
I'm still on 0.2. I can gladly install a cvs snapshot to mess around
with...if i could find one. there is no package in the downloads, and i'm
not sure what to grab out of cvs.
As I don't hack my install too much...in order to be able to easily update
new changes...there are some small things I've changed, which you might be
interested in incorporating upstream.
1. Enable gzip in PHP. When viewing a list of ~80 feed-items, the
downloaded page was ca. 350k. With gzip it turns out to be about 110k.
Performance increase all around. All you need to do is add
"ob_start("ob_gzhandler");" to the top of init.php, and it's done. gzip
handling is supported by all modern browsers...i've used it with IE6 and
various firefox versions.
2. In the database, since your release - i didn't have it with steve's fof
- whenever an item is marked read, we change the cache time on update to
current timestamp. I yanked this effect out of the DB so that cache time
remains simply cache time. I'd suggest adding an extra field for
read_time. I'd find it useful myself...but haven't gotten around to doing
anything yet.
3. My 3rd change is not so simple. As I tend to pile up items
occasionally, i have to force fofr to remain in paged mode. If I don't, an=
d
try to load a large page, PHP blows up as it goes over it's allotted 8MB of
RAM. I won't go into details, since it was a messy change...but i'd
recommend putting some kind of $default_view into the config, so the user
can choose what he wants to see.
Anyway, enough with that. point me in the right direction to the cvs
snapshot to download, and i'd be happy test/debug the new version...nothing
better to do this week with no work - "=FCberstunden" as they say here.
Regards,
Evan
On 3/17/06, Andrew Turner <ajt...@hi...> wrote:
>
> On 3/16/06, Kevin <ke...@dr...> wrote:
> > Evan Roth wrote:
> > > Hi Guys,
> and gal(s)
>
> > >
> > >
> > Welcome Evan. The list has been a bit slow over the last few weeks. I
> > expect things will pick up after the 0.3 release. (hint hint: Andrew)
>
> Yes yes, welcome - we always welcome new users/supporters.
>
> Oh - yeah, I haven't pushed that - my bad. I will be out of town this
> weekend, but I think the last thing I had done was thrown out there
> the question to all devs if they felt good about an 0.3 release. I
> think we should go ahead with it, but not necessarily push it as a big
> release, since we know we're making severe (and grand) UI changes and
> some DB changes (specifically search)
>
> I will tag the CVS (or Subversion, you said they're linked, right) as
> Release Candidate 0.3, we'll do testing this weekend and early next
> week. Plan on release by the end of next week.
>
> > > After installing fofredux, and importing my feeds and their
> > > corresponding items (i don't delete anything), i noticed that updates
> > > on some of the feeds were losing their dates. They were in fact
> > > defaulting to 01-01-1970, and it was all atom (see anything at
> > > blogger.com <http://blogger.com>) and RSS1.0 (e.g. slashdot), which i=
s
> > > actually the Atom format. RSS 2.0 feeds are fine.
> > >
> > What version of fofredux are you using? I subscribe to slashdot also
> > and don't see this behavior with a recent CVS snapshot.
> >
> > > Digging through the code, i realized it's not entirely our problem, a=
s
> > > we call the magpie parsing function parse_w3cdtf for those two types
> > > of feeds, which is not functioning as desired. In fact, these dates
> > > are not w3c standard: w3c contains no : in the timezone, whereas Ato=
m
> > > time does.
> > >
> > > The fix for this problem...at least so far on my side...is to simply
> > > replace this function with the native PHP strtotime() twice in
> init.php.
> > > I've tested on PHP 5.0.5 and 5.1.2 with success. Don't have any php
> > > 4.X.X running here.
> > >
> > > I guess the real fix would be to repair the magpie function...but I'm
> > > not sure it's necessary when php provides native support.
> > >
> > > Hope that might help anyone else out there with a similar
> > > problem...and let me know if i've destroyed something in the process.
> > I'm fine with converting to a php builtin function as long as it behave=
s
> > the same. According to the php site, strtotime() has been around sinc=
e
> > version 3. There has to be some reason why magpierss includes it's own
> > implementation.
>
> Perhaps they didn't realize it was available? But that would be suprising=
.
> We should submit this fix to our respository, and include it in testing.
>
> Andrew
>
> --
> Andrew Turner
> ajt...@hi... 42.4266N x 83.4931W
> http://highearthorbit.com Northville, Michigan, USA
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by xPML, a groundbreaking scripting
> language
> that extends applications into web and mobile media. Attend the live
> webcast
> and join the prime developer group breaking into this new coding
> territory!
> http://sel.as-us.falkag.net/sel?cmdlnk&kid=110944&bid$1720&dat=121642
> _______________________________________________
> Fofredux-devel mailing list
> Fof...@li...
> https://lists.sourceforge.net/lists/listinfo/fofredux-devel
>
|
|
From: Andrew T. <ajt...@hi...> - 2006-03-17 12:33:18
|
On 3/17/06, Evan Roth <eva...@gm...> wrote:
> I'm still on 0.2. I can gladly install a cvs snapshot to mess around
> with...if i could find one. there is no package in the downloads, and i'=
m
> not sure what to grab out of cvs.
Kevin has created an automatic tarball generation, but for some reason
I can't find the link - Kevin?
We probably want to put one up for the RC-0_3
> All you need to do is add
> "ob_start("ob_gzhandler");" to the top of init.php, and it's done. gzip
> handling is supported by all modern browsers...i've used it with IE6 and
> various firefox versions.
Cool, we'll try this.
> 2. In the database, since your release - i didn't have it with steve's f=
of
> - whenever an item is marked read, we change the cache time on update to
> current timestamp. I yanked this effect out of the DB so that cache time
> remains simply cache time. I'd suggest adding an extra field for read_ti=
me.
> I'd find it useful myself...but haven't gotten around to doing anything
> yet.
Good point - we should investigate this. Can you file a bug? (it keeps
us more honest ;) and allows us to also know what major features were
added in which versions - same with the gzip as a feature request)
> 3. My 3rd change is not so simple. As I tend to pile up items
> occasionally, i have to force fofr to remain in paged mode. If I don't, =
and
> try to load a large page, PHP blows up as it goes over it's allotted 8MB =
of
> RAM. I won't go into details, since it was a messy change...but i'd
> recommend putting some kind of $default_view into the config, so the user
> can choose what he wants to see.
This is fixed in the newest version. I too had the problem and
suprised more people didn't in the original FoF.
> Anyway, enough with that. point me in the right direction to the cvs
> snapshot to download, and i'd be happy test/debug the new version...nothi=
ng
> better to do this week with no work - "=FCberstunden" as they say here.
>
Wo wohnst du?
--
Andrew Turner
ajt...@hi... 42.4266N x 83.4931W
http://highearthorbit.com Northville, Michigan, USA
|
|
From: Evan R. <eva...@gm...> - 2006-03-17 12:44:14
|
> > Kevin has created an automatic tarball generation, but for some reason http://fofredux.sourceforge.net/snapshots/ > Good point - we should investigate this. Can you file a bug? (it keeps > us more honest ;) and allows us to also know what major features were > added in which versions - same with the gzip as a feature request) Can do. Side note...how do i get access to the CVS (or SVN) itself. When = i try to connect with Eclipse, i'm getting a "no such user" error, even thoug= h i've got a valid account at sourceforge. Would be happy to change this stuff myself. > 3. My 3rd change is not so simple. As I tend to pile up items > > occasionally, i have to force fofr to remain in paged mode. If I don't= , > and > This is fixed in the newest version. I too had the problem and > suprised more people didn't in the original FoF. Ok, just installed it...so i'll try an import of the data and give it a test. Wo wohnst du? Braunschweig...but i wish i was still in chicago...blame the wife. |
×
Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.
Thanks for helping keep SourceForge clean.
X