the "welcome" item has disapeared all together, not that i mind, i don't know if it's intentional, just thought i'd state my observations
I'll post here again when a new item is added with what happened
Yes, I got rid of the welcome from the news feed. It was only put there as a simple way to get it onto the front page.
The GUID were but I've now changed them to "true" ones. Is this better?
I'll post another news item in the next few days to see if things then work better for you.
Bruce
had a look at the feed, and yes, looks perfect :)
I wasn't actually expecting permalinks to be used, just proper s, but it's great that you have, and the numbering scheme looks like it'll work properly now :)
thanks bruce
Lee, unless you post again to this thread I will assume that all is now OK with the RSS feed. Since I don't use RSS feeds, I would appreciate it if you raise a fuss in the future if you spot anything amiss.
Lee, unless you post again to this thread I will assume that all is now OK with the RSS feed. Since I don't use RSS feeds, I would appreciate it if you raise a fuss in the future if you spot anything amiss.
I'll post some feedback when the next news item is published (don't do a special one, only when you'd normally post a news item for something about the project)
indeed i will then :) thanks for spending the time fixing these relatively minor things, but it sure helps with keeping track of news at all the projects :)
could i make a request...
i know you're a busy guy, so it's not a high priority, but as you've gone thru the trouble of fixing einsteins feed yourself, and no-one seems to be listening over at SETI (i've had no feedback about the feed issues at least),
would it be possible to bring this up on the boinc_dev mailing list, and possibly submit your changes into the CVS (i don't know how easy that would be)
it's just that as more projects upgrade, this problem will affect everyone until/unless things are changed at the source, and it would be helpful to the other projects too if they didn't each have to fix their own feeds
but again, many thanks for sorting einsteins feed :)
feeds should be served as applicatin/xml (at least, if not their content-specific type such as application/rss+xml for RSS)
rather than text/xml
the difference is that application/xml means that it needs to be processed/parsed before it's meaningful, and before it should be displayed to the user (which is correct for feeds)
text/xml means that it's ok as it is (which isn't true of most cases where xml is used)
text/xml will work, but application/xml is the correct type to use
feeds should be served as applicatin/xml (at least, if not their content-specific type such as application/rss+xml for RSS)
rather than text/xml
the difference is that application/xml means that it needs to be processed/parsed before it's meaningful, and before it should be displayed to the user (which is correct for feeds)
text/xml means that it's ok as it is (which isn't true of most cases where xml is used)
text/xml will work, but application/xml is the correct type to use
if i may be so bold, i've got another suggestion which will help both yourselves and users with bandwidth and performance related to the feed (but this could be applied to other php pages too)
basically, sharpreader uses conditional GET requests, but the problem with the feed is that the "last-modified" value keeps changing, and is the current date/time when the feed is served (use a HTTP header viewer to see)
so this makes the conditional requests useless, and instead of a nice short "HTTP 304 Not Modified", a full "HTTP 200 OK" response is sent along with the full feed
to correct this behaviour, the last-modified value needs to remain constant untill such the feed is changed (new item, or an existing item is updated, or any other change really), when it is changed, it should be replaced with the date/time that the feed was modified (so clients redownload it)
also the of the feed should be the same as the last-modified value, because that keeps changing as well, take a look and refresh it after a couple of seconds
i helped chris malton of BOINC Project Status fame with getting his feed to report last-modified correctly, so you may wish to speak to him about what code he used, but i think the way he did it requires that a new field be created in the DB
to correct this behaviour, the last-modified value needs to remain constant untill such the feed is changed (new item, or an existing item is updated, or any other change really), when it is changed, it should be replaced with the date/time that the feed was modified (so clients redownload it)
The fundamental problem is that the page (I assume the relevant page is rss_main.php, right?) is being generated dynamically by a .php script, so it's not static. It may be possible to fix this using the BOINC page caching mechanism -- I'll check.
The fundamental problem is that the page (I assume the relevant page is rss_main.php, right?) is being generated dynamically by a .php script, so it's not static. It may be possible to fix this using the BOINC page caching mechanism -- I'll check.
yup, i was refering to rss_main.php
the content that it generates is the same until a new item is posted, so if the right php code is used (although i don't know a whole lot about PHP) the last-modified header being sent can be the same until the fed changes, making it behave like a static file and allowing conditional GET requests to be effective
because at the moment, every time the feed is requested, it's sent with a full "HTTP 200 OK" rather than just a header containing "HTTP 304 Not Modified"
i'm sure you could check in the server logs for how much the feed is used, and if it's a significant number then the bandwidth savings to projects could be quite a lot, not to mention server load, becuase as you know, each time it's requested it has to go through the PHP engine to be generated
i would try to do this myself, but i'm having trouble accessing the CVS, so untill i can get the source files i can't do much
My RSS Reader (add-in to MS Outlook) downloads the “news article page�. What happens when I’ve now subscribed to the news is that I have a load of articles (9 actually), each with exactly the same web page (the one with all articles).
More normal with RSS stuff to put each article of a different page. Not a major issue but thought I would mention is if things are being “re-structured at some point.
RE: RE: RE: the
)
Lee, unless you post again to this thread I will assume that all is now OK with the RSS feed. Since I don't use RSS feeds, I would appreciate it if you raise a fuss in the future if you spot anything amiss.
Bruce
Director, Einstein@Home
RE: Lee, unless you post
)
I'll post some feedback when the next news item is published (don't do a special one, only when you'd normally post a news item for something about the project)
indeed i will then :) thanks for spending the time fixing these relatively minor things, but it sure helps with keeping track of news at all the projects :)
could i make a request...
i know you're a busy guy, so it's not a high priority, but as you've gone thru the trouble of fixing einsteins feed yourself, and no-one seems to be listening over at SETI (i've had no feedback about the feed issues at least),
would it be possible to bring this up on the boinc_dev mailing list, and possibly submit your changes into the CVS (i don't know how easy that would be)
it's just that as more projects upgrade, this problem will affect everyone until/unless things are changed at the source, and it would be helpful to the other projects too if they didn't each have to fix their own feeds
but again, many thanks for sorting einsteins feed :)
Want to search the BOINC Wiki, BOINCstats, or various BOINC forums from within firefox? Try the BOINC related Firefox Search Plugins
Oh, another thing... feeds
)
Oh, another thing...
feeds should be served as applicatin/xml (at least, if not their content-specific type such as application/rss+xml for RSS)
rather than text/xml
the difference is that application/xml means that it needs to be processed/parsed before it's meaningful, and before it should be displayed to the user (which is correct for feeds)
text/xml means that it's ok as it is (which isn't true of most cases where xml is used)
text/xml will work, but application/xml is the correct type to use
Want to search the BOINC Wiki, BOINCstats, or various BOINC forums from within firefox? Try the BOINC related Firefox Search Plugins
RE: Oh, another
)
Fixed, please confirm.
Bruce
Director, Einstein@Home
RE: RE: feeds should be
)
lovely :)
a shining example to all the other projects ;)
Want to search the BOINC Wiki, BOINCstats, or various BOINC forums from within firefox? Try the BOINC related Firefox Search Plugins
Bruce, Just to let you
)
Bruce,
Just to let you know that the recent news item appeared exactly as it should :)
Thanks for all the work on the feed, hopefully others will follow your fine example
Want to search the BOINC Wiki, BOINCstats, or various BOINC forums from within firefox? Try the BOINC related Firefox Search Plugins
if i may be so bold, i've got
)
if i may be so bold, i've got another suggestion which will help both yourselves and users with bandwidth and performance related to the feed (but this could be applied to other php pages too)
basically, sharpreader uses conditional GET requests, but the problem with the feed is that the "last-modified" value keeps changing, and is the current date/time when the feed is served (use a HTTP header viewer to see)
so this makes the conditional requests useless, and instead of a nice short "HTTP 304 Not Modified", a full "HTTP 200 OK" response is sent along with the full feed
to correct this behaviour, the last-modified value needs to remain constant untill such the feed is changed (new item, or an existing item is updated, or any other change really), when it is changed, it should be replaced with the date/time that the feed was modified (so clients redownload it)
also the of the feed should be the same as the last-modified value, because that keeps changing as well, take a look and refresh it after a couple of seconds
i helped chris malton of BOINC Project Status fame with getting his feed to report last-modified correctly, so you may wish to speak to him about what code he used, but i think the way he did it requires that a new field be created in the DB
Want to search the BOINC Wiki, BOINCstats, or various BOINC forums from within firefox? Try the BOINC related Firefox Search Plugins
RE: to correct this
)
The fundamental problem is that the page (I assume the relevant page is rss_main.php, right?) is being generated dynamically by a .php script, so it's not static. It may be possible to fix this using the BOINC page caching mechanism -- I'll check.
Cheers,
Bruce
Director, Einstein@Home
RE: The fundamental problem
)
yup, i was refering to rss_main.php
the content that it generates is the same until a new item is posted, so if the right php code is used (although i don't know a whole lot about PHP) the last-modified header being sent can be the same until the fed changes, making it behave like a static file and allowing conditional GET requests to be effective
because at the moment, every time the feed is requested, it's sent with a full "HTTP 200 OK" rather than just a header containing "HTTP 304 Not Modified"
i'm sure you could check in the server logs for how much the feed is used, and if it's a significant number then the bandwidth savings to projects could be quite a lot, not to mention server load, becuase as you know, each time it's requested it has to go through the PHP engine to be generated
i would try to do this myself, but i'm having trouble accessing the CVS, so untill i can get the source files i can't do much
Want to search the BOINC Wiki, BOINCstats, or various BOINC forums from within firefox? Try the BOINC related Firefox Search Plugins
My RSS Reader (add-in to MS
)
My RSS Reader (add-in to MS Outlook) downloads the “news article page�. What happens when I’ve now subscribed to the news is that I have a load of articles (9 actually), each with exactly the same web page (the one with all articles).
More normal with RSS stuff to put each article of a different page. Not a major issue but thought I would mention is if things are being “re-structured at some point.
Ian