Hi!
Is anyone here writing facebook apps on termite? or even scheme in general? If so, can you point me at what scheme <-> facebook library you're using? (If you converted over the php one, and are willing to donate the code, even better!)
Thanks!
Afficher les réponses par date
ah, alternative, and potentially 'better' idea --- anyone here got gambit (or scheme in general) to talk to php in apache / lighttpd? i don't mind the slight inefficiency if it puts alot of libraries at my finger tips
On Sun, Mar 8, 2009 at 4:14 PM, lowly coder lowlycoder@huoyanjinjing.comwrote:
Hi!
Is anyone here writing facebook apps on termite? or even scheme in general? If so, can you point me at what scheme <-> facebook library you're using? (If you converted over the php one, and are willing to donate the code, even better!)
Thanks!
Well it kind of depends on your app. For what I've done in the past, C talking to a php app, I did by keeping both completely separate and have them use the same data (both query same mysql db). The other time I just had C daemon and had php connect to it using a socket.
Depending on your project you might try something similar, Pavel
On Sun, Mar 8, 2009 at 4:23 PM, lowly coder lowlycoder@huoyanjinjing.com wrote:
ah, alternative, and potentially 'better' idea --- anyone here got gambit (or scheme in general) to talk to php in apache / lighttpd? i don't mind the slight inefficiency if it puts alot of libraries at my finger tips
On Sun, Mar 8, 2009 at 4:14 PM, lowly coder lowlycoder@huoyanjinjing.com wrote:
Hi!
Is anyone here writing facebook apps on termite? or even scheme in general? If so, can you point me at what scheme <-> facebook library you're using? (If you converted over the php one, and are willing to donate the code, even better!)
Thanks!
Gambit-list mailing list Gambit-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
Interesting.
So for neither of these, did you use FFI -- you pretty much said "okay, what communication mechanisms does unix give me, and I'll use those" --> thus pipe/sockets & writing to mysql (which is almost like shared memory).
Can you say some more about your experiences with this approach?
On Sun, Mar 8, 2009 at 8:20 PM, Pavel Dudrenov dudrenov@gmail.com wrote:
Well it kind of depends on your app. For what I've done in the past, C talking to a php app, I did by keeping both completely separate and have them use the same data (both query same mysql db). The other time I just had C daemon and had php connect to it using a socket.
Depending on your project you might try something similar, Pavel
On Sun, Mar 8, 2009 at 4:23 PM, lowly coder lowlycoder@huoyanjinjing.com wrote:
ah, alternative, and potentially 'better' idea --- anyone here got gambit (or scheme in general) to talk to php in apache / lighttpd? i don't mind
the
slight inefficiency if it puts alot of libraries at my finger tips
On Sun, Mar 8, 2009 at 4:14 PM, lowly coder <
lowlycoder@huoyanjinjing.com>
wrote:
Hi!
Is anyone here writing facebook apps on termite? or even scheme in general? If so, can you point me at what scheme <-> facebook library
you're
using? (If you converted over the php one, and are willing to donate the code, even better!)
Thanks!
Gambit-list mailing list Gambit-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
OK, Not much more to say there, but anyway. All you have is some standardized intermediate data representation that you can use to communicate between different systems. It's like xmlhttprequests in js with json or xml returned. Same principle.
For the mysql stuff. I had to make an IM notification for a wiki and forum systems for a client. They wanted not only email notifications, but IM notifications too. So I used libyahoo and such to make a little daemon that reads from db for pending messages, mark them as read, and send them. So whenever new post replies happen, or a wiki page was edited, they would be written to a table and the im notifier would just read from it periodically. It could be done other ways, but that was comfortable coz' the team I worked with. They were not really comfortable with any other methods. It worked rather well in the end actually. It was an internal tool, so I did not even have to wory about IM providers blocking the bot, since it had small traffic.
Second think I did was for a kind of a big site. We needed fast search based on different tags/keywords. So we made a little C server/daemon that on boot up would read up the whole DB, initialize bunch of arrays on for each keyword we wanted to search by, and sort the data for each array based on keywords. After that you would just query it using a socket and it would return you just mysql IDs, which you use to get the actual data. It proved much more efficient then doing it all in mysql. Ofcaurse whenever you would create new record in DB you would update the daemon too.
Any method will have it's pluses and minuses use whatever you feel most comfortable, whatever you understand best.
Pavel P.S. I might still have the source for the later one, but it's such a simple thing that I don't think it's even worth pasting here. I might have the IM thingy code somewhere too.
On Mon, Mar 9, 2009 at 12:29 PM, lowly coder lowlycoder@huoyanjinjing.com wrote:
Interesting.
So for neither of these, did you use FFI -- you pretty much said "okay, what communication mechanisms does unix give me, and I'll use those" --> thus pipe/sockets & writing to mysql (which is almost like shared memory).
Can you say some more about your experiences with this approach?
On Sun, Mar 8, 2009 at 8:20 PM, Pavel Dudrenov dudrenov@gmail.com wrote:
Well it kind of depends on your app. For what I've done in the past, C talking to a php app, I did by keeping both completely separate and have them use the same data (both query same mysql db). The other time I just had C daemon and had php connect to it using a socket.
Depending on your project you might try something similar, Pavel
On Sun, Mar 8, 2009 at 4:23 PM, lowly coder lowlycoder@huoyanjinjing.com wrote:
ah, alternative, and potentially 'better' idea --- anyone here got gambit (or scheme in general) to talk to php in apache / lighttpd? i don't mind the slight inefficiency if it puts alot of libraries at my finger tips
On Sun, Mar 8, 2009 at 4:14 PM, lowly coder lowlycoder@huoyanjinjing.com wrote:
Hi!
Is anyone here writing facebook apps on termite? or even scheme in general? If so, can you point me at what scheme <-> facebook library you're using? (If you converted over the php one, and are willing to donate the code, even better!)
Thanks!
Gambit-list mailing list Gambit-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
I want to apologize for my previous email. It's full with spelling errors and half of it don't make sense. I blame it on my lack of sleep/coffee and my inferior knowledge of English.
Pavel
On Mon, Mar 9, 2009 at 12:51 PM, Pavel Dudrenov dudrenov@gmail.com wrote:
OK, Not much more to say there, but anyway. All you have is some standardized intermediate data representation that you can use to communicate between different systems. It's like xmlhttprequests in js with json or xml returned. Same principle.
For the mysql stuff. I had to make an IM notification for a wiki and forum systems for a client. They wanted not only email notifications, but IM notifications too. So I used libyahoo and such to make a little daemon that reads from db for pending messages, mark them as read, and send them. So whenever new post replies happen, or a wiki page was edited, they would be written to a table and the im notifier would just read from it periodically. It could be done other ways, but that was comfortable coz' the team I worked with. They were not really comfortable with any other methods. It worked rather well in the end actually. It was an internal tool, so I did not even have to wory about IM providers blocking the bot, since it had small traffic.
Second think I did was for a kind of a big site. We needed fast search based on different tags/keywords. So we made a little C server/daemon that on boot up would read up the whole DB, initialize bunch of arrays on for each keyword we wanted to search by, and sort the data for each array based on keywords. After that you would just query it using a socket and it would return you just mysql IDs, which you use to get the actual data. It proved much more efficient then doing it all in mysql. Ofcaurse whenever you would create new record in DB you would update the daemon too.
Any method will have it's pluses and minuses use whatever you feel most comfortable, whatever you understand best.
Pavel P.S. I might still have the source for the later one, but it's such a simple thing that I don't think it's even worth pasting here. I might have the IM thingy code somewhere too.
On Mon, Mar 9, 2009 at 12:29 PM, lowly coder lowlycoder@huoyanjinjing.com wrote:
Interesting.
So for neither of these, did you use FFI -- you pretty much said "okay, what communication mechanisms does unix give me, and I'll use those" --> thus pipe/sockets & writing to mysql (which is almost like shared memory).
Can you say some more about your experiences with this approach?
On Sun, Mar 8, 2009 at 8:20 PM, Pavel Dudrenov dudrenov@gmail.com wrote:
Well it kind of depends on your app. For what I've done in the past, C talking to a php app, I did by keeping both completely separate and have them use the same data (both query same mysql db). The other time I just had C daemon and had php connect to it using a socket.
Depending on your project you might try something similar, Pavel
On Sun, Mar 8, 2009 at 4:23 PM, lowly coder lowlycoder@huoyanjinjing.com wrote:
ah, alternative, and potentially 'better' idea --- anyone here got gambit (or scheme in general) to talk to php in apache / lighttpd? i don't mind the slight inefficiency if it puts alot of libraries at my finger tips
On Sun, Mar 8, 2009 at 4:14 PM, lowly coder lowlycoder@huoyanjinjing.com wrote:
Hi!
Is anyone here writing facebook apps on termite? or even scheme in general? If so, can you point me at what scheme <-> facebook library you're using? (If you converted over the php one, and are willing to donate the code, even better!)
Thanks!
Gambit-list mailing list Gambit-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
Thanks! The english was fine, plus the content was much more than what I would have achieved wasting time reading slashdot/reddit/your-favorite-news-site.
On Mon, Mar 9, 2009 at 12:54 PM, Pavel Dudrenov dudrenov@gmail.com wrote:
I want to apologize for my previous email. It's full with spelling errors and half of it don't make sense. I blame it on my lack of sleep/coffee and my inferior knowledge of English.
Pavel
On Mon, Mar 9, 2009 at 12:51 PM, Pavel Dudrenov dudrenov@gmail.com wrote:
OK, Not much more to say there, but anyway. All you have is some standardized intermediate data representation that you can use to communicate between different systems. It's like xmlhttprequests in js with json or xml returned. Same principle.
For the mysql stuff. I had to make an IM notification for a wiki and forum systems for a client. They wanted not only email notifications, but IM notifications too. So I used libyahoo and such to make a little daemon that reads from db for pending messages, mark them as read, and send them. So whenever new post replies happen, or a wiki page was edited, they would be written to a table and the im notifier would just read from it periodically. It could be done other ways, but that was comfortable coz' the team I worked with. They were not really comfortable with any other methods. It worked rather well in the end actually. It was an internal tool, so I did not even have to wory about IM providers blocking the bot, since it had small traffic.
Second think I did was for a kind of a big site. We needed fast search based on different tags/keywords. So we made a little C server/daemon that on boot up would read up the whole DB, initialize bunch of arrays on for each keyword we wanted to search by, and sort the data for each array based on keywords. After that you would just query it using a socket and it would return you just mysql IDs, which you use to get the actual data. It proved much more efficient then doing it all in mysql. Ofcaurse whenever you would create new record in DB you would update the daemon too.
Any method will have it's pluses and minuses use whatever you feel most comfortable, whatever you understand best.
Pavel P.S. I might still have the source for the later one, but it's such a simple thing that I don't think it's even worth pasting here. I might have the IM thingy code somewhere too.
On Mon, Mar 9, 2009 at 12:29 PM, lowly coder lowlycoder@huoyanjinjing.com wrote:
Interesting.
So for neither of these, did you use FFI -- you pretty much said "okay,
what
communication mechanisms does unix give me, and I'll use those" --> thus pipe/sockets & writing to mysql (which is almost like shared memory).
Can you say some more about your experiences with this approach?
On Sun, Mar 8, 2009 at 8:20 PM, Pavel Dudrenov dudrenov@gmail.com
wrote:
Well it kind of depends on your app. For what I've done in the past, C talking to a php app, I did by keeping both completely separate and have them use the same data (both query same mysql db). The other time I just had C daemon and had php connect to it using a socket.
Depending on your project you might try something similar, Pavel
On Sun, Mar 8, 2009 at 4:23 PM, lowly coder lowlycoder@huoyanjinjing.com wrote:
ah, alternative, and potentially 'better' idea --- anyone here got gambit (or scheme in general) to talk to php in apache / lighttpd? i don't
mind
the slight inefficiency if it puts alot of libraries at my finger tips
On Sun, Mar 8, 2009 at 4:14 PM, lowly coder lowlycoder@huoyanjinjing.com wrote:
Hi!
Is anyone here writing facebook apps on termite? or even scheme in general? If so, can you point me at what scheme <-> facebook library you're using? (If you converted over the php one, and are willing to donate the code, even better!)
Thanks!
Gambit-list mailing list Gambit-list@iro.umontreal.ca https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
-
lowly coder -
Pavel Dudrenov