gmane.comp.lang.pike.user

C API: accessing to the elements of an array

2008-09-30T11:45:02

Yep, but possibly with a typecheck first. (I still don't know if the method is internal, then it might just as well *return* the array.)

C API: accessing to the elements of an array

2008-09-30T09:05:03

Accessing the elements is not that difficult either. This should give the array_index item (svalue) from the array at stack_index position on the stack. ITEM(Pike_sp[-stack_index])[array_index]

C API: accessing to the elements of an array

2008-09-30T08:25:03

Do you need to access the elements, or do you need to just index them out and leave them on the stack? The last is easier, you simply make sure you have the array on top of the stack, push an integer and do f_index(2), which gives the equivalent of arr[i]. But I'm not sure what kind of problem you are trying to solve. It looks like you're calling matrixX(_lu) internally, not via the method calls, and you wouldn't need the args counter then...

C API: accessing to the elements of an array

Felip Alàez Nadal — 2008-09-30T07:08:33

Hello: I'm adding a solveLinearEquations method to Math.Matrix, writting new code in matrix_code.h. There's something I don't know how to do. Well, from inside solveLinearEquations I call the method which does the LU decompositon, by typing static void solveLinearEquations ( INT32 args ) { //SOme code to access to the parameters matrixX(_lu) (0 ); // the lu decomposition doesn't need any parameters //Here I need something to access to the elements of the array which matrixX(_lu) returns into pike. //More code which solve the equations system } The matrixX(_lu) method pushes an array with for elements into the stack calling something similar to push_object( L); push_object ( U); push_object ( Permutations ); push_int ( Sign ); f_aggregate (4); /* Make an array of the 4 top elements on the stack. */ Now I want to access those elements from solveLinearEquations in order to use the L and U matrices to solve a linear equations system. How can I "unaggregate" those elements, or pop the complete array

Re: Filesystem.System()->mkdir()

2008-09-23T20:40:04

You should probably wonder "why mkdir() and Filesystem...mkdir()". Filesystem solves other problems, for instance when you have a tar file or a remote filesystem. It was created as a base of making systems-that-contains-files with a common interface. mkdir() and Stdio.* exist because it's what you *usually* want to do without going circles before you get there. (mkdirhier() solves a superproblem of mkdir().)

Re: Filesystem.System()->mkdir()

Bertrand LUPART — 2008-09-23T15:42:50

http://pike.ida.liu.se/generated/manual/modref/ex/predef_3A_3A/Filesyste m/Base/mkdir.html Out of curiosity, why Filesystem.Base()->mkdir() and Stdio.mkdirhier()?

Re: Filesystem.System()->mkdir()

Martin Bähr — 2008-09-23T15:19:44

what is missing is links to functions of parent classes in the docs: http://pike.ida.liu.se/generated/manual/modref/ex/predef_3A_3A/Filesystem/Base/mkdir.html greetings, martin.

Filesystem.System()->mkdir()

Arne Goedeke — 2008-09-23T14:15:10

not there. any reason for that or is it just missing?

Re: Public.Tools.RRDtool

Lance Dillon — 2008-09-22T09:35:20

I've got some somewhere... I'll try to dig some up. It isn't very good, however. The thing about rrdtool is that they don't have much of an api. The api is nothing more than passing arguments to the function, the same way as you pass arguments to the command line application. It does all the parsing itself. Everything else was the same, including parts that dumped to or to a file, still did so. You couldn't request to get the information back as a string buffer or something. You had to dump to the file (or stdout), then parse that yourself, rather than getting something more in a machine format (like arrays or something). It turned out to be almost easier to skip it and just spawn processes and capture the output. ----- Original Message ---- From: James Harton helicopter.geek.nz> To: pike< at >roxen.com Sent: Sunday, September 21, 2008 10:55:34 PM Subject: Public.Tools.RRDtool Hi there. There are no module docs for this module on modules.gotpike.org - does anyone have any code they can show

Re: Public.Network.Pcap

Bertrand LUPART — 2008-09-22T07:24:41

Hello, Looks like your are using the package pike7.6-7.6.93-1 from Debian etch which is known to be buggy. Maybe try the ones from Henrik. Add the following to your /etc/sources.list: ----8<----8<----8<----8<---- deb http://debian.han.pp.se/debian/ stable main ---->8---->8---->8---->8---- and install 7.6.112 As a side note, all the Debian etch packages installing code from modules.gotpike.org (pike7.6-public*) are compiled for 7.6.93. They aren't available from debian.han.pp.se for 7.6.112 and you'll have to rebuild them yourself using monger: http://modules.gotpike.org/docs/UsingMonger.html

Public.Tools.RRDtool

James Harton — 2008-09-22T02:55:34

Hi there. There are no module docs for this module on modules.gotpike.org - does anyone have any code they can show me? Cheers. James.

Re: Public.Network.Pcap

James Harton — 2008-09-21T22:43:10

Hi guys. I get the following error from Thread.Queue when tracing fast transfers - the Queue seems very bursty even with 10 workers, and seems to build up before unblocking the read() calls. It's only using about 0.5% CPU and hardly any RAM, so it's not that the machine can't handle it: Svalue stack overflow. (99624 of 100000 entries on stack, needed 256 more entries) /usr/lib/pike/7.6.93/lib/modules/Thread.pmod:373: Thread.Thread(0).Queue(38950)->write(mapping[5]) /usr/lib/pike/7.6.93/lib/modules/Public.pmod/Network.pmod/Pcap.pmod:6: Public.Network.Pcap.Pcap()->loop(0) The code I'm using is: int main() { queue = Thread.Queue(); threads = ({}); for (int i; i < THREADS; i++) threads += ({ Thread.thread_create(worker) }); fe = IP.FlowEngine(); object cap = Public.Network.Pcap.Pcap(); cap->set_capture_callback(queue->write); //cap->set_capture_callback(capture_cb); cap->set_capture_length(MTU); cap->open_live(IF); threads += ({ Thread.thread_create(cap->loop, 0) }); return

dwim - format not handled

2008-09-18T21:05:03

They are actually getting more common in recent geek/web2 dev circles. Surprising but pleasant.

dwim - format not handled

2008-09-18T20:45:03

That's an ISO format... I didn't know anyone *actually* used those. :)

dwim - format not handled

Jeff Ober — 2008-09-18T19:12:05

The Calendar module FAQ asked for notifications for standard formats that the dwim functions do not successfully parse. Here is a sample one, used in some RSS-type feeds (notably, Flickr/Yahoo's Media RSS format): 2008-08-08T15:13:53-08:00 Replace the T with a space and it parses correctly.

Re: Math.Matrix revisited

2008-09-18T18:30:04

There hasn't been a lot of man-hours dedicated to it lately. Feel free to pick a FIXME from CHANGES and fill out the details.

Re: Math.Matrix revisited

2008-09-18T07:35:04

Anyone working on that?

Re: Math.Matrix revisited

2008-09-17T15:55:02

Changelog. Absolutely only the changelog. Has been for weeks.

Re: Math.Matrix revisited

2008-09-17T12:55:02

What is the delay for 7.8 release right now?

Re: Math.Matrix revisited

Martin Stjernholm — 2008-09-17T12:13:53

/.../ There's a newline missing at the end of the error message that apparently makes you misinterpret it (the "Unknown program" bit comes from that the backtrace formatter couldn't find the name of the program that contains norm()). I've now fixed newlines for several error messages in matrix_code.h. Thanks. As for your math patches, I hope someone else more familiar with that area will review your contribution so that we get it in. Note however that 7.8 is in strict feature freeze, so it's unlikely that it'll get into that version.

Re: Math.Matrix revisited

Felip Alàez Nadal — 2008-09-17T11:38:03