Pike-devel

pike-devel@lists.lysator.liu.se

2 participants
3046 discussions

What are the consequences of adding a subtype?
by Chris Angelico 20 Oct '21

20 Oct '21

Currently functions have two subtypes: normal ones (0) and builtin functions (USHRT_MAX). What would be the implications of adding two more? I propose adding FUNCTION_GENERATOR and FUNCTION_GENSTATE to differentiate continue functions, and the state functions that they return; and also a Pike-callable Function.function_type which will return the corresponding type. Thus: continue int demo() { } Function.function_type(demo) == 1 Function.function_type(demo()) == 2 Alternatively, if this doesn't need to be a subtype, is there a way to recognize them already? I'm poking around, trying to implement this, but I think more information needs to be carried through to runtime - might be wrong though. ChrisA

2 1

Confusing results in continue function - overaggressive optimization?
by Chris Angelico 19 Oct '21

19 Oct '21

continue int fibonacci() { int a = 0, b = 1; while (1) { yield(a += b); yield(b += a); } } int main() { function fib = fibonacci(); while (1) { write("--> %d\n", fib()); sleep(0.125); } } In theory, this should yield a stream of Fibonacci numbers (1, 1, 2, 3, 5, 8, 13, etc). In practice - on Pike 8.1.14 - it yields counting numbers (1, 2, 3, 4, 5, etc). I've tried to defeat the optimization in a few ways but to no avail. The only way to get Fibonacci numbers out of it is to avoid yield altogether: #define yield(x) write("%d\n", x) Any ideas? ChrisA

1 0

Segfault from Pike 8.1 - possible type check issue?
by Chris Angelico 06 Sep '21

06 Sep '21

float thing_sum(array(string) things) { return `+(@(array(float))things) + 1.0; } Without the "+ 1.0" at the end (or some other arithmetic operation), it doesn't bomb out. GDB is pointing to src/pike_types.cmod:11185 which is just an assignment, but I suspect the issue is the add_ref above it. Unfortunately that line gets hit a LOT, so I didn't get very far with debugging. ChrisA

1 0

Bleeding-edge Pike broadly operable - code review appreciated
by Chris Angelico 25 Aug '21

25 Aug '21

I just pushed a few changes to master that get the build working (on 64-bit Debian Linux, no other platforms tested). They're mostly small changes and I hope they won't break anything. This actually gets "make doc" happy again after a good while, so it would be really awesome if I haven't just broken something unrelated :) There's another, more significant, change that happened in the past few months. It's not a simple fix. The default behaviour of Protocols.DNS.async_host_to_ip, which is used by Protocols.HTTP, changed from "only look up A records" to "look up AAAA and A records, and return the first AAAA if found, otherwise the first A". Unfortunately everything is built to assume that there's at most one IP address for that host name. The ultimate solution would be to have an asynchronous "establish connection" call, which will do an async DNS lookup, then attempt to connect to each IP in turn until one succeeds. That would require some other changes, and I'm not sure how it would interact with HTTP keep-alive, for instance. As a lesser solution, would it be reasonable to have a method like Protocols.DNS.prefer_ipv4() to invert the return order of DNS records? ChrisA

4 12

Type checking segfault in Pike 8.1
by Chris Angelico 28 Jun '21

28 Jun '21

void boom(mapping info) {m_delete(info[0]);} It ought to give a compilation error (not enough args to m_delete) but it segfaults the interpreter on compilation. Correct behaviour is restored if the argument is explicitly cast to mapping: void boom(mapping info) {m_delete((mapping)info[0]);} boom.pike:1:Too few arguments to m_delete (got 1). boom.pike:1:Expected: mixed. boom.pike:1:Function type: boom.pike:1:Got : scope(0,function(object, mixed : mixed) | function(mapping((0=mixed):(1=mixed)), 0 : 1) | function(multiset((0=mixed)), 0 : int(1bit))). Pike: Failed to compile script. The crash happens in src/pike_ctypes.cmod, get_first_arg_type, in the T_MANY case - fun_type->car is NULL. ChrisA

1 0

Re: Bleeding-edge Pike broadly operable - code review appreciated
by Chris Angelico 10 Jun '21

10 Jun '21

On Thu, Jun 10, 2021 at 11:49 AM Martin Nilsson (Coppermist) @ Pike (-) developers forum <10353(a)lyskom.lysator.liu.se> wrote: > > > > >Okay. Here's a proposal: > > > >1) Have a Protocols.DNS.prefer_ipv6() that chooses whether IPv6 is > >preferred over IPv4 for simple lookups. Calling prefer_ipv6(1) will > >result in the current behaviour, calling prefer_ipv6(0) will return to > >previous behaviour. This just changes the order of responses, nothing > >else. > > No, add new methods for new behavior. The old function continues to > return IPv4 and the new one does IPv6+IPv4. Or have some new parameter > to control preference on a per call basis. We don't want global states > in modules. Trouble with that is that it'd need to be carried through as part of the API for everything that uses it. So Protocols.HTTP.Query would need to have something to control which one it prefers, and Protocols.HTTP.get_url would need to have something to set that parameter, etc. But maybe that won't matter, if the other parts of the plan are worth doing. ChrisA

1 0

hapcanmonitor in Pike
by Stephen R. van den Berg 24 May '21

24 May '21

It's still work in progress, but in case anyone is looking for something to run devices off and is programmed in Pike. I bought some equipment from www.hapcan.com and then programmed a hapcanmonitor daemon. The system is successfully running my house. A quick rundown on the source: - hapcanmonitor + hapcanmonitor.pike (40 KB), the full hapcan core communication driver. + www/x/wildcard.pike (7 KB), the REST/JSON api to the core communication driver. - spike (55 KB), the webserver daemon. hapcanmonitor source: git://devel2.cuci.nl/hapcanmonitor SPike daemon source: git://devel2.cuci.nl/spike -- Stephen.

1 0

HTTP server regression
by Martin Nilsson (Coppermist) ＠ Pike (-) developers forum 10 May '21

10 May '21

Looks like the Protocols.HTTP.Server is broken, though I don't know how recently as it was a while ago I installed the previous Pike 8.1. It's never the first request that fails, so I'm suspecting something like keep-alive, though it could be inside the SSL state (tickets?) things are messed up. The problem appears to be writes after finished() has been called. As I said the code used to work in 8.1 and works in 8.0. Anyone got any obvious idea? Indexing the NULL value with "output_to". /usr/local/pike/8.1.13/lib/modules/Protocols.pmod/HTTP.pmod/Server.pmod/Request.pike:929: Protocols.HTTP.Server.Request("GET" "/test.html")->send_write() /usr/local/pike/8.1.13/lib/modules/SSL.pmod/File.pike:1315: SSL.File(Stdio.File("socket", "85.228.101.175 55103", 777 /* fd=14 */), SSL.ServerConnection(ready))->internal_poll() -:1: Pike.Backend(0)->`()(3600.0)

2 9

Generic async DNS calls with multiple results
by Stephen R. van den Berg 09 May '21

09 May '21

I'm writing a generic HTTP server and will probably be writing a full-fledged SMTP server as well (queueing messages in a database, automatic DKIM signage, full MX-DNS support with retries and automatic fallback/backoff for recipient servers that don't allow many recipients or many connections or many messages per second). For the HTTP server I need multiple IP addresses per domainquery (I need to bind to all interfaces), and for the SMTP server I need to query multiple TXT records (SPF validation) and need to get back preference-sorted MX records, of which I then need multiple IP addresses again to try and connect to all IP addresses. For the both servers I need to know the difference between a (possibly temporary) failure of the query and a definitive result that says the record does not exist. For this purpose I created a small generic dnsquery() function (see below). Is it something worth including in the Protocols.DNS module, or is it better left outside? The problem was/is, none of the existing convenience functions seems to allow me to do this. Protocols.DNS.async_client dnsclient = Protocols.DNS.async_client(); private void dnscb(string domain, mapping res, function(array, array:void) cb, array restargs) { array an = res->an; switch (res->rcode) { case Protocols.DNS.SERVFAIL: case Protocols.DNS.NOTIMP: case Protocols.DNS.REFUSED: case Protocols.DNS.YXDOMAIN: case Protocols.DNS.YXRRSET: case Protocols.DNS.NXRRSET: case Protocols.DNS.NOTZONE: if (!sizeof(an)) { cb(0, restargs); return; } } sort(an->preference, an); cb(an->aaaa + an->a + an->mx + an->txt - ({ 0 }), restargs); } private mapping dnstypetonum = ([ "A": Protocols.DNS.T_A, "MX": Protocols.DNS.T_MX, "TXT": Protocols.DNS.T_TXT, "AAAA": Protocols.DNS.T_AAAA, ]); private void collect2aaaa(array|zero results, array restargs) { array|zero prevresults = restargs[1]; if (results) { if (prevresults) results -= prevresults; else prevresults = ({}); prevresults += results; } if (--restargs[0]) restargs[1] = prevresults; else restargs[2](prevresults, restargs[3..]); } //! Asynchronous DNS query with multiple results and a distinction //! between failure and empty results. //! //! @param type //! DNS query type (case sensitive). Currently supported: A, MX, TXT, AAAA. //! Querying for AAAA gives both IPv4 and IPv6 results. //! //! @param domain //! The domain name we are querying. //! //! @param result_cb //! The callback function that receives the result of the DNS query. //! It is called as follows: //! @expr{void result_cb(array(string)|zero results,array restargs);@} //! If the request fails it will return @expr{zero@} for @expr{results@}. //! @expr{restargs@} is passed unaltered to the callback function. //! //! @note //! There is a notable difference between @expr{results@} equal //! to @expr{zero@} (= request failed and can be retried) and //! @expr{({})@} (= request definitively answered the record //! does not exist; retries are pointless). void dnsquery(string type, string domain, function(array(string)|zero, array:void) result_cb, array restargs) { int itype = dnstypetonum[type]; if (!itype) { result_cb(0, restargs); return; } if (type == "AAAA") { restargs = ({ 2, 0, result_cb }) + restargs; dnsclient->do_query(domain, Protocols.DNS.C_IN, dnstypetonum["A"], dnscb, collect2aaaa, restargs); result_cb = collect2aaaa; } dnsclient->do_query(domain, Protocols.DNS.C_IN, itype, dnscb, result_cb, restargs); } -- Stephen.

1 0

Replacing a single function/method in an existing live class?
by Stephen R. van den Berg 02 May '21

02 May '21

Say I have this: class A { int k; void B() { write("foo %d\n", k); } void C() { k = 3; write("bar\n"); } } int main() { A a = A(); a->C(); // Displays: bar a->B(); // Displays: foo 3 // At this point I want to replace the function B // in the running/compiled instance of A in a. // I.e. I want to call compile() or similar // on the following code: // void B() { write("FOO %d\n", k); } // such that I can subsequently run: a->B(); // Should display: FOO 3 return 0; } Any way this can be accomplished? An alternate way would be to compile the whole class of A again, but then run method B() in it with a custom this argument pointing to the old instance in a. Is that possible? I know javascript can do this, but it seems like Pike does not allow/support it. -- Stephen.

3 7

← Newer
1
...
4
5
6
7
8
9
10
...
305
Older →

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

Pike-devel