Pike-devel

pike-devel@lists.lysator.liu.se

2 participants
3046 discussions

System.PAM module
by Bill Welliver 01 Dec '03

01 Dec '03

I put together a glue to the Pluggable Authentication Modules framework over the weekend. I've made it available to anyone who'd like to test it at the following address: ftp://ftp.riverweb.com/pub/hww3/pike_modules/System_PAM-1.0.tar.gz This requires a relatively recent version of Pike 7.5, ie .12 or .13 should be fine. The module is documented, though the autodoc extraction for C modules isn't currently working properly. Look at pam.cmod for the raw autodocs. If you're familiar with PAM, you shouldn't have any problems using this, and there are a few examples included. I'd also be interested to hear of any design improvement type suggestions that might be floating out there, as this is my first "real" C module. I've built this module on Solaris and Linux, and have tested it under Solaris. I've tested it with dmalloc, so it shouldn't have any built in memory leaks during normal operation, but I've noticed the Solaris PAM implimentation seems to leak regardless depending on your authentication configuration (eg LDAP). Bill

2 1

module documentation
by Bill Welliver 30 Nov '03

30 Nov '03

I've run into a problem with building module documentation outside the pike tree: when a documentation build directory is within the documentation source directory, you end up with a recursion loop. this is because extract_autodoc creates a parallel tree inside of the documentation build directory, and once the extractor descends into the documentation build directory, it never reaches the end of the directories to recurse. The problem came up because autodoc markup is not extracted from C/CMOD source files using the new extract targets. I wanted to add this, but ran into the problem above. As far as I can see, there are 3 avenues: 1. build documentation outside of the documentation source directory (ie the module build directory) i don't think that this is a very good option, as it only masks the problem. 2. add a check in extract_autodoc.pike to check for the problem and stop recursing into the documentation build directory this stops the problem above, but seems to create a problem with duplicate documentation for the module in question (it ends with an error to this effect). 3. add a separate target that will extract documentation for files specified in the makefile, as say AUTODOC_C_INPUT or something, then extract the pike autodoc as normal. this would require extract_autodoc to be modified to be able to extract from a source file, rather than a directory. Anyone have any constructive thoughts about the problem in general and a solution to it? Bill

1 0

Checking threads on ppc
by Martin Stjernholm, Roxen IS ＠ Pike developers forum 25 Nov '03

25 Nov '03

When ppc machine code is used, this program won't terminate: int do_work = 1, rounds; void work() { while (do_work) { rounds++; } } int main() { for (int i = 0; i < 5; i++) Thread.thread_create (work); sleep (10); do_work = 0; sleep (1); werror ("rounds: %d\n", rounds); } The reason is that it uses OPCODE_INLINE_BRANCH and doesn't generate the appropriate calls to check_threads_etc() on backward jumps after branch opcodes. (Normally the opcodes do that themselves since they have the address to jump to.)

2 3

Re: A bug(?) in lfun::`[]() handling
by Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) ＠ Pike (-) developers forum 25 Nov '03

25 Nov '03

> I want to parse something (from specific subset) that specified inside of > [], and pass control to default `[]() operator when this something doesn't > fit to specific rules. But you can't pass something to the default [] operator which doesn't fit to the rules of the default [] operator. So it's your code which is faulty. > But I wonder why `->() was called... It shouldn't be there (at least I see > no reason why it should). The default [] implementation uses the -> implementation to save code duplication, since they have the same semantics (on objects). / Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) Previous text: >2003-11-25 21:31: >Subject: Re: A bug(?) in lfun::`[]() handling >-------------------------------------------------------------------- >On Tue, Nov 25, 2003 at 09:05:03PM +0100, Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum wrote: > >> there, although I'm not sure what. What semantics did you intend for >> your [] operator? > > I want to parse something (from specific subset) that specified inside of > [], and pass control to default `[]() operator when this something doesn't > fit to specific rules. > > But I wonder why `->() was called... It shouldn't be there (at least I see > no reason why it should). > >Regards, >/Al > > > / Brevbäraren >

1 0

A bug(?) in lfun::`[]() handling
by Alexander Demenshin 25 Nov '03

25 Nov '03

Hi, Doing some experiments (Pike 7.4.28)... So: ---snip--- class Parent { mixed `[](mixed ... args) { Stdio.stderr->write("Called `[](%O) in %O: %O\n", args, this, ::`[](@args)); return ::`[](@args); } } class Child { inherit Parent; } int main(int argc, array(string)argv) { Child obj = Child(); mixed val; val = obj[-1..1]; // Just in case - optimizer will cut the code when "val =" is removed. } ---snip--- The attempt to execute gives: Bad argument 1 to ::->(). Expected string Unknown program: ::->(-1,1) object.c:2105: object.c:2105()->()(-1,1) inh2.pike:5: inh2()->Child()->](-1,1) inh2.pike:20: inh2()->main(1,({"/home/aldem/src/pike/inh2.pike"})) I don't have any definition of `->() nor any references to it... And no references in docs that `->() and `[]() are equivalent... In fact, inheritance doesn't matter - it is just my standard template for test cases... Substitute Child to Parent and the error will be exactly the same. Not sure if problem exists with 7.5 or 7.4-cvs version - didn't check it out (yet)... Regards, /Al

2 1

Bootstraping pike-cvs and stripped down source dist
by Alexander Demenshin 25 Nov '03

25 Nov '03

Hi, Is there any way to build pike-cvs _without_ working Pike installed? And second question... Some time ago someone posted here a minimum set of of modules which are really necessary to install/run Pike, so the question is - is it possible to "strip" all unnecessary modules from the source tree so they won't be compiled at all? Will simply removal of modules from the source work? For those who wonder why - I just want to release an application (naturally, written in Pike), but I don't want to force users to install full version of Pike (they don't need it at all but to run my app, and I know exactly what I (don't) use). Regards, /Al

3 2

fake_dynamic_load.pike
by Martin Nilsson (saturator) ＠ Pike (-) developers forum 22 Nov '03

22 Nov '03

When looking over the bin folder manifest I found that fake_dynamic_load.pike isn't documented. What is it for and when is it used?

4 7

alloca usage
by nisse＠lysator.liu.se 20 Nov '03

20 Nov '03

It seems that the default configuration of GMP, with WANT_TMP_ALLOCA, allocates arbitrarily large temporaries on the stack. For example mpn_tdiv_qr allocates temporary storage proportional to the size of the input. Fredrik Hübinette (Pike author) reports that his pike program that calculates a million decimals of Pi crashes in gmp with a default configuration of gmp-4.1.2, but works fine if gmp is reconfigured with --enable-alloca=malloc-reentrant. The problem is made worse by linux threads (and perhaps threads on other systems as well), which limits stack space to only 4 MB. (See http://fredrik.hubbe.net/hacker/viewtopic.php?t=26). Perhaps it would be a good idea to define TMP_ALLOC as something like #define TMP_ALLOC(size) \ ((size) < PAGESIZE*2) ? alloca((size)) : heap_alloc((size), &tmp_marker) Then for small inputs, we get only a single comparison of extra overhead, while for large inputs, the allocation overhead should be small compared to the cost of the actual computation. Due to the way the "red zone" stack extension mechanism works, I don't think it's reliable to allocate more than a few pages of data on the stack at a time. If you allocate too many pages you may get a segmentation violation of you access the pages in an unfortunate order. Regards, /Niels

1 0

Strange GMP coredumps
by Fredrik (Naranek) Hubinette (Real Build Master) ＠ Pike (-) developers forum 20 Nov '03

20 Nov '03

I've figured out why Pike dumps core when using very large numbers in GMP. See this link for more info: http://fredrik.hubbe.net/hacker/viewtopic.php?t=26

3 5

Strange GMP coredumps
by Johan Sundstr�m (f�ngar heffaklumpar) ＠ Pike (-) developers forum 20 Nov '03

20 Nov '03

(For Linux systems, at least.)

1 0

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