Pike-devel February 2004

pike-devel@lists.lysator.liu.se

8 participants
62 discussions

builtin.cmod: backtrace_frame->fill_in_file_and_line
by Niels M�ller (v�ssar r�dpennan) ＠ Pike (-) developers forum 10 Feb '04

10 Feb '04

Another compiler warning: static void fill_in_file_and_line() { struct pike_string *file; if (THIS->pc && THIS->prog) file = ... else if (THIS->fun.type == PIKE_T_FUNCTION) file = .... else if (THIS->prog) file = ... if (!THIS->filename) THIS->filename = file; else free_string (file); Compiler complains that file isn't initialized, and it isn't, if none the three tests is true.

2 1

Protocols.HTTP.Server.* under heavy load
by Mirar ＠ Pike developers forum 10 Feb '04

10 Feb '04

Umm, someone might want to backport these fixes to 7.4 from 7.5 :) (That is, HTTP.pmod/Server.pmod/*, probably?) / Mirar Previous text: >2004-02-10 22:08: >Subject: Protocols.HTTP.Server.* under heavy load >-------------------------------------------------------------------- >Yes. It leaked fd's prior to the checkin I did a few weeks ago. >(I run it as my main server at www.mirar.org...) > >---------------------------- >revision 1.18 >date: 2004/01/24 10:03:31; author: mirar; state: Exp; lines: +73 -4 >o more paranoid closing of connections >o support for if-modified-since and range in preparation of http/1.1 >---------------------------- > >Seems to work fine since that fix. > > / Mirar >

1 0

7.6
by Martin Nilsson (saturator) ＠ Pike (-) developers forum 09 Feb '04

09 Feb '04

The most difficult part of releasing a new Pike, besides deciding when to actually do it, is to make the changelog. Could people please take a look at CHANGES. It contains a quick list of changes up to 2003-09-30. These needs to be elaborated and everything that has happened in the quarter since needs to be added.

2 1

sscanf and Gmp.mpq
by Johan Sundstr�m (Achtung Liebe!) ＠ Pike (-) developers forum 07 Feb '04

07 Feb '04

Would a sscanf format %q for turning numbers (possibly with decimals) to Gmp.mpq objects be a good, bad, easy and/or hard addition to make? (I was just considering looking at it when I wake up later on today.)

5 45

sscanf and Gmp.mpq
by Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) ＠ Pike (-) developers forum 07 Feb '04

07 Feb '04

In fact, the object could be given by the lvalue: Gmp.mpq q = Gmp.mpq(); sscanf("17", "%¤f", q); That won't work for array_sscanf() though.

1 0

wordbits.h
by Niels M�ller (v�ssar r�dpennan) ＠ Pike (-) developers forum 06 Feb '04

06 Feb '04

Should the end of the ranges array really look like {0x01d1aa,0x01d1ad}, {0x0e0100-0x10fffd} }; ? (I don't know what this is used for).

4 4

Nettle 1.7 and pike 7.5
by Xavier Beaudouin 06 Feb '04

06 Feb '04

Hello there, Seems that Pike 7.5 is not compatible with nettle 1.7... because it look for nettle_md5_init() and since it doesn't exist on Nettle 1.7.... It fails... Ok, now this is bad that 1.8 has some api changes.... and/or that Pike 7.5 cannot cope with Nettle 1.7 ;( My question is why do pike doen't support both api ??? /Xavier

3 14

block_alloc(_h).h
by Niels M�ller (v�ssar r�dpennan) ＠ Pike (-) developers forum 06 Feb '04

06 Feb '04

I get lots of errors like Compiling /home/nisse/hack/pike/src/pike_search.c In file included from /home/nisse/hack/pike/src/stralloc.h:12, from /home/nisse/hack/pike/src/pike_memory.h:12, from /home/nisse/hack/pike/src/pike_search.c:15: /home/nisse/hack/pike/src/block_alloc_h.h:20:2: #error "block_alloc.h must be included after all uses of block_alloc_h.h" What's this about?

3 2

Nettle
by Niels M�ller (v�ssar r�dpennan) ＠ Pike (-) developers forum 06 Feb '04

06 Feb '04

I think it may be possible to do the sha1 compression function all in x86 registers. Five registers for the state, one for pointing to the input, and then one free temporary. Benchmark for the C implementation of various hashes: md2 (Update): 2.327MB/s md4 (Update): 171.846MB/s md5 (Update): 114.488MB/s sha1 (Update): 81.916MB/s sha256 (Update): 43.055MB/s / Niels Möller (vässar rödpennan) Previous text: >2004-02-05 17:43: >Subject: Nettle >-------------------------------------------------------------------- >If you are in a good hacking mood, any speedup in SHA1 would be most >welcome. Since it is used in bittorrent, quite a few gigabytes runs >through it. > > / Martin Nilsson (saturator) >

6 13

SSL part two
by Bill Welliver 05 Feb '04

05 Feb '04

I noticed that SSL.handshake supports certificate authentication. That's actually rather nice, however, I've got a problem with the way it's controlled. If the settings are in handshake, there's no way to change them when using sslport, or any other function that uses sslfile directly. Would it make more sense to move those settings to the context object? Would anyone be terribly put out if I did that (or at least added the option)? Any thoughts? Bill

3 13

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 February 2004