Pike-devel June 2008

pike-devel@lists.lysator.liu.se

9 participants
20 discussions

Start a nNew thread

Nettle
by Magnus Holmgren, Millnet AB (nu med fast anst�llning) ＠ Pike (-) developers forum 17 Jun '08

17 Jun '08

Could you upgrade the bundled version of Nettle to 1.15, which among other things, gets rid of a non-free RFC?

4 7

pike crashes
by Peter Bortas ＠ Pike developers forum 13 Jun '08

13 Jun '08

For the record: This did not get exported to the list, Nilssons answer did.

1 0

Valgrinding 7.7
by Martin Nilsson (Opera Mini - AFK!) ＠ Pike (-) developers forum 12 Jun '08

12 Jun '08

I'm getting some valgrind errors that appears to be because of programs not being properly initiated. Anyone care to look at it? ==7004== Conditional jump or move depends on uninitialised value(s) ==7004== at 0x81208DD: index_modules (program.c:1521) ==7004== by 0x8120B4E: find_module_identifier (program.c:1596) ==7004== by 0x80738E5: yyparse (language.yacc:3878) ==7004== by 0x8122997: f_compilation_compile (program.c:7854) ==7004== by 0x807C5BA: low_mega_apply (apply_low.h:225) ==7004== by 0x807DA5A: mega_apply (interpret.c:2200) ==7004== by 0x807DDEF: apply (interpret.c:2591) ==7004== by 0x811721E: f_compilation_env_compile (program.c:7520) ==7004== by 0x807C5BA: low_mega_apply (apply_low.h:225) ==7004== by 0x807DA5A: mega_apply (interpret.c:2200) ==7004== by 0x816EA79: f_compile (builtin_functions.c:4145) ==7004== by 0x8106550: get_master (object.c:636) ==7004== Uninitialised value was created by a stack allocation ==7004== at 0x8122CE9: init_program (program.c:9565) ==7004== Conditional jump or move depends on uninitialised value(s) ==7004== at 0x8163DB0: optimize (treeopt.h:520) ==7004== by 0x8167703: dooptcode (las.c:5443) ==7004== by 0x807319C: yyparse (language.yacc:1056) ==7004== by 0x8122693: run_pass2 (program.c:7924) ==7004== by 0x812281D: f_compilation_compile (program.c:8374) ==7004== by 0x807C5BA: low_mega_apply (apply_low.h:225) ==7004== by 0x807DA5A: mega_apply (interpret.c:2200) ==7004== by 0x807DDEF: apply (interpret.c:2591) ==7004== by 0x811721E: f_compilation_env_compile (program.c:7520) ==7004== by 0x807C5BA: low_mega_apply (apply_low.h:225) ==7004== by 0x807DA5A: mega_apply (interpret.c:2200) ==7004== by 0x816EA79: f_compile (builtin_functions.c:4145) ==7004== Uninitialised value was created by a stack allocation ==7004== at 0x8122CE9: init_program (program.c:9565) ==7004== Conditional jump or move depends on uninitialised value(s) ==7004== at 0x8165C3B: optimize (treeopt.h:2638) ==7004== by 0x8167703: dooptcode (las.c:5443) ==7004== by 0x807319C: yyparse (language.yacc:1056) ==7004== by 0x8122693: run_pass2 (program.c:7924) ==7004== by 0x812281D: f_compilation_compile (program.c:8374) ==7004== by 0x807C5BA: low_mega_apply (apply_low.h:225) ==7004== by 0x807DA5A: mega_apply (interpret.c:2200) ==7004== by 0x807DDEF: apply (interpret.c:2591) ==7004== by 0x811721E: f_compilation_env_compile (program.c:7520) ==7004== by 0x807C5BA: low_mega_apply (apply_low.h:225) ==7004== by 0x807DA5A: mega_apply (interpret.c:2200) ==7004== by 0x816EA79: f_compile (builtin_functions.c:4145) ==7004== Uninitialised value was created by a stack allocation ==7004== at 0x815E4F3: debug_mkprgnode (las.c:984)

1 0

pike crashes
by Martin Nilsson (Opera Mini - AFK!) ＠ Pike (-) developers forum 12 Jun '08

12 Jun '08

> This one is a bug in Hilfe: > > > sscanf("abc"[0..2],"%s",s); > Compiler Error: 1: syntax error, unexpected TOK_DOT_DOT > Compiler Error: 1: syntax error, unexpected ']', expecting ',' or ')' > Compiler Error: 1: syntax error, unexpected TOK_NUMBER > Compiler Error: 1: Missing ']'. > Compiler Error: 1: syntax error, unexpected ']', expecting TOK_LEX_EOF or ';' Fixed.

1 0

pike crashes
by Martin Bähr 12 Jun '08

12 Jun '08

hi, just stumbled accross this problem: this is pikefarm build 2898 Pike v7.7 release 47 running Hilfe v3.5 (Incremental Pike Frontend) > sscanf(("abc"[0..2]),"%s",s); fish: Job 1, #~pikefarm/bin/pike# terminated by signal SIGSEGV (Address boundary error) also the following issues: Pike v7.7 release 47 running Hilfe v3.5 (Incremental Pike Frontend) > string s; sscanf(("abc"[0..2]),"%s",s); (1) Result: 1 > sscanf(("abc"[0..2]),"%s",s); (2) Result: 1 > sscanf("abc"[0..2],"%s",s); Compiler Error: 1: parse error, unexpected `TOK_DOT_DOT' Compiler Error: 1: parse error, unexpected `']'', expecting `','' or `')'' Compiler Error: 1: parse error, unexpected `TOK_NUMBER' Compiler Error: 1: Missing ']'. Compiler Error: 1: parse error, unexpected `']'', expecting `TOK_LEX_EOF' or `';'' > sscanf(("abc"[0..2]),"%s",s); Compiler Error: 1: parse error, unexpected `TOK_NUMBER' Compiler Error: 1: parse error, unexpected `']'', expecting `')'' Compiler Error: 1: parse error, unexpected `TOK_DOT_DOT' Compiler Error: 1: Missing ']'. Compiler Error: 1: parse error, unexpected `TOK_DOT_DOT' Compiler Error: 1: parse error, unexpected `']'', expecting `')'' Compiler Error: 1: parse error, unexpected `TOK_NUMBER' Compiler Error: 1: parse error, unexpected `']'', expecting `error' or `','' Compiler Error: 1: parse error, unexpected `')'', expecting `TOK_LEX_EOF' or `';'' this happens in 7.6 as well: Pike v7.6 release 66 running Hilfe v3.5 (Incremental Pike Frontend) > "abc"[0..2]; (1) Result: "abc" > sscanf("abc"[0..2],"%s",s); Compiler Error: 1:syntax error, unexpected TOK_DOT_DOT Compiler Error: 1:syntax error, unexpected ']', expecting ',' or ')' Compiler Error: 1:syntax error, unexpected TOK_NUMBER Compiler Error: 1:Missing ']'. > "abc"[0..2]; Compiler Error: 1:syntax error, unexpected TOK_NUMBER Compiler Error: 1:syntax error, unexpected ']', expecting ')' Compiler Error: 1:syntax error, unexpected TOK_DOT_DOT Compiler Error: 1:Missing ']'. Compiler Error: 1:syntax error, unexpected TOK_DOT_DOT Compiler Error: 1:syntax error, unexpected ']', expecting ')' Compiler Error: 1:syntax error, unexpected TOK_NUMBER greetings, martin.

1 0

installing documentation missing html?
by Martin Bähr 10 Jun '08

10 Jun '08

hi, i played around with installing pike out of a pikefarm build and i somehow can't seem to find the generated documentation. i do find an xml file, but no html files. what's missing? i was running make install_interactive giving the users homedirectory as install-target the html files do get built when xenofarm runs. greetings, martin.

1 1

Re: _main call in OCPikeInterpreter.m
by Martin Stjernholm, Roxen IS ＠ Pike developers forum 06 Jun '08

06 Jun '08

It also does a SETJMP and several lines of error handling, only to be followed by a block that effectively executes no code that could have use for it. So it's not only priming for something that happens somewhere else. To me it looks just as if the very call to _main itself is missing there. And no, _main no longer uses that, that's why the function is gone. If someone wants the environment in that (bulky) form, (s)he now has to do it by converting the mapping returned by __builtin._getenv.

1 0

_main call OCPikeInterpreter.m
by Martin Stjernholm, Roxen IS ＠ Pike developers forum 06 Jun '08

06 Jun '08

I've just committed a fix so that putenv and getenv actually do what they claim to do. As part of this, master()->_main no longer takes an environment mapping. I think I've made the corresponding change for that in OCPikeInterpreter.m too, but someone who knows what's going on there better take a look at it, because to me the call to master()->_main seems to be missing altogether.

2 2

7.8 blockers
by Martin Nilsson (Opera Mini - AFK!) ＠ Pike (-) developers forum 01 Jun '08

01 Jun '08

For Tools.Standalone.dump constant dont_dump_program=1 is a solution.

1 0

Setting things to NULL after free
by Martin Stjernholm, Roxen IS ＠ Pike developers forum 01 Jun '08

01 Jun '08

It's been customary to set pointers to NULL after freeing them in object (and module) exit functions. I'd like to question whether that really makes much sense. The object storage is only going to be freed just afterwards anyway (same goes for dynamically loaded modules), so the only bugs that this kind of "hygiene" could help nail is if destruct_object or exit_dynamic_load misbehaves. That's a bit unlikely by now, and anyway it's an awful lot of code everywhere just for that. The assignments are not even inside #ifdef PIKE_DEBUG as such belts-and-suspenders things usually are. Or am I missing something here? Also, when a pointer is not supposed to be used anymore, I usually do something like this instead: free (x->foo); #ifdef PIKE_DEBUG x->foo = (void *) (ptrdiff_t) -1; #endif That is more likely to trig a helpful coredump if the pointer is being used incorrectly.

1 0

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 June 2008