Pike-devel July 2012

pike-devel@lists.lysator.liu.se

7 participants
21 discussions

Re: Filesystem.System: get_dir and get_stat behave differently?
by Tobias S. Josefowitz 21 Jul '12

21 Jul '12

On Fri, Jul 20, 2012 at 11:55 PM, Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum <10353(a)lyskom.lysator.liu.se> wrote: > should be relative to the chroot point. The correct value is the of > full before the initial / was removed. But you could also use > "/"+full. Apparently full does not neccessarily include the / anyway. :)

2 1

Re: Filesystem.System: get_dir and get_stat behave differently?
by Tobias S. Josefowitz 20 Jul '12

20 Jul '12

On Fri, Jul 20, 2012 at 2:20 PM, Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum <10353(a)lyskom.lysator.liu.se> wrote: > variable "full"). So putting back the / should restore compatibility > with how it used to (and was supposed to) work before May 2001. .) - ok then.

3 2

Re: Filesystem.System: get_dir and get_stat behave differently?
by Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) ＠ Pike (-) developers forum 20 Jul '12

20 Jul '12

Before dcde88c3, the fullpath was set including the /. Mirar added stripping of the initial / to fix a chroot bug, which presumably means that he wanted to change the third parameter to combine_path(), not the value stored in stat->fullpath (which is set from the same local variable "full"). So putting back the / should restore compatibility with how it used to (and was supposed to) work before May 2001.

1 2

Re: Filesystem.System: get_dir and get_stat behave differently?
by Tobias S. Josefowitz 20 Jul '12

20 Jul '12

Hmm. I doubt I can just make fullpath contain the leading / because of compat etc, and the Filesystem.System()->cd() cannot just add it either. Any suggestions?

2 1

Filesystem.System: get_dir and get_stat behave differently?
by Bill Welliver 20 Jul '12

20 Jul '12

I noticed that if I have a Filesystem.System object and I chroot(), calling get_dir() with an absolute path will break out of the chroot. This doesn't happen with get_stat(), and it seems like it might be a bug. For example: Filesystem.System("/etc")->chroot("/etc")->get_dir("/") returns a list of items in the system root (/). If I do the same thing but replace the get_dir with get_stat(), I get a list of stat objects for /etc. The latter (get_stat) behavior seems correct to me. Can anyone confirm that the current behavior is actually correct? If not, I'll look into fixing it. Best, Bill

3 2

Failed test on 7.8.695
by Bill Welliver 18 Jul '12

18 Jul '12

I'm getting a test failure on the latest 7.8 build, and it looks like it exists to check a fix from last week. Please have a look and let me know if I should hold off on doing a new build for a fix. Thanks! Bill testsuite:2: syntax error, unexpected TOK_CONSTANT /Users/hww3/pike-git/src/testsuite.in:544: Test 138 (shift 0) failed. 1: mixed a() { return 2: constant foo = 5; 3: 4: // Test that it survives a F_ARG_LIST with CAR == CDR. 5: return foo%foo; 6: ; } 7: mixed b() { return 0; } 8:

4 4

proposed backport from 7.9 to 7.8
by Bill Welliver 18 Jul '12

18 Jul '12

I've made a number of commits to 7.9 surrounding the external module support tools (monger and pmar_install) that I'd like to include in the next 7.8 release. Also, I'd like to backport the fixes and additions to spawn_pike and the new get_user/get_home functions (but not FreeConsole and friends) in System as well. The code is fairly well isolated and don't think there will be any downside. Considering the utility of the functionality and that 7.9 is likely to be in development for the foreseeable future, I think it's worth doing. Does anyone have any objection? Bill

3 2

New Pike Beta: 7.8.694
by H. William Welliver III 12 Jul '12

12 Jul '12

I've made a new beta build which for your consideration. If no problems are found, this build will become the next stable release. Please download, build, test, make binary packages, etc. As always, please report back with any problems (or successes, too!) http://pike.ida.liu.se/download/pub/pike/beta/7.8.694/ Best, Bill

8 24

Pike type regression
by Martin Nilsson (Opera Mini - AFK!) ＠ Pike (-) developers forum 11 Jul '12

11 Jul '12

nilsson@riaa:~/pike$ make run_hilfe ARGS="-V7.6" Configure arguments: Use `make CONFIGUREARGS="..." ...' to change them. They will be retained in the build directory. Making run_hilfe in build/linux-3.0.0-22-generic-x86_64 /home/nilsson/pike/build/linux-3.0.0-22-generic-x86_64/pike -DNOT_INSTALLED -DPRECOMPILED_SEARCH_MORE -m/home/nilsson/pike/build/linux-3.0.0-22-generic-x86_64/master.pike -V7.6 Pike v7.9 release 5 running Hilfe v3.5 (Incremental Pike Frontend) (running in Pike 7.6 compat mode) > Calendar.Events; make[2]: *** [run_hilfe] Segmentation fault make[1]: *** [_make_in_builddir] Error 2 make: *** [run_hilfe] Error 2 (perhaps I should get a bugs.roxen.com account again...)

1 0

roxen.c
by Martin Nilsson (Opera Mini - AFK!) ＠ Pike (-) developers forum 07 Jul '12

07 Jul '12

When not to use dash as punctuation :) i = -i; /* Always defined for unsigned - signed overflows are not. */

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 July 2012