August 2013 - Dillo-dev - lists.mailman3.com

Proposed Patch for Zombies Left by Dpi_start_dpid() in dpi.c
by jgaffney＠lawrenceville.org 02 Sep '13

02 Sep '13

It seems that the dpid children forked by Dpi_start_dpid() in dpi.c are not reaped when they terminate themselves after a period of inactivity while the parent dillo process is still running. The patch below is simply to reap these zombies. The SIGCHLD handler should probably do more, but I am not yet familiar enough with Dillo to know what. I believe this patch adheres to the Dillo rules of coding style, but am new to the list and a novice at Dillo patching. And hence also apologies to Jorge for this duplication of a previous communication by private email (although with one small change to get the patch to compile under OpenBSD as well as Linux). ######## BEGIN PATCH ######## --- a/src/IO/dpi.c 2013-01-27 12:26:38.000000000 -0500 +++ b/src/IO/dpi.c 2013-08-27 14:58:07.995235867 -0400 @@ -33,6 +33,8 @@ #include <netinet/tcp.h> #include <arpa/inet.h> #include <netdb.h> +#include <sys/wait.h> +#include <signal.h> #include "../msg.h" #include "../klist.h" @@ -331,16 +333,33 @@ } } +static void Sigchld_handler (int signum) { + int status; + waitpid (-1, &status, WNOHANG); + } + /* * Start dpid. * Return: 0 starting now, 1 Error. */ static int Dpi_start_dpid(void) { + static int install_sigchld_handler_flag = TRUE; pid_t pid; int st_pipe[2], ret = 1; char *answer; + if ( install_sigchld_handler_flag ) { + /* Install SIGCHLD handler. */ + struct sigaction sa; + sa.sa_handler = Sigchld_handler; + sigemptyset(&sa.sa_mask); + sa.sa_flags = SA_RESTART | SA_NOCLDSTOP; + if (sigaction(SIGCHLD, &sa, 0)) + MSG ("Unable to install SIGCHLD handler.\n"); + install_sigchld_handler_flag = FALSE; + } + /* create a pipe to track our child's status */ if (pipe(st_pipe)) return 1; ######## END PATCH ########

4 7

Proposed Patch for Zombies Left by Dpi_start_dpid() in dpi.c
by jgaffney＠lawrenceville.org 31 Aug '13

31 Aug '13

Jorge Arellano Cid <jcid at dillo.org> writes: > Hi Johannes, JG: > [...] > The attached program is a stress test for SIGCHLD. The idea is to > find a formula that works well even in such cases. IIRC both of you > have access to BSD machines so please test it there. > > The program spawns 255 children which sleep 5 seconds and > terminate generating a SIGCHLD avalanche, which should be handled > cleanly. It works OK in Debian. [...] 1) OpenBSD is very strict about compliance with standards, so that, whereas on my (and I believe most if not all) Linux systems wait.h occurs redundantly in /usr/include/ and /usr/include/sys, on OpenBSD wait.h is solely in /usr/include/sys and your code will not compile unless the `#include<wait.h>' line is removed. Then `gcc -W -Wall' compiles it without complaint. 2) OpenSBD is not so extravagant with the resources allocated to unprivileged users (max 128 processes). But spawning 64 rather than 256 children the test is passed with the expected behavior on OpenBSD 5.2 GENERIC.MP#368 amd64. The OpenBSD 5.2 ports include Dillo 2.2 and FLTK2, but I just recently compiled FLTK 1.3.2 and Dillo 3.0.3 from their unmodified source on that machine without (IIRC) any errors. If you were curious more generally to see how they are modifying the Dillo source, the OpenBSD patches for Dillo 3.0.3 (the version currently in ports for OpenBSD 5.3) may be found at http://www.openbsd.org/cgi-bin/cvsweb/ports/www/dillo/patches/

2 1

Proposed Patch for Zombies Left by Dpi_start_dpid() in dpi.c
by jgaffney＠lawrenceville.org 29 Aug '13

29 Aug '13

Johannes Hofmann <Johannes.Hofmann at gmx.de> writes: > Hi JG, > > On Tue, Aug 27, 2013 at 03:15:14PM -0400, JG wrote: >> It seems that the dpid children forked by Dpi_start_dpid() in dpi.c >> are not reaped when they terminate themselves after a period of >> inactivity while the parent dillo process is still running. The patch >> below is simply to reap these zombies. The SIGCHLD handler should >> probably do more, but I am not yet familiar enough with Dillo to know >> what. >> >> I believe this patch adheres to the Dillo rules of coding style, but >> am new to the list and a novice at Dillo patching. And hence also >> apologies to Jorge for this duplication of a previous communication by >> private email (although with one small change to get the patch to >> compile under OpenBSD as well as Linux). > > I can reproduce the issue here by killing dpid manually. > For me the following also fixes it: > > diff -r 659dc205c377 src/dillo.cc > --- a/src/dillo.cc Wed Aug 21 11:42:38 2013 +0200 > +++ b/src/dillo.cc Wed Aug 28 19:48:10 2013 +0200 > @@ -329,6 +329,7 @@ > > // Some OSes exit dillo without this (not GNU/Linux). > signal(SIGPIPE, SIG_IGN); > + signal(SIGCHLD, SIG_IGN); > > /* Handle command line options */ > opt_argv = dNew0(char*, numOptions(Options) + 1); > > Can you please check whether it works for you too? It would have the > advantage that we have all signal related stuff in one place. > > Cheers, > Johannes Thank you: your solution is of course not only correct but far simpler and better. (Though I thought that in general sigaction() is to be preferred to signal()?)

1 0

allow_white_bg
by sgeerken＠dillo.org 21 Aug '13

21 Aug '13

Hi! It seems that I'm getting old that I tried to set the option "allow_white_bg" to "false"; to find out that the replacement color is hardcoded! If I haven't missed something, and noone objets, I'll add a prefs variable "white_bg_replacement". BTW: is there a reason that there is still a prefs variable "bg_color", while fonts (the actual fonts, not the font classes) are now defined by CSS? Sebastian

2 3