- Dillo-dev - lists.mailman3.com

Re: image loading from menus

by jcid＠dillo.org

Hi, On Sat, Nov 03, 2007 at 11:18:36PM +0000, place wrote: > Jorge wrote: > > After some study and a reimplentation of plain.cc with classes > > (getting rid of its linkblock and hooking a callback for memory > > handling), it looks simpler and better. It looks like the same > > may be done with html.cc. > > I was just wondering how you weigh the options of adding a callback > vs. adding a signal. I had experimented with a destroy signal in > my tree recently. It seemed more consistent to me, although I suspect > the various child widgets were wasting cycles on signal emission code > when they didn't have any receivers... I'd say that for me, the key is simplicity, and elegance. In a sense, accurately captured by this quote: "The key to performance is elegance, not batalions of special cases" -- Jon Bentley and Doug McIlroy In the case of Dillo, I've seen several times the simplicity pay off. When code gets simpler, debugging and improving accelerates. With complexity it stagnates. Elegance helps to keep the design simple and flexible. Having flexibilty in the design, helps to tackle unforeseen problems at the design level, without workarounds, reducing complexity and special cases. -- Cheers Jorge.-

Nov. 4, 2007

Crashes on OpenBSD

by Johannes.Hofmann＠gmx.de

Hello, as I'm also using a BSD system, I converted the D_DNS_THREADED code to use the inherently thread-safe getaddrinfo(3) interface that also has some other advantages (e.g. a predictable ordering of the returned addresses). Jorge, what do you think about this? Are there supported platforms that do not have getaddrinfo? Should we add a compile option for this? Johannes diff --git a/src/dns.c b/src/dns.c --- a/src/dns.c +++ b/src/dns.c @@ -239,58 +239,86 @@ static void Dns_note_hosts(Dlist *list, } } +static void Dns_note_hosts2(Dlist *list, struct addrinfo *res0) +{ + struct addrinfo *res; + DilloHost *dh; + + for (res = res0; res; res = res->ai_next) { + + if (res->ai_family == AF_INET) { + struct sockaddr_in *in_addr; + + if (res->ai_addrlen < sizeof(struct sockaddr_in)) { + continue; + } + + dh = dNew0(DilloHost, 1); + dh->af = AF_INET; + + in_addr = (struct sockaddr_in*) res->ai_addr; + dh->alen = sizeof (struct in_addr); + memcpy(&dh->data[0], &in_addr->sin_addr.s_addr, dh->alen); + + dList_append(list, dh); +#ifdef ENABLE_IPV6 + } else if (res->ai_family == AF_INET6) { + struct sockaddr_in6 *in6_addr; + + if (res->ai_addrlen < sizeof(struct sockaddr_in6)) { + continue; + } + + dh = dNew0(DilloHost, 1); + dh->af = AF_INET6; + + in6_addr = (struct sockaddr_in6*) res->ai_addr; + dh->alen = sizeof (struct in6_addr); + memcpy(&dh->data[0], &in6_addr->sin6_addr.s6_addr, dh->alen); + + dList_append(list, dh); +#endif + } + } +} + #ifdef D_DNS_THREADED /* * Server function (runs on its own thread) */ static void *Dns_server(void *data) { - struct hostent *host; int channel = VOIDP2INT(data); -#ifdef LIBC5 - int h_err; - char buff[1024]; - struct hostent sh; -#endif + struct addrinfo hints, *res0; + int error; + + memset(&hints, 0, sizeof(hints)); + hints.ai_family = PF_INET; + hints.ai_socktype = SOCK_STREAM; + Dlist *hosts = dList_new(2); DEBUG_MSG(3, "Dns_server: starting...\n ch: %d host: %s\n", channel, dns_server[channel].hostname); -#ifdef ENABLE_IPV6 - if (ipv6_enabled) { - host = gethostbyname2(dns_server[channel].hostname, AF_INET6); - if (host) { - Dns_note_hosts(hosts, AF_INET6, host); - } - } -#endif - -#ifdef LIBC5 - host = gethostbyname_r(dns_server[channel].hostname, &sh, buff, - sizeof(buff), &h_err); -#else - host = gethostbyname(dns_server[channel].hostname); -#endif - - if (!host) { -#ifdef LIBC5 - dns_server[channel].status = h_err; -#else - dns_server[channel].status = h_errno; - if (h_errno == HOST_NOT_FOUND) + error = getaddrinfo(dns_server[channel].hostname, NULL, &hints, &res0); + + if (error != 0) { + dns_server[channel].status = error; + if (error == EAI_NONAME) MSG("DNS error: HOST_NOT_FOUND\n"); - else if (h_errno == TRY_AGAIN) + else if (error == EAI_AGAIN) MSG("DNS error: TRY_AGAIN\n"); - else if (h_errno == NO_RECOVERY) - MSG("DNS error: NO_RECOVERY\n"); - else if (h_errno == NO_ADDRESS) + else if (error == EAI_NODATA) MSG("DNS error: NO_ADDRESS\n"); -#endif - } else { + else if (h_errno == EAI_FAIL) + MSG("DNS error: NO_RECOVERY\n"); + } else { + Dns_note_hosts2(hosts, res0); dns_server[channel].status = 0; - Dns_note_hosts(hosts, AF_INET, host); - } + freeaddrinfo(res0); + } + if (dList_length(hosts) > 0) { dns_server[channel].status = 0; } else { diff --git a/src/dns.h b/src/dns.h --- a/src/dns.h +++ b/src/dns.h @@ -1,5 +1,7 @@ #ifndef __DNS_H__ #define __DNS_H__ + +#include <netinet/in.h> #include "chain.h" @@ -14,7 +16,7 @@ void a_Dns_freeall(void); void a_Dns_freeall(void); void a_Dns_resolve(const char *hostname, DnsCallback_t cb_func, void *cb_data); -#define DILLO_ADDR_MAX 16 +#define DILLO_ADDR_MAX sizeof(struct in6_addr) typedef struct _DilloHost {

Nov. 4, 2007

PATCH: adds the ' entity

by onepoint＠starurchin.org

Is it dillo's policy to include *only* HTML 4.01 entities. If not, here's a patch to add the ' entity, which appears all over the place despite not being official HTML 4.01. Regards, Jeremy Henty Only in dillo2: CVS Only in dillo2-cur: Makefile Only in dillo2-cur: Makefile.in Only in dillo2-cur: aclocal.m4 Only in dillo2-cur: autom4te.cache Only in dillo2-cur: config.guess Only in dillo2-cur: config.h Only in dillo2-cur: config.log Only in dillo2-cur: config.status Only in dillo2-cur: config.sub Only in dillo2-cur: configure Only in dillo2-cur: d_size.h Only in dillo2-cur: depcomp Only in dillo2-cur/dlib: .deps Only in dillo2/dlib: CVS Only in dillo2-cur/dlib: Makefile Only in dillo2-cur/dlib: Makefile.in Only in dillo2-cur/dlib: dlib.o Only in dillo2-cur/dlib: libDlib.a Only in dillo2/doc: CVS Only in dillo2-cur/doc: Makefile Only in dillo2-cur/doc: Makefile.in Only in dillo2-cur/dpi: .deps Only in dillo2/dpi: CVS Only in dillo2-cur/dpi: Makefile Only in dillo2-cur/dpi: Makefile.in Only in dillo2-cur/dpi: bookmarks.dpi Only in dillo2-cur/dpi: bookmarks.o Only in dillo2-cur/dpi: cookies.dpi Only in dillo2-cur/dpi: cookies.o Only in dillo2-cur/dpi: datauri.filter.dpi Only in dillo2-cur/dpi: datauri.o Only in dillo2-cur/dpi: downloads.dpi Only in dillo2-cur/dpi: downloads.o Only in dillo2-cur/dpi: dpiutil.o Only in dillo2-cur/dpi: file.dpi Only in dillo2-cur/dpi: file.o Only in dillo2-cur/dpi: ftp.filter.dpi Only in dillo2-cur/dpi: ftp.o Only in dillo2-cur/dpi: hello.filter.dpi Only in dillo2-cur/dpi: hello.o Only in dillo2-cur/dpi: https.filter.dpi Only in dillo2-cur/dpi: https.o Only in dillo2-cur/dpid: .deps Only in dillo2/dpid: CVS Only in dillo2-cur/dpid: Makefile Only in dillo2-cur/dpid: Makefile.in Only in dillo2-cur/dpid: dpi.o Only in dillo2-cur/dpid: dpi_service.o Only in dillo2-cur/dpid: dpi_socket_dir.o Only in dillo2-cur/dpid: dpid Only in dillo2-cur/dpid: dpid.o Only in dillo2-cur/dpid: dpid_common.o Only in dillo2-cur/dpid: main.o Only in dillo2-cur/dpid: misc_new.o Only in dillo2-cur/dpip: .deps Only in dillo2/dpip: CVS Only in dillo2-cur/dpip: Makefile Only in dillo2-cur/dpip: Makefile.in Only in dillo2-cur/dpip: dpip.o Only in dillo2-cur/dpip: libDpip.a Only in dillo2-cur: install-sh Only in dillo2-cur: missing Only in dillo2-cur/src: .deps Only in dillo2/src: CVS Only in dillo2-cur/src/IO: .deps Only in dillo2/src/IO: CVS Only in dillo2-cur/src/IO: IO.o Only in dillo2-cur/src/IO: Makefile Only in dillo2-cur/src/IO: Makefile.in Only in dillo2-cur/src/IO: about.o Only in dillo2-cur/src/IO: dpi.o Only in dillo2-cur/src/IO: http.o Only in dillo2-cur/src/IO: iowatch.o Only in dillo2-cur/src/IO: libDiof.a Only in dillo2-cur/src/IO: mime.o Only in dillo2-cur/src/IO: proto.o Only in dillo2-cur/src: Makefile Only in dillo2-cur/src: Makefile.in Only in dillo2-cur/src: bitvec.o Only in dillo2-cur/src: bookmark.o Only in dillo2-cur/src: bw.o Only in dillo2-cur/src: cache.o Only in dillo2-cur/src: capi.o Only in dillo2-cur/src: chain.o Only in dillo2-cur/src: colors.o Only in dillo2-cur/src: cookies.o Only in dillo2-cur/src: dialog.o Only in dillo2-cur/src: dicache.o Only in dillo2-cur/src: dillo-fltk Only in dillo2-cur/src: dillo.o Only in dillo2-cur/src: dir.o Only in dillo2-cur/src: dns.o Only in dillo2-cur/src: dpiapi.o Only in dillo2-cur/src: form.o Only in dillo2-cur/src: gif.o Only in dillo2-cur/src: history.o diff -pru -- dillo2/src/html.cc dillo2-cur/src/html.cc --- dillo2/src/html.cc 2007-11-03 22:16:26.000000000 +0000 +++ dillo2-cur/src/html.cc 2007-11-04 08:44:02.000000000 +0000 @@ -814,7 +814,7 @@ typedef struct { int isocode; } Ent_t; -#define NumEnt 252 +#define NumEnt 253 static const Ent_t Entities[NumEnt] = { {"AElig",0306}, {"Aacute",0301}, {"Acirc",0302}, {"Agrave",0300}, {"Alpha",01621},{"Aring",0305}, {"Atilde",0303}, {"Auml",0304}, @@ -833,7 +833,9 @@ static const Ent_t Entities[NumEnt] = { {"Yacute",0335},{"Yuml",0570}, {"Zeta",01626}, {"aacute",0341}, {"acirc",0342}, {"acute",0264}, {"aelig",0346}, {"agrave",0340}, {"alefsym",020465},{"alpha",01661},{"amp",38}, {"and",021047}, - {"ang",021040}, {"aring",0345}, {"asymp",021110},{"atilde",0343}, + {"ang",021040}, + {"apos",39}, /* not in HTML 4.01 */ + {"aring",0345}, {"asymp",021110},{"atilde",0343}, {"auml",0344}, {"bdquo",020036},{"beta",01662}, {"brvbar",0246}, {"bull",020042},{"cap",021051}, {"ccedil",0347}, {"cedil",0270}, {"cent",0242}, {"chi",01707}, {"circ",01306}, {"clubs",023143}, Only in dillo2-cur/src: html.o Only in dillo2-cur/src: image.o Only in dillo2-cur/src: jpeg.o Only in dillo2-cur/src: klist.o Only in dillo2-cur/src: menu.o Only in dillo2-cur/src: misc.o Only in dillo2-cur/src: nav.o Only in dillo2-cur/src: plain.o Only in dillo2-cur/src: png.o Only in dillo2-cur/src: prefs.o Only in dillo2-cur/src: timeout.o Only in dillo2-cur/src: ui.o Only in dillo2-cur/src: uicmd.o Only in dillo2-cur/src: url.o Only in dillo2-cur/src: web.o Only in dillo2-cur: stamp-h1

Nov. 4, 2007

Quick question about font_factor in dillorc2

by onepoint＠starurchin.org

Just out of interest, is there any reason why the default dillorc2 sets font_factor to 1.5? It seems unreasonably large to me, and I personally much prefer to comment it out. Anyone else have feelings one way or the other? Regards, Jeremy Henty

Nov. 4, 2007

Re: image loading from menus

by place＠gobigwest.com

Jorge wrote: > After some study and a reimplentation of plain.cc with classes > (getting rid of its linkblock and hooking a callback for memory > handling), it looks simpler and better. It looks like the same > may be done with html.cc. I was just wondering how you weigh the options of adding a callback vs. adding a signal. I had experimented with a destroy signal in my tree recently. It seemed more consistent to me, although I suspect the various child widgets were wasting cycles on signal emission code when they didn't have any receivers...

Nov. 4, 2007

Re: image loading from menus

by jcid＠dillo.org

Hi, On Fri, Nov 02, 2007 at 02:32:49PM +0000, place wrote: > > > Now I have to send a linkblock ptr around. > > > > Where, what are you working on? > > Please give me more background. > > The right-click menus. I am giving them the linkblock ptr via > a_UIcmd_*_popup so that they can call Html_load_images > (well, a_Html_load_images now, by way of a_UIcmd_load_images -- > if that's the right way to route it.) After some study and a reimplentation of plain.cc with classes (getting rid of its linkblock and hooking a callback for memory handling), it looks simpler and better. It looks like the same may be done with html.cc. Please check out the CVS. Comments are welcomed. Although the patch is already in CVS, I'm attaching the dw2 patch text so Sebastian can review it easily when he has time. -- Cheers Jorge.-

Nov. 3, 2007

Re: patch: images: working menu, deferred loading; page menu works

by jcid＠dillo.org

On Fri, Nov 02, 2007 at 07:45:01PM +0000, place wrote: > > BTW, the "reload" button does not reload the current page, but the > > most recent one in the history list. This is not the same if you have > > gone back one or more pages. > > I've attached a small patch that appears to work for me. > This bug is just one in a set, but it has a simple solution: diff -pru dillo2/src/nav.c dillo2-cur/src/nav.c --- dillo2/src/nav.c 2007-10-25 17:18:04.000000000 -0300 +++ dillo2-cur/src/nav.c 2007-11-02 19:14:03.000000000 -0300 @@ -70,7 +70,7 @@ int a_Nav_get_top_uidx(BrowserWindow *bw { nav_stack_item *nsi; - nsi = dList_nth_data (bw->nav_stack, a_Nav_stack_size(bw) - 1); + nsi = dList_nth_data (bw->nav_stack, a_Nav_stack_ptr(bw)); return (nsi) ? nsi->url_idx : -1; } Commited. ;) -- Cheers Jorge.-

Nov. 2, 2007

Another handy patch.

by jcid＠dillo.org

Hi, + Enabled vertical mouse wheel scrolling. -- Cheers Jorge.-

Nov. 2, 2007

Re: patch: images: working menu, deferred loading; page menu works

by place＠gobigwest.com

> BTW, the "reload" button does not reload the current page, but the > most recent one in the history list. This is not the same if you have > gone back one or more pages. I've attached a small patch that appears to work for me.

Nov. 2, 2007

Re: patch: images: working menu, deferred loading; page menu works

by matthias.franz＠ujf-grenoble.fr

On Thu, Nov 01, 2007 at 03:03:31PM -0300, Jorge Arellano Cid wrote: > On Thu, Nov 01, 2007 at 04:53:48PM +0000, place wrote: > > Are you saying that you would prefer that it be moved to the menu anyway? > > I've been using left click to load for years now, and I would find > > it significantly less convenient to use the menu for every image... > > The point is that it may be handy to follow an image-link without > loading the image (I'm not sure how useful though). I think it's definitely useful, for example when you have a slow internet connection. Moreover, text browsers link elinks show that browsing without images has its uses. BTW, the "reload" button does not reload the current page, but the most recent one in the history list. This is not the same if you have gone back one or more pages. Cheers, -- Matthias Franz

Nov. 2, 2007