- Dillo-dev - lists.mailman3.com

Re: [Dillo-dev] dillo-0.7.2 compile probs SuSE 8.1

by Victor Sahlstedt

On Tue, 29 Apr 2003 21:38:36 -0500 Patrick Shanahan <paka(a)MyRealBox.com> wrote: > Having successfully compiled 0.7.0, 0.7.1 & 0.7.12, I am unable to > compile 0.7.2. I get the following errors which mean very little to > me. Any help or direction will be appreciated. Perhaps outdated libpng? Get the latest stable at; http://sourceforge.net/project/showfiles.php?group_id=5624 - Victor

April 30, 2003

dillo-0.7.2 compile probs SuSE 8.1

by Patrick Shanahan

Having successfully compiled 0.7.0, 0.7.1 & 0.7.12, I am unable to compile 0.7.2. I get the following errors which mean very little to me. Any help or direction will be appreciated. checking png.h usability... no checking png.h presence... yes configure: WARNING: png.h: present but cannot be compiled configure: WARNING: png.h: check for missing prerequisite headers? configure: WARNING: png.h: proceeding with the preprocessor's result configure: WARNING: ## ------------------------------------ ## configure: WARNING: ## Report this to bug-autoconf(a)gnu.org. ## configure: WARNING: ## ------------------------------------ ## checking for png.h... yes checking for png_check_sig in -lpng... no checking for pthread_create in -lpthread... no checking for ANSI C header files... (cached) yes checking fcntl.h usability... no checking fcntl.h presence... yes configure: WARNING: fcntl.h: present but cannot be compiled configure: WARNING: fcntl.h: check for missing prerequisite headers? configure: WARNING: fcntl.h: proceeding with the preprocessor's result configure: WARNING: ## ------------------------------------ ## configure: WARNING: ## Report this to bug-autoconf(a)gnu.org. ## configure: WARNING: ## ------------------------------------ ## checking for fcntl.h... yes checking for unistd.h... (cached) yes checking sys/uio.h usability... no checking sys/uio.h presence... yes configure: WARNING: sys/uio.h: present but cannot be compiled configure: WARNING: sys/uio.h: check for missing prerequisite headers? configure: WARNING: sys/uio.h: proceeding with the preprocessor's result configure: WARNING: ## ------------------------------------ ## configure: WARNING: ## Report this to bug-autoconf(a)gnu.org. ## configure: WARNING: ## ------------------------------------ ## checking for sys/uio.h... yes checking for socket... no configure: creating ./config.status config.status: creating Makefile sed: file ./confstatT5yymW/subs-2.sed line 19: Unterminated ' command config.status: creating dpi/Makefile sed: file ./confstatT5yymW/subs-2.sed line 19: Unterminated ' command config.status: creating doc/Makefile sed: file ./confstatT5yymW/subs-2.sed line 19: Unterminated ' command config.status: creating src/Makefile sed: file ./confstatT5yymW/subs-2.sed line 19: Unterminated ' command config.status: creating src/IO/Makefile sed: file ./confstatT5yymW/subs-2.sed line 19: Unterminated ' command config.status: creating config.h config.status: executing depfiles commands And I end up with a Makefile of 0 bytes. System is SuSE 8.1. tks, -- Patrick Shanahan Please avoid TOFU and trim >quotes< http://wahoo.no-ip.org Registered Linux User #207535 icq#173753138 @ http://counter.li.org Linux, a continuous *learning* experience

April 30, 2003

Re: [Dillo-dev] Thank you..

by Victor Sahlstedt

On Tue, 29 Apr 2003 21:58:32 -0400 Bas van Besouw <bvanbesouw(a)wideopenwest.com> wrote: > It has probably been said a million times (however I could only find 1) but I like to thank all of you guys for writing Dillo. > > It is a great piece of work, that makes the internet actually look fast.:) > > Thanks again I second that :-) - Victor

April 30, 2003

Thank you..

by Bas van Besouw

Hi. It has probably been said a million times (however I could only find 1) but I like to thank all of you guys for writing Dillo. It is a great piece of work, that makes the internet actually look fast.:) Thanks again Bas

April 30, 2003

GUI library changing

by Isaque Galdino

Hi guys, I would like to know how hard would be changing the GUI library. I would like to change GTK+ stuff by FLTK and run dillo on my PDA that uses mainly FLTK library. Thnx. Isaque Galdino "Free as a penguim" - maybe The Beatles were trying to say that! :) --- UOL, o melhor da Internet http://www.uol.com.br/

April 30, 2003

Re: [Dillo-dev] dillo-0.7.2 release

by Kelson Vibber

On Tuesday 29 April 2003 03:24 pm, Jorge Arellano Cid wrote: > Dillo-0.7.2 is ready for download at the website. The > announcement is delayed one day or two to give enough time for > the .deb (and maybe a .rpm) to be already packaged when it goes > public. RPMs for Red Hat 7.3 and 9 are up at http://www.hyperborea.org/software/dillo.html -- Kelson Vibber www.hyperborea.org

April 29, 2003

dillo-0.7.2 release

by Jorge Arellano Cid

Hi there, Dillo-0.7.2 is ready for download at the website. The announcement is delayed one day or two to give enough time for the .deb (and maybe a .rpm) to be already packaged when it goes public. Enjoy Jorge.-

April 29, 2003

Re: [Dillo-dev] Dillo on the iPAQ

by john＠utzweb.net

# xhost + see what that gets you you could also do a man xhost and see what help that gives > I'm trying to install dillo on the iPAQ running > familiar linux. I've tried it on several versions > of familiar, and keep getting the same result: > # ./dillo > Setting locale to C > > Gtk-WARNING **: cannot open display: :0.0 > # > > any hints on how to fix this? > > Thanks, > Ivan > > > _______________________________________________ > Dillo-dev mailing list > Dillo-dev(a)lists.auriga.wearlab.de > http://lists.auriga.wearlab.de/cgi-bin/mailman/listinfo/dillo-dev > >

April 29, 2003

Re: [Dillo-dev] dillo-0.7.2 compile probs SuSE 8.1

by William Trenker

On Wed, 30 Apr 2003 05:16:24 +0200, Victor Sahlstedt <salan(a)myrealbox.com> wrote: > On Tue, 29 Apr 2003 21:38:36 -0500 > Patrick Shanahan <paka(a)MyRealBox.com> wrote: > >> Having successfully compiled 0.7.0, 0.7.1 & 0.7.12, I am unable to >> compile 0.7.2. I get the following errors which mean very little to >> me. Any help or direction will be appreciated. > > Perhaps outdated libpng? Get the latest stable at; > http://sourceforge.net/project/showfiles.php?group_id=5624 > <snip> I just downloaded the latest libpng and built it. I still get the following segfault when I link to a web page with a png file on it: # dillo Setting locale to C... dillo_dns_init: Here we go! Disabling cookies. Nav_open_url: Url=>about:splash< Nav_open_url: Url=>http://dillo.auriga.wearlab.de/ChangeLog.html< Dns_server [0]: dillo.auriga.wearlab.de is 0x80d3608 Connecting to 134.102.206.165 Png_error_handling: Application was compiled with png.h from libpng-1.0.14 Segmentation fault If the web page does not have a png file, dillo does not crash. Checking the dillo binary I get: # ldd /usr/bin/dillo libpthread.so.0 => /lib/libpthread.so.0 (0x40013000) libpng.so.2 => /usr/local/lib/libpng.so.2 (0x40028000) <snip> So my question is how does libpng-1.0.14 relate to libpng.so.2? Thanks for any help you can provide, Bill

April 29, 2003

Dillo on the iPAQ

by Ivan Gevirtz

I'm trying to install dillo on the iPAQ running familiar linux. I've tried it on several versions of familiar, and keep getting the same result: # ./dillo Setting locale to C Gtk-WARNING **: cannot open display: :0.0 # any hints on how to fix this? Thanks, Ivan

April 29, 2003

Re: [Dillo-dev] dillo-0.7.2 compile probs SuSE 8.1

by William Trenker

On Tue, 29 Apr 2003 21:38:36 -0500, Patrick Shanahan <paka(a)MyRealBox.com> wrote: > Having successfully compiled 0.7.0, 0.7.1 & 0.7.12, I am unable to > compile 0.7.2. I get the following errors which mean very little to > me. Any help or direction will be appreciated. > > checking png.h usability... no > checking png.h presence... yes > configure: WARNING: png.h: present but cannot be compiled > configure: WARNING: png.h: check for missing prerequisite headers? > configure: WARNING: png.h: proceeding with the preprocessor's result > configure: WARNING: ## ------------------------------------ ## > configure: WARNING: ## Report this to bug-autoconf(a)gnu.org. ## > configure: WARNING: ## ------------------------------------ ## <snip> Yes, I got that problem too when I tried to build Dillo 0.7.2 earlier today. Since it didn't work I downloaded the rpm (http://www.hyperborea.org/software/repackaged/dillo-0.7.2-1.rh7.i386.rpm) which installed but when I run Dillo it segfaults with a message saying the Dillo was built for a different version of libpng (which I've already reported on this list today). So there seems to be a common thread here. Bill -- Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/

April 29, 2003

Bad view page usiing overlib

by Diego Sáenz

I view this page bad (scroll down and see the table) http://www.riojaparty.com/riojaparty2003/modules.php?name=Content&pa=showpa… They are testing overlib, a javascript lib that is free software, so maybe the page changes fast. Before using that lib i saw the page well. The html of the table seems well to my brain, can be that dillo fail to ignore some javascript stuff or must i send a mail to the webmaster? Another question: What program do you use to test the html syntaxis of the pages? Thanks for this time. Diego.

April 29, 2003

0.7.2 installation problem

by William Trenker

If been using Dillo for several months now. I just downloaded http://www.hyperborea.org/software/repackaged/dillo-0.7.2-1.rh7.i386.rpm and installed it using this command: rpm -Uvh --nodeps dillo-0.7.2-1.rh7.i386.rpm (To be fair, I had to use the --nodeps option because without it rpm complains that I don't have libpng.so.2. But I've done other installs with this same arrangement and they work fine.) When I run dillo I get the Dillo 0.7.2 spash screen. I then click on the Change Log link and dillo loads the changelog page from the dillo web site. But when the graphic at the top of the page gets about 1/2 done dillo segfaults with this: # dillo Setting locale to C... dillo_dns_init: Here we go! Disabling cookies. Nav_open_url: Url=>about:splash< Nav_open_url: Url=>http://dillo.auriga.wearlab.de/ChangeLog.html< Dns_server [0]: dillo.auriga.wearlab.de is 0x80d33d8 Connecting to 134.102.206.165 Png_error_handling: Application was compiled with png.h from libpng-1.0.14 Segmentation fault I seem to have a problem with my libpng. Here is what I have installed: # l /usr/lib/libpng* lrwxrwxrwx 1 root root 10 Jan 8 13:59 /usr/lib/libpng.a - > libpng12.a lrwxrwxrwx 1 root root 19 Jan 8 14:42 /usr/lib/libpng.so - > libpng12.so.0.1.2.4* lrwxrwxrwx 1 root root 17 Jan 5 14:53 /usr/lib/libpng.so.2 -> libpng.so.3.1.2.1* lrwxrwxrwx 1 root root 17 Jan 8 16:10 /usr/lib/libpng.so.3 -> libpng.so.3.1.2.1* -rwxr-xr-x 1 root root 127212 Jun 10 2002 /usr/lib/libpng.so.3.1.2.1* lrwxrwxrwx 1 root root 19 Jan 8 14:42 /usr/lib/libpng.so.3.1.2.4 -> libpng12.so.0.1.2.4* -rw-r--r-- 1 root root 203154 Nov 3 00:00 /usr/lib/libpng12.a lrwxrwxrwx 1 root root 13 Jan 8 14:42 /usr/lib/libpng12.so -> libpng12.so.0* lrwxrwxrwx 1 root root 19 Jan 8 14:42 /usr/lib/libpng12.so.0 -> libpng12.so.0.1.2.4* -rwxr-xr-x 1 root root 176576 Aug 24 2002 /usr/lib/libpng12.so.0.1.2.4* Do you know where I can get a libpng that meets dillo's needs? By the way, Dillo 0.7.1 didn't have this problem, I use it all the time. Thanks for Dillo and any help you can provide, Bill -- Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/

April 29, 2003

Re: [Dillo-dev] CVS and Automake/Autoconf

by Joe Hill

On Mon, 28 Apr 2003 23:48:00+0200 Andreas Schweitzer<Andreas.Schweitzer(a)hs.uni-hamburg.de> wrote: > Hi, > > Since it probably won't help a lot, I only answer off-list. > > > I am running autoconf2.5-2.53 and automake1.6-1.63-1 on Mandrake > > 9.0. > > I'm compiling dillo with older and newer version of autoconf and > automake. So it *should* be ok. Yet, I do suspect some issues > with your version. > Thanks, you were right on the money. One thing, I tried to reply to you directly, but it looks like you have even more restrictive mail filters than I do ;), so I'll send this to the list just to mainly say thanks for the tip. I reinstalled automake and autoconf, this time with ./configure--prefix=/usr (seems I have to do that with everything on Mandrake), and all seems well. > > Any help appreciated, I am really excited about this project. I have > > .70 running now, but...I would like some of the bugfixes and newer > > featues. I tried installing dillo-0.7.1.2 from source, but it > > crashed every time I hit the bookmark icon. I'll keep on pushin' > > ahead if someone can give > > That should NOT be the case. You probably still have an old > bm_srv running. Exit all dillos, then do a "ps ax | egrep bm_srv" > and kill all matching processes. Then make sure, you have the latest > bm_srv12 (of dillo-0.7.1.2) in your $PATH, and no OLD bm_srv12 in > ~/.dillo/ > but why would that not happen with the rpm version that I installed? and it doesn't happen with the CVS ver I just built either. bah, it doesn't matter, anyway, right? It's workin now! Once again, thanks, this is one fantastic browser. The bookmark process isn't my favourite (no pun intended), but I think I saw on the list archives that might be changing? ++++ ..| ..\ /\ / ../ ..\ /\ / ../ ..| ..\/ ..\/ ../ ..\/ \/ ../ _| __/\__/ __/\__/ "The philosophers have only interpreted the world in various ways. The point however is to change it." -Chuck Marx > HTH, > Andreas > > -- > **************************** NEW ADDRESS > ****************************** Hamburger Sternwarte > Universitaet Hamburg > Gojenbergsweg 112 Tel. ++49 40 42891 > 4016 D-21029 Hamburg, Germany Fax. ++49 40 > 42891 4198

April 28, 2003

CVS and Automake/Autoconf

by Joe Hill

When I try to build from CVS, ./autogen.sh gives me the following errors: aclocal: configure.in: 50: macro `AM_PATH_GTK' not found in library autoheader: error: AC_CONFIG_HEADERS not found in configure.in configure.in:8: error: possibly undefined macro: AM_INIT_AUTOMAKE If this token and others are legitimate, please use m4_pattern_allow. See the Autoconf documentation. configure.in:9: error: possibly undefined macro: AM_CONFIG_HEADER configure.in:29: error: possibly undefined macro: AM_PROG_CC_STDC configure.in:50: error: possibly undefined macro: AM_PATH_GTK configure.in:8: no proper implementation of AM_INIT_AUTOMAKE was found, configure.in:8: probably because aclocal.m4 is missing... configure.in:8: You should run aclocal to create this file, then configure.in:8: run automake again. configure.in: installing `./install-sh' configure.in: installing `./mkinstalldirs' configure.in: installing `./missing' configure.in:6: installing `./config.guess' configure.in:6: installing `./config.sub' configure.in:9: required file `./config.h.in' not found dpi/Makefile.am: installing `./depcomp' /usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/lang-compile.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/lang-compile.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/lang-compile.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/lang-compile.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/lang-compile.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL/usr/local/share/automake-1.6/am/lang-compile.am: AMDEP does not appear in AM_CONDITIONAL a configure script is built, but of course it doesn't work: ./configure: line 1351: syntax error near unexpected token `dillo,' ./configure: line 1351: `AM_INIT_AUTOMAKE(dillo, 0.7.1)' I tried running aclocal as advised in the error message, however, more errors: aclocal: configure.in: 50: macro `AM_PATH_GTK' not found in library Since I have successfully built from CVS with other software using an autogen script, I am wondering if maybe this is just a version issue? I am running autoconf2.5-2.53 and automake1.6-1.63-1 on Mandrake 9.0. Any help appreciated, I am really excited about this project. I have .70 running now, but...I would like some of the bugfixes and newer featues. I tried installing dillo-0.7.1.2 from source, but it crashed every time I hit the bookmark icon. I'll keep on pushin' ahead if someone can give me a bit of a nudge about this autogen.sh thing. Thanks all, -J.

April 28, 2003

Re: Dillo-dev digest, Vol 1 #100 - 6 msgs

by linux seaq

> Onto another topic: > > A last minute leak fix (iterators) left dillo a bit unstable. > We knew the iterators had a leak and it's great to have found it. > Now it will take some extra time to try to fix the problem and, > after that, stress-test the Copy&Paste code again. > > We haven't got an accurate procedure to reproduce the crash > (yet). This explanation may help: > > <q> > I think I have found a relatively new segfault bug : > You need some sort of thumbtags like the screenshots on > http://www.incagold.com/products/Speed/speedthief.php > > Now you have to move down the mouse and at the same time click > on one image (i.e. click while still aiming for the image). > If the button-push occures outside the image and the button-release > inside the image, chances are high to get a segfault. This > is a bit hard to reproduce, but I couldn't get an older version > with selection to behave like that. It also happened a few > times to me last night (after the last 'cvs up') but never before. > </q> > HI i'm having that problem when the page is not entirely downloaded and i'm scrolling down the page and selecting some text, it segfaults.. it happens sometimes not always, so the conditions may vary.. i`m using the cvs from 06/04 when i used an earlier cvs update (just after select cvs commit) it ocurred*a lot more times* than now. --- Andres Mauricio Mujica SEAQ SERVICIOS CIA LTDA www.seaq.com.co

April 28, 2003

Re: [Dillo-dev] Copy&Paste bug BINGO!!!

by Jorge Arellano Cid

Hi there, > > [...] > > I believe the functions that may have side effects on their > > arguments should use an API that somehow reflects that. For > > instance: > > > > DwExtIterator* a_Dw_ext_iterator_new (DwIterator *it); > > Should be > > DwExtIterator* a_Dw_ext_iterator_new (DwIterator **it); > ^^ Yeah, just a typo. > Another possibility would be to define wrapper macros: > > #define a_Dw_ext_iterator_new(it) _a_Dw_ext_iterator_new(&it) > > DwExtIterator* _a_Dw_ext_iterator_new (DwIterator **it); > > but that's a bit kludgy for my taste. Yep. > > The same applies for a_Dw_iterator_free() and friends. > > It should be quite obvious that functions ending on "_free", > "_destroy" etc. have a side-effect of their argument, so I'd keep this > and regard it as an exception. a_Dw_ext_iterator_new is another case, > and very confusing, so this one should be fixed. Hem, that's what I meant (rush writing). > > BTW, the > > first approach I wrote was to force a_Dw_iterator_free to make > > NULL its argument). > > Very useful for debugging! :-) That was the idea! > > Well, that's it. As I don't know the iterators code as well as > > Sebastian, I'll let him make the decisions. > > Well, first I'd use this change for the release, and fix it for > 0.7.3. A grep for "a_Dw_ext_iterator_new" shows that this is handled > correctly elsewhere. OK guys, please give Copy&Paste some hard testing. If it shows no problems, I'll probably make the final touches this Monday and try to make a release on Tuesday. > Perhaps it is cleaner to remove the side-effect completely, That sounds good and clean. > by letting > a_Dw_ext_iterator_new putting a clone on top of the stack, so that the > callers must free the simple iterator. This would be a bit slower, but > compared to that, a_Dw_ext_iterator_new creates anyway quite a couple > of DwIterator's. Has to be tested. Yes, probably 0.7.3 will require some profiling and tunning. > (Thorben: As far as I can say from a first gprof run, this is not the > critical code, so the problem is not made worse.) This problem certainly is worth an investigation. Let it be on 0.7.3. Cheers Jorge.-

April 26, 2003

Re: [Dillo-dev] Copy&Paste bug BINGO!!!

by Sebastian Geerken

On Fri, Apr 25, Jorge Arellano Cid wrote: > After some hard code revisions, I decided to "eat" Valgrind's > manual, install it and BINGO!: > > [...] > > What does this mean? > > As a_Dw_ext_iterator_new(it) may _free_ 'it', 'it' should no > longer be used after that call. That was the case with the page > (right-click test case). > > I commited a short _code_change_ that avoids the problem. It is > not a "patch" but just illustrative code. > > I believe the functions that may have side effects on their > arguments should use an API that somehow reflects that. For > instance: > > DwExtIterator* a_Dw_ext_iterator_new (DwIterator *it); Should be DwExtIterator* a_Dw_ext_iterator_new (DwIterator **it); ^^ > and be used like: > > a_Dw_ext_iterator_new(&it); > > so 'it' can be NULL at return time. Another possibility would be to define wrapper macros: #define a_Dw_ext_iterator_new(it) _a_Dw_ext_iterator_new(&it) DwExtIterator* _a_Dw_ext_iterator_new (DwIterator **it); but that's a bit kludgy for my taste. > The same applies for a_Dw_iterator_free() and friends. It should be quite obvious that functions ending on "_free", "_destroy" etc. have a side-effect of their argument, so I'd keep this and regard it as an exception. a_Dw_ext_iterator_new is another case, and very confusing, so this one should be fixed. > BTW, the > first approach I wrote was to force a_Dw_iterator_free to make > NULL its argument). Very useful for debugging! :-) > Well, that's it. As I don't know the iterators code as well as > Sebastian, I'll let him make the decisions. Well, first I'd use this change for the release, and fix it for 0.7.3. A grep for "a_Dw_ext_iterator_new" shows that this is handled correctly elsewhere. Perhaps it is cleaner to remove the side-effect completely, by letting a_Dw_ext_iterator_new putting a clone on top of the stack, so that the callers must free the simple iterator. This would be a bit slower, but compared to that, a_Dw_ext_iterator_new creates anyway quite a couple of DwIterator's. Has to be tested. (Thorben: As far as I can say from a first gprof run, this is not the critical code, so the problem is not made worse.) Sebastian

April 26, 2003

Copy&Paste bug BINGO!!!

by Jorge Arellano Cid

Hi there, After some hard code revisions, I decided to "eat" Valgrind's manual, install it and BINGO!: <q> ==6797== ==6797== Invalid memory access of size 4 ==6797== at 0x80662B3: a_Selection_button_press (selection.c:142) ==6797== by 0x805EB8F: Dw_page_send_selection_event (dw_page.c:1486) ==6797== by 0x805EE47: Dw_page_button_press (dw_page.c:1499) ==6797== by 0x806474D: Dw_widget_mouse_event (dw_widget.c:740) ==6797== Address 0x40D714F0 is 0 bytes inside a block of size 52 free'd ==6797== at 0x40148050: free (in /usr/local/lib/valgrind/valgrind.so) ==6797== by 0x403C1FE8: g_free (in /usr/lib/libglib-1.2.so.0.0.10) ==6797== by 0x8057CA1: a_Dw_ext_iterator_new (dw_ext_iterator.c:171) ==6797== by 0x8066286: a_Selection_button_press (selection.c:136) </q> What does this mean? As a_Dw_ext_iterator_new(it) may _free_ 'it', 'it' should no longer be used after that call. That was the case with the page (right-click test case). I commited a short _code_change_ that avoids the problem. It is not a "patch" but just illustrative code. I believe the functions that may have side effects on their arguments should use an API that somehow reflects that. For instance: DwExtIterator* a_Dw_ext_iterator_new (DwIterator *it); and be used like: a_Dw_ext_iterator_new(&it); so 'it' can be NULL at return time. The same applies for a_Dw_iterator_free() and friends. BTW, the first approach I wrote was to force a_Dw_iterator_free to make NULL its argument). Well, that's it. As I don't know the iterators code as well as Sebastian, I'll let him make the decisions. Hope this helps Happy :) Jorge.-

April 25, 2003

Re: [Dillo-dev] Bookmarks

by Andreas Schweitzer

Hi > > Are there any plans to reintroduce a bookmarks menu? Any patches for it? > > Not really. Although dpis will have a way of setting menu items > someday, it's very unlikely it will be enough to make a good > implementation from it. I think, if the dpi sticks to one bookmark file syntax, dillo should be able to read it - maybe even via a dpi call - and then *only* display it when - let's say - clicking the right mouse button on the bookmark icon. > It is possible to make a dpi that uses a plain GTK+ interface > for bookmarks (think of a dpi as an external program that > communicates with dillo using the dpi protocol), but this > probably would be an overkill for your problem. Actually, in my quest to learn a bit of GTK, I modified the current bookmark server, so that it accepts one more command, which then will pop up a GTK-Gui. So far it only pops up some parts of the GTK tutorial :-) .... but it is kind of cool. The bm server itself returns a webpage that has a link which will fire up the GUI (just like the "modify" link). Alternatively, I modified dillo to directly fire up the bm-gui by pressing the middle mouse button on the bookmark icon (hey there are all three buttons used now ;-) ...) Cheers, Andreas -- **************************** NEW ADDRESS ****************************** Hamburger Sternwarte Universitaet Hamburg Gojenbergsweg 112 Tel. ++49 40 42891 4016 D-21029 Hamburg, Germany Fax. ++49 40 42891 4198

April 25, 2003

Re: [Dillo-dev] Bookmarks

by Victor Sahlstedt

On Tue, 22 Apr 2003 16:45:25 -0500 Lars Clausen <lrclause(a)cs.uiuc.edu> wrote: > While the page looks nifty for the editing, I really don't like getting an > extra bookmark page into my history stack every time I want to go to a > bookmark. Agreed. The way Dillo handle bookmarks nowadays is such a hassle that I usually end up just typing the url (whenever I can remember it). Editing bookmarks is quite unconvenient too, unfortunately. - Victor

April 25, 2003

Re: [Dillo-dev] Delayed release

by Phil Pennock

On 2003-04-24 at 20:31 +0200, Andreas Schweitzer wrote: > Here is an alternative method : simply RIGHT-click > on the thumbnails in said page : > http://www.incagold.com/products/Speed/speedthief.php Ah, that did it, I was finally able to reproduce. > Curiously, on FreeBSD I "only" get a > Gtk-WARNING **: gtk_signal_emit_by_name(): could not find signal "link_pressed" in the `DwTable' class ancestry > a second attempt to right click the image will pop up > the menu. FreeBSD 4.7, x86, dumped core on sigsegv. Stack backtrace: -----------------------------< cut here >------------------------------- #0 0x283df1d6 in strncpy () from /usr/lib/libc_r.so.4 #1 0x28273f0e in g_strndup () from /usr/local/lib/libglib12.so.3 #2 0x80671e9 in Selection_copy (selection=0x80eb560) at selection.c:379 #3 0x8066dfc in a_Selection_button_release (selection=0x80eb560, it=0x8281500, char_pos=0, link=39, event=0x80f6018, within_content=1) at selection.c:202 #4 0x805f916 in Dw_page_send_selection_event (page=0x8123e00, fn=0x8066cb8 <a_Selection_button_release>, x=178, y=320, event=0x80f6018) at dw_page.c:1486 #5 0x805f98b in Dw_page_button_release (widget=0x8123e00, x=178, y=320, event=0x80f6018) at dw_page.c:1513 #6 0x80651c3 in Dw_widget_mouse_event (widget=0x811b700, viewwidget=0x80c0b00, x=283, y=1175, event=0x80f6018) at dw_widget.c:740 #7 0x805b86c in Dw_gtk_viewport_mouse_event (widget=0x80c0b00, x=283, y=330, event=0x80f6018) at dw_gtk_viewport.c:358 #8 0x805b940 in Dw_gtk_viewport_button_release (widget=0x80c0b00, event=0x80f6018) at dw_gtk_viewport.c:389 #9 0x2818ce93 in gtk_marshal_BOOL__POINTER () from /usr/X11R6/lib/libgtk12.so.2 #10 0x281bced6 in gtk_signal_real_emit () from /usr/X11R6/lib/libgtk12.so.2 #11 0x281baef3 in gtk_signal_emit () from /usr/X11R6/lib/libgtk12.so.2 #12 0x281f2f6c in gtk_widget_event () from /usr/X11R6/lib/libgtk12.so.2 #13 0x2818cde5 in gtk_propagate_event () from /usr/X11R6/lib/libgtk12.so.2 #14 0x2818bf56 in gtk_main_do_event () from /usr/X11R6/lib/libgtk12.so.2 #15 0x2823c87c in gdk_event_dispatch () from /usr/X11R6/lib/libgdk12.so.2 #16 0x2826acf3 in g_main_dispatch () from /usr/local/lib/libglib12.so.3 #17 0x2826b31c in g_main_iterate () from /usr/local/lib/libglib12.so.3 #18 0x2826b4b4 in g_main_run () from /usr/local/lib/libglib12.so.3 #19 0x2818b813 in gtk_main () from /usr/X11R6/lib/libgtk12.so.2 #20 0x8067e25 in main () #21 0x8053269 in _start () -----------------------------< cut here >------------------------------- At level #2, g_strndup() is called: 379 else if (cmp == 0) 380 /* the end */ 381 txt = g_strndup(si->content.data.text, 382 b_char); (gdb) print si->content.data.text $4 = 0x81265b0 "3D" (gdb) print b_char $5 = 0 Are you supposed to be able to dup zero-length of a string?

April 25, 2003

Re: [Dillo-dev] Cookie locks?

by Lars Clausen

On Fri, 25 Apr 2003, Jorge Arellano Cid wrote: > > > Lars, > >> >> I understand unrestricted access to cookie jars can be bad for your >> health, but it's annoying that you can't have two Dillos using cookies >> at the same time. I usually have a Dillo running permanently, and when >> I get a URL in my mail, I can click on it to open another Dillo >> (quickly!) -- but that Dillo can't use cookies, which is frequently >> (e.g. in bugzilla) necessary. Could we do a write-lock on it? > > Actually, we plan (with Jörgen) to make of cookies a dpi server. Now *that* would be a good use of the dpi server. Shouldn't be horribly hard, and then the plugin can have a nice admin interface if it wants to. -Lars is still miffed about the bookmarks menu. -- Lars Clausen (http://shasta.cs.uiuc.edu/~lrclause)| Hårdgrim of Numenor "I do not agree with a word that you say, but I |---------------------------- will defend to the death your right to say it." | Where are we going, and --Evelyn Beatrice Hall paraphrasing Voltaire | what's with the handbasket?

April 25, 2003

Re: [Dillo-dev] Cookie locks?

by Jorge Arellano Cid

Lars, > > I understand unrestricted access to cookie jars can be bad for your health, > but it's annoying that you can't have two Dillos using cookies at the same > time. I usually have a Dillo running permanently, and when I get a URL in > my mail, I can click on it to open another Dillo (quickly!) -- but that > Dillo can't use cookies, which is frequently (e.g. in bugzilla) necessary. > Could we do a write-lock on it? Actually, we plan (with Jörgen) to make of cookies a dpi server. Cheers Jorge.-

April 25, 2003

Re: [Dillo-dev] Bookmarks

by Jorge Arellano Cid

On Tue, 22 Apr 2003, Lars Clausen wrote: > > Are there any plans to reintroduce a bookmarks menu? Any patches for it? Not really. Although dpis will have a way of setting menu items someday, it's very unlikely it will be enough to make a good implementation from it. It is possible to make a dpi that uses a plain GTK+ interface for bookmarks (think of a dpi as an external program that communicates with dillo using the dpi protocol), but this probably would be an overkill for your problem. > While the page looks nifty for the editing, I really don't like getting an > extra bookmark page into my history stack every time I want to go to a > bookmark. One way to solve this is to make dpi:/bm/ URIs "phantom URIs". That is, to flag them and tell the stack manager to get rid of them once they're no longer displayed. This needs some thought though, because you may have more than one dillo window. > In other news: Yay! Cut-and-paste! One of my main reasons to use other > browsers is now gone! Thanks! :) Cheers Jorge.-

April 25, 2003

Re: [Dillo-dev] Delayed release

by Chris Palmer

Andreas Schweitzer writes: > > We haven't got an accurate procedure to reproduce the crash > > (yet). This explanation may help: > > Here is an alternative method : simply RIGHT-click on the thumbnails > in said page : http://www.incagold.com/products/Speed/speedthief.php > > Curiously, on FreeBSD I "only" get a Gtk-WARNING **: > gtk_signal_emit_by_name(): could not find signal "link_pressed" in the > `DwTable' class ancestry a second attempt to right click the image > will pop up the menu. For what it's worth, on OpenBSD with Dillo from CVS about a week old, I get no warnings and no crashes when I try this test. -- Chris music is what numbers feel like San Francisco, CA http://www.nodewarrior.org/chris

April 25, 2003

selection and performance

by Thorben Thuermer

One of dillo's main advantages is it's incredible speed and efficiency, but somehow the selection code doesn't seem to keep up with this. it seems to have some scalability problem that makes it use excessive amounts of cpu when selecting large parts of complex pages (tables), and slows dillo down to a crawl even on my p3/750 (which dillo usually runs lightning fast on). to reproduce the problem, load: http://users.auriga.wearlab.de/~r00t/table.html (a table of 21x21 cells), and then try to select all cells by dragging from the top left to the lower right. cpu usage grows immensely, and near the end it slows down to a crawl. the script used to create the table (to avoid duplicate work if you need other test-cases): { echo '<html><body><table border=1>';for a in $(seq 0 20);do echo '<tr>'; for b in $(seq 0 20);do echo '<td>'"$a-$b"'</td>';done;echo '</tr>';done; echo '</table></body></html>';}>table.html if this is a known limitation, i'm sorry to bug you with it, but it seemed worth reporting. Greetings, Thorben Thuermer

April 24, 2003

[PATCH] Pth library support for threads

by Patrice Mandin

Hello dillo coders, I made a simple patch to add pth support for threads for Dillo. So platforms that do not have the pthread library can also use it. I am not very clever regarding thread programming, it may needs a bit more of work. I tested it on very simple pages and it worked for me. -- Patrice Mandin WWW: http://membres.lycos.fr/pmandin/ Programmeur Linux, Atari Spécialité: Développement, jeux

April 24, 2003

Re: [Dillo-dev] Delayed release

by Andreas Schweitzer

> Onto another topic: > > A last minute leak fix (iterators) left dillo a bit unstable. > We knew the iterators had a leak and it's great to have found it. > Now it will take some extra time to try to fix the problem and, > after that, stress-test the Copy&Paste code again. > > We haven't got an accurate procedure to reproduce the crash > (yet). This explanation may help: Here is an alternative method : simply RIGHT-click on the thumbnails in said page : http://www.incagold.com/products/Speed/speedthief.php Curiously, on FreeBSD I "only" get a Gtk-WARNING **: gtk_signal_emit_by_name(): could not find signal "link_pressed" in the `DwTable' class ancestry a second attempt to right click the image will pop up the menu. Cheers, Andreas -- **************************** NEW ADDRESS ****************************** Hamburger Sternwarte Universitaet Hamburg Gojenbergsweg 112 Tel. ++49 40 42891 4016 D-21029 Hamburg, Germany Fax. ++49 40 42891 4198

April 24, 2003

Delayed release

by Jorge Arellano Cid

Hi there, I just commited the code that cleanups the bookmarks server. It works like this: <q> /* * If this is our first run, kill the (old) bookmarks server, * and create the '~/.dillo/.072' file to flag what we did. */ static void Dillo_check_bm_srv12(char *dir) [...] </q> Yes, it uses the "[b]m_srv12" trick and some other precautions. Please test it: the whole point is that when installing the new version of dillo (072), the old bookmarks server may keep running and survive the updated. The new code detects the first dillo-072 run and issues a kill command to "bm_srv12". Onto another topic: A last minute leak fix (iterators) left dillo a bit unstable. We knew the iterators had a leak and it's great to have found it. Now it will take some extra time to try to fix the problem and, after that, stress-test the Copy&Paste code again. We haven't got an accurate procedure to reproduce the crash (yet). This explanation may help: <q> I think I have found a relatively new segfault bug : You need some sort of thumbtags like the screenshots on http://www.incagold.com/products/Speed/speedthief.php Now you have to move down the mouse and at the same time click on one image (i.e. click while still aiming for the image). If the button-push occures outside the image and the button-release inside the image, chances are high to get a segfault. This is a bit hard to reproduce, but I couldn't get an older version with selection to behave like that. It also happened a few times to me last night (after the last 'cvs up') but never before. </q> We'll be working on it. Cheers Jorge.-

April 24, 2003

Re: [Dillo-dev] Cookie locks?

by Gabriel Kihlman

Hi, Lars Clausen <lrclause(a)cs.uiuc.edu> writes: > I understand unrestricted access to cookie jars can be bad for your health, > but it's annoying that you can't have two Dillos using cookies at the same > time. I usually have a Dillo running permanently, and when I get a URL in > my mail, I can click on it to open another Dillo (quickly!) -- but that > Dillo can't use cookies, which is frequently (e.g. in bugzilla) necessary. > Could we do a write-lock on it? That is something I would appreciate as well. I have been running with this ugly patch that disables locking for atleast 6-7 months without a problem: --- src/cookies.c.orig Sun Jan 12 01:25:20 2003 +++ src/cookies.c Mon Feb 24 09:43:37 2003 @@ -31,7 +31,6 @@ #include "list.h" #include "cookies.h" -#define DEBUG_LEVEL 8 #include "debug.h" /* The maximum length of a line in the cookie file */ @@ -140,7 +139,7 @@ void a_Cookies_init() char line[LINE_MAXLEN]; /* Default setting */ - disabled = TRUE; + disabled = FALSE; /* Read and parse the cookie control file (cookiesrc) */ if (Cookie_control_init() != 0) { @@ -158,6 +157,7 @@ void a_Cookies_init() return; } +#if 0 /* Try to get a lock from the file descriptor */ disabled = (lockf(fileno(file_stream), F_TLOCK, 0) == -1); if (disabled) { @@ -166,7 +166,7 @@ void a_Cookies_init() fclose(file_stream); return; } - +#endif DEBUG_MSG(10, "Enabling cookies as from cookiesrc...\n"); cookies = g_hash_table_new(g_str_hash, g_str_equal);

April 24, 2003

Re: [Dillo-dev] Vera

by Nicola Girardi

> > Yes, I'm using it now. > > I can use it too, but it's not anti-aliased. xterm is able to use > anti-aliased TTF fonts (seemingly ONLY Luxi, but that's another issue), > so I know my X setup can use it. What do you have to do to get > anti-aliasing? Apply my patch, look in the archives. Should apply with minor tweaks. (Again, I hope this creeps in the official version.)

April 23, 2003

Re: [Dillo-dev] Vera works now

by Andreas Zwinkau

> Sounds like the "FontPath" settings in your XF86Config file do not > match the "catalogue" settings in your xfs config > (<X11etcdir>/fs/config). xfs is simply a decent cache, which can > provide fonts over the network(only need to install on one machine, > etc etc). Update your server config to use more font paths. You can > augment the list at run time with "xset fp+ /path/to/more/fonts" if > you don't want to touch XF86Config. > > One thing is to make sure that you run "xset fp rehash" after adding > new fonts and running "mkfontdir", since X doesn't automatically see > new fonts. I should have said that before, but only gave "xfstt > --sync". Sorry. I thought it was a xfs problem, so i did run it. But now i checked my X output and it says "Could not init font path element unix/:7100, removing from list!". So xfs is not used! I think the problem was mkfontdir. I thought fc-cache (gentoo) would do this, but it does not seem so. Xfs did mkfontdir or something similar on all directories and so Dillo can find them now. Thats my assumption. I checked .../fs/config and it is there. Thanks for help :D -- Andreas Zwinkau | web: andi.dasstellenwirinsinternet.de | mail: andi(a)buxach.de | jabber: beza1e1(a)amessage.de

April 23, 2003

Re: [Dillo-dev] Vera

by Phil Pennock

On 2003-04-22 at 18:25 +0200, Andreas Zwinkau wrote: > I did not use xfs, was that the problem? Mozilla and xchat support Vera > nevertheless. Now i run xfs Dillo and xlsfonts do find Vera, too. Sounds like the "FontPath" settings in your XF86Config file do not match the "catalogue" settings in your xfs config (<X11etcdir>/fs/config). xfs is simply a decent cache, which can provide fonts over the network (only need to install on one machine, etc etc). Update your server config to use more font paths. You can augment the list at run time with "xset fp+ /path/to/more/fonts" if you don't want to touch XF86Config. One thing is to make sure that you run "xset fp rehash" after adding new fonts and running "mkfontdir", since X doesn't automatically see new fonts. I should have said that before, but only gave "xfstt --sync". Sorry. So it could simply be fixed by having restarted X, since that effectively did a rehash. ;^) -Phil

April 23, 2003

Re: [Dillo-dev] Unix domain sockets

by Phil Pennock

On 2003-04-22 at 16:20 -0400, Jorge Arellano Cid wrote: > kill `ps -U <USER> | grep bm_srv12 | sed 's/^ *//g' | cut -d' ' -f1` Depending upon line buffering and how your PIDs are ordered (wrapped around recently?) that can end up killing the "grep" process before its given the PID for the bookmark server. A neat little Unix shell idiom for that is: grep '[b]m_srv12' The pattern there will no longer match the grep command-line, since the ']' adds an extra character. I think that I first learnt this from Aeleen Frisch's Essential System Adminstration book -- an excellent resource now in its third edition (was 1st ed when I learnt that one -- it's an old idiom). ps -U <USER> | grep '[b]m_srv' | awk '{print $1}' | xargs kill > That is, issued from dillo to force the shut down of a running > bookmark server. Suggestion: put it into a two-line shell-script which Dillo invokes; then portability simply becomes a package-maintainer issue of changing a very short shell script. > The point is: Does this command work well on *BSDs, and > Solaris? The basic stuff tested successfully on FreeBSD 4.7 ia32, OpenBSD 3.2 sparc and Solaris 2.5.1 sparc. Both the sed/cut version and the awk version. ;^) That's using the default /usr/bin/ps on Solaris; if someone has put the *BSD commands in their path first, then they'll have the really ancient syntax of /usr/ucb/ps and -U isn't supported. Such people are probably best just deemed "on their own".

April 23, 2003

Re: [Dillo-dev] Unix domain sockets

by Andreas Schweitzer

On Wed, Apr 23, 2003 at 11:35:16AM +0200, Lars Segerlund wrote: > > This is actually a newbie error on solaris :-) ... killall does kill > ALL ! same as HPUX, think AIX is the same. Yep, AIX is just as "destructive" :-) > >>> kill `ps -U <USER> | grep bm_srv12 | sed 's/^ *//g' | cut -d' ' -f1` Oh, and while being with AIX, this will NOT work on AIX. The first field on AIX is the UID ... But dillo doesn't compile easily on AIX anyways :-). Back when I was working on an AIX machine I did have an executable, but there were always some gotchas. Cheers, Andreas -- **************************** NEW ADDRESS ****************************** Hamburger Sternwarte Universitaet Hamburg Gojenbergsweg 112 Tel. ++49 40 42891 4016 D-21029 Hamburg, Germany Fax. ++49 40 42891 4198

April 23, 2003

Re: [Dillo-dev] Unix domain sockets

by Lars Segerlund

This is actually a newbie error on solaris :-) ... killall does kill ALL ! same as HPUX, think AIX is the same. / lars Segerlund Lars Clausen wrote: > On Tue, 22 Apr 2003, Andreas Schweitzer wrote: > >>Hi, >> >> >>> I kept cosidering the alternatives, and apart from the UDS idea >>>(not feasible), the incomplete "bye" command, there's also >>>another option! >>> >>> kill `ps -U <USER> | grep bm_srv12 | sed 's/^ *//g' | cut -d' ' -f1` >> >>On OpenBSD and FreeBSD, this construct does return the PID >>indeed. I'm just not sure how portable that is in general. >> >>Quite less portable is killall (which does not even exist on OpenBSD), >>but on FreeBSD it has an -u for user option. I.e. "killall -u andy >>bm_srv12" will do it, but only on FreeBSD :-( > > > Don't ever, *ever* consider killall portable. Under at least HPUX, killall > silently ignores its arguments and kills all the processes it can get its > murderous little hands on. A bad day to be root... > > -Lars >

April 23, 2003

Re: [Dillo-dev] Some thoughts

by Lars Segerlund

I agree, I have been trying to 'break out' the bookmark dpi into a loadable object file, ( loads at runtime, and installs the button and such ), in order to do work on a dpi plugin myself. If something about loading dpi modules were in place it would be much nicer. I have tried to hack a list of dpi modules in .dillo/dpis which is loaded at runtime, but so far not much sucess. Is there a plan on how these things will work in the future ? Because then one can progress with the dpi itself and the infrastructure. / Lars Segerlund. Andreas Schweitzer wrote: > Hi, > > Just a comment on one of your plans : > > >> Appart from that, I still have to finish the design of dpi >>initialization, addition, managing etc. (needed to support >>several dpis at the same time). > > > I think this will be important, since it kept me somewhat from > playing with dpi's. I did some experiments, but an automatic > dpi initilization makes coding much easier. And I have the feeling > others think the same :-) > > Something like > dpi:/foo/ > starts the dpi foo[possible-extra-name-part] from some location like > $PATH or some other "repository". And > dpi:/foo/command > passes "command" to the dpi foo. > Just some rambling :-) > > Cheers, > Andreas >

April 23, 2003

Re: [Dillo-dev] Vera

by Chris Palmer

Phil Pennock writes: > > Example: http://www.bitstream.com/categories/products/fonts/vera/ > > Is it possible to use it in Dillo? > > Yes, I'm using it now. I can use it too, but it's not anti-aliased. xterm is able to use anti-aliased TTF fonts (seemingly ONLY Luxi, but that's another issue), so I know my X setup can use it. What do you have to do to get anti-aliasing? -- Chris music is what numbers feel like San Francisco, CA http://www.nodewarrior.org/chris

April 23, 2003

Re: [Dillo-dev] Unix domain sockets

by higuita

On Tue, 22 Apr 2003 18:55:18 -0300, Livio Baldini Soares <livio(a)ime.usp.br> wrote: > > kill `ps -U <USER> | grep bm_srv12 | sed 's/^ *//g' | cut -d' ' -f1` > This works fine on the Solaris I have access to... it also works in irix 6.5 (although i still didnt try to compile dilo there, but i want to 8) higuita -- Naturally the common people don't want war... but after all it is the leaders of a country who determine the policy, and it is always a simple matter to drag the people along, whether it is a democracy, or a fascist dictatorship, or a parliament, or a communist dictatorship. Voice or no voice, the people can always be brought to the bidding of the leaders. That is easy. All you have to do is tell them they are being attacked, and denounce the pacifists for lack of patriotism and exposing the country to danger. It works the same in every country. -- Hermann Goering, Nazi and war criminal, 1883-1946

April 22, 2003

External viewers patch

by Lars Clausen

Attached is an updated version of my old external viewers patch. It's pretty rough, in that it'll just pick the first thing it finds in /etc/mailcap for a viewer, but I've used it for a long time now and it's made Dillo much more usable for me. -Lars -- Lars Clausen (http://shasta.cs.uiuc.edu/~lrclause)| Hårdgrim of Numenor "I do not agree with a word that you say, but I |---------------------------- will defend to the death your right to say it." | Where are we going, and --Evelyn Beatrice Hall paraphrasing Voltaire | what's with the handbasket?

April 22, 2003

Re: [Dillo-dev] Unix domain sockets

by Livio Baldini Soares

Hey Patrick Patrick Shanahan writes: > * Jorge Arellano Cid <jcid(a)softhome.net> [04-22-03 15:47]: > > > > I kept cosidering the alternatives, and apart from the UDS idea > > (not feasible), the incomplete "bye" command, there's also > > another option! > > > > kill `ps -U <USER> | grep bm_srv12 | sed 's/^ *//g' | cut -d' ' -f1` This works fine on the Solaris I have access to... > BUT, also and cleaner, > kill `pidof bm_srv12` > also works, but I do not know about other ?nix versions. ... but this does not. Sorry. best regards, -- Livio B. Soares

April 22, 2003

Cookie locks?

by Lars Clausen

I understand unrestricted access to cookie jars can be bad for your health, but it's annoying that you can't have two Dillos using cookies at the same time. I usually have a Dillo running permanently, and when I get a URL in my mail, I can click on it to open another Dillo (quickly!) -- but that Dillo can't use cookies, which is frequently (e.g. in bugzilla) necessary. Could we do a write-lock on it? -Lars -- Lars Clausen (http://shasta.cs.uiuc.edu/~lrclause)| Hårdgrim of Numenor "I do not agree with a word that you say, but I |---------------------------- will defend to the death your right to say it." | Where are we going, and --Evelyn Beatrice Hall paraphrasing Voltaire | what's with the handbasket?

April 22, 2003

Re: [Dillo-dev] Unix domain sockets

by Patrick Shanahan

* Jorge Arellano Cid <jcid(a)softhome.net> [04-22-03 15:47]: > > I kept cosidering the alternatives, and apart from the UDS idea > (not feasible), the incomplete "bye" command, there's also > another option! > > kill `ps -U <USER> | grep bm_srv12 | sed 's/^ *//g' | cut -d' ' -f1` > > That is, issued from dillo to force the shut down of a running > bookmark server. This works correctly on SuSE 8.1 uname -a: Linux wahoo 2.4.19-4GB #1 Wed Nov 27 00:56:40 UTC 2002 i686 unknown BUT, also and cleaner, kill `pidof bm_srv12` also works, but I do not know about other ?nix versions. sorry bout the repost -- Patrick Shanahan Please avoid TOFU and trim >quotes< http://wahoo.no-ip.org Registered Linux User #207535 icq#173753138 @ http://counter.li.org Linux, a continuous *learning* experience

April 22, 2003

Bookmarks

by Lars Clausen

Are there any plans to reintroduce a bookmarks menu? Any patches for it? While the page looks nifty for the editing, I really don't like getting an extra bookmark page into my history stack every time I want to go to a bookmark. In other news: Yay! Cut-and-paste! One of my main reasons to use other browsers is now gone! Thanks! Thanks, -Lars -- Lars Clausen (http://shasta.cs.uiuc.edu/~lrclause)| Hårdgrim of Numenor "I do not agree with a word that you say, but I |---------------------------- will defend to the death your right to say it." | Where are we going, and --Evelyn Beatrice Hall paraphrasing Voltaire | what's with the handbasket?

April 22, 2003

Re: [Dillo-dev] Unix domain sockets

by Lars Clausen

On Tue, 22 Apr 2003, Andreas Schweitzer wrote: > Hi, > >> I kept cosidering the alternatives, and apart from the UDS idea >> (not feasible), the incomplete "bye" command, there's also >> another option! >> >> kill `ps -U <USER> | grep bm_srv12 | sed 's/^ *//g' | cut -d' ' -f1` > > On OpenBSD and FreeBSD, this construct does return the PID > indeed. I'm just not sure how portable that is in general. > > Quite less portable is killall (which does not even exist on OpenBSD), > but on FreeBSD it has an -u for user option. I.e. "killall -u andy > bm_srv12" will do it, but only on FreeBSD :-( Don't ever, *ever* consider killall portable. Under at least HPUX, killall silently ignores its arguments and kills all the processes it can get its murderous little hands on. A bad day to be root... -Lars -- Lars Clausen (http://shasta.cs.uiuc.edu/~lrclause)| Hårdgrim of Numenor "I do not agree with a word that you say, but I |---------------------------- will defend to the death your right to say it." | Where are we going, and --Evelyn Beatrice Hall paraphrasing Voltaire | what's with the handbasket?

April 22, 2003

Re: [Dillo-dev] Unix domain sockets

by Patrick Shanahan

* Jorge Arellano Cid <jcid(a)softhome.net> [04-22-03 15:47]: > > I kept cosidering the alternatives, and apart from the UDS idea > (not feasible), the incomplete "bye" command, there's also > another option! > > kill `ps -U <USER> | grep bm_srv12 | sed 's/^ *//g' | cut -d' ' -f1` > > That is, issued from dillo to force the shut down of a running > bookmark server. This works correctly on SuSE 8.1 uname -a: Linux wahoo 2.4.19-4GB #1 Wed Nov 27 00:56:40 UTC 2002 i686 unknown -- Patrick Shanahan Please avoid TOFU and trim >quotes< http://wahoo.no-ip.org Registered Linux User #207535 icq#173753138 @ http://counter.li.org Linux, a continuous *learning* experience

April 22, 2003

Re: [Dillo-dev] Some thoughts

by Andreas Schweitzer

Hi, Just a comment on one of your plans : > Appart from that, I still have to finish the design of dpi > initialization, addition, managing etc. (needed to support > several dpis at the same time). I think this will be important, since it kept me somewhat from playing with dpi's. I did some experiments, but an automatic dpi initilization makes coding much easier. And I have the feeling others think the same :-) Something like dpi:/foo/ starts the dpi foo[possible-extra-name-part] from some location like $PATH or some other "repository". And dpi:/foo/command passes "command" to the dpi foo. Just some rambling :-) Cheers, Andreas -- **************************** NEW ADDRESS ****************************** Hamburger Sternwarte Universitaet Hamburg Gojenbergsweg 112 Tel. ++49 40 42891 4016 D-21029 Hamburg, Germany Fax. ++49 40 42891 4198

April 22, 2003

Re: [Dillo-dev] Unix domain sockets

by Andreas Schweitzer

Hi, > I kept cosidering the alternatives, and apart from the UDS idea > (not feasible), the incomplete "bye" command, there's also > another option! > > kill `ps -U <USER> | grep bm_srv12 | sed 's/^ *//g' | cut -d' ' -f1` On OpenBSD and FreeBSD, this construct does return the PID indeed. I'm just not sure how portable that is in general. Quite less portable is killall (which does not even exist on OpenBSD), but on FreeBSD it has an -u for user option. I.e. "killall -u andy bm_srv12" will do it, but only on FreeBSD :-( Cheers, Andreas -- **************************** NEW ADDRESS ****************************** Hamburger Sternwarte Universitaet Hamburg Gojenbergsweg 112 Tel. ++49 40 42891 4016 D-21029 Hamburg, Germany Fax. ++49 40 42891 4198

April 22, 2003

dillo-0.7.1 (fwd)

by Jorge Arellano Cid

---------- Forwarded message ---------- Date: Sat, 19 Apr 2003 12:40:31 -0300 (BRT) From: Frank Hrebabetzky <frankh(a)terra.com.br> To: jcid(a)softhome.net Subject: dillo-0.7.1 I folks, It is just to tell you that it is really a decent piece of software that you have written. It is lean, fast and free. For web browsing and writing of simple html texts, I use mozilla. But I was looking for a reader for documentation written in html, and loading mozilla takes far too long for that. So - congratulations. Should you ever decide to inflate dillo to a full fledged web browser, stuffed with all kind of gizmos, please keep offering a small, fast version such as the actual one. Cheers Frank ------------------------------------------------- Frank Hrebabetzky Tel.: +55 / 48 / 235 1106 Florianopolis +55 / 48 / 9998 7686 Brazil email: frankh(a)terra.com.br

April 22, 2003