August 2003 - Dillo-dev - lists.mailman3.com

Some points about the dpi's
by Andreas Schweitzer Aug. 22, 2003

Aug. 22, 2003

Hi, All right, I did it. I made my first dpi ! :-) (heavily based on the hello dpi, though). I will poilsh it a bit and release it soon. It can handle finger URLs. Two caveats : there is no standard for finger URL's, only some drafts http://www.watersprings.org/pub/id/draft-ietf-uri-url-finger-00.txt http://www.watersprings.org/pub/id/draft-ietf-uri-url-finger-01.txt http://www.watersprings.org/pub/id/draft-ietf-uri-url-finger-02.txt http://ftp.ics.uci.edu/pub/ietf/uri/draft-ietf-uri-url-finger-03.txt I implemented version 01 and 03 (if you check the documents you'll know why). However, I did find real finger URL's in the wild. The second caveat : dillo does not automatically call this plugin upon such a URL. It is easy to patch dillo, though. However, this brings me to my questions : 1) will there be a generic URL->dpi mapping ? I could imagine that dpid could actually handle that. I think, one would need to only modify IO/Url.c and capi.c to simply pass through EVERYTHING to the dpi framework. However, some "core URL handling dpi's" that will become part of the dillo package can probably stay hardcoded in dillo. 1b) Either way, I am planing on writing a "catch-all-URL handler" or "fallback handler" or "univeral URL handler" or the like. It will give the user the possibility to start an external program for handling certain URL's. In fact, with my current dpi, this is quite easy to do :-). 2) Does it make sense to write a library for all the dpi's ? I, for myself, did start simply collecting all the functions that I found in dpi/ and are used by many dpi's and put them in a library. Like the various send_stream functions or the get_attr_value function, or the encoding and decoding functions. 3) About configuration (and I think I read comments about this in the code), shouldn't there be only one config file ? and not dpidrc and possibly different configs for all the dpis ? maybe even merge cookiesrc with dillorc ? I guess some of this (or all of it ?) is being worked on by some of you. Maybe this mail helps to avoid duplication :-) Cheers, Andreas P.S.: I also have some ideas about improving the ftp plugin. There is a pretty simple ftp library, and I think it does make sense to actually log in an ftp server and stay logged in for all the operations. (unlike wget which logs in and out for every file). This is not only faster but helps with busy ftp sites, too ! -- **************************** NEW ADDRESS ****************************** Hamburger Sternwarte Universitaet Hamburg Gojenbergsweg 112 Tel. ++49 40 42891 4016 D-21029 Hamburg, Germany Fax. ++49 40 42891 4198

3 3

[PATCH] Tentative solution of #448, #449
by tomas＠fabula.de Aug. 20, 2003

Aug. 20, 2003

Hi, I've got a problem with 8 bit depth visuals (the images have the wrong palette). This is similar to bug #449 of your bug tracking database (thus I tentatively offered to fix it. Tentative patch is below. Besides, this looks very much like bug #448, so this might be related. I call the patch ``tentative'' because (a) I don't really know 100% what I'm doing here (b) I haven't tested this on other color depths yet. Nevertheless images show correctly now; the price I pay for this is Technicolor. Regards -- tomas ------------------------------------------------- diff -Naur dillo-0.7.3/src/dillo.c dillo-0.7.3-color/src/dillo.c --- dillo-0.7.3/src/dillo.c Tue Jul 22 22:51:22 2003 +++ dillo-0.7.3-color/src/dillo.c Fri Aug 8 19:01:32 2003 @@ -245,6 +245,9 @@ signal(SIGCHLD, Dillo_sigchld_handler); + gtk_widget_push_visual(gdk_rgb_get_visual()); + gtk_widget_push_colormap(gdk_rgb_get_cmap()); + /* check that ~/.dillo exists, create it if it doesn't */ dir = a_Misc_prepend_user_home(".dillo"); Dillo_check_home_dir(dir); @@ -304,6 +307,9 @@ a_Prefs_freeall(); a_Dw_freeall(); a_History_free(); + + gtk_widget_pop_colormap(); + gtk_widget_pop_visual(); g_print("Dillo: normal exit!\n"); return 0;

2 1

Re: Another review of Dillo 0.7.2
by Jorge Arellano Cid Aug. 18, 2003

Aug. 18, 2003

On Sun, 17 Aug 2003, Pete Spicer wrote: > Hi there! > > There was a review of Dillo 0.7.2 in Linux Format issue 42 (July 2003) which > featured it in their "Hot Picks" section (it got a half-page review with > screenshot browsing Slashdot.org on page 41) Good! Thanks for the information. > It was written by Mike Saunders, and quotes it as "starting up amazingly > quickly" and "works marvellously well for simple browsing tasks and the > rendering engine rivals Links for pure speed". It'd be very nice to have a link to it, but I don't know whether LinuxFormat has it online. Does anyone know the link? > Oh, and by the way, I use it myself on my machine when administering my > database with phpMyAdmin (just because it uses frames doesn't mean it can't > cope... provided you pick the right frame to be displayed, Dillo is fine > with PMA!) This is also very good to know. We may surprise you again soon! > Thanks for a great browser, > Pete :-) Cheers Jorge.-

1 0

dillo - my favorite browser (fwd)
by Jorge Arellano Cid Aug. 18, 2003

Aug. 18, 2003

---------- Forwarded message ---------- Date: Thu, 14 Aug 2003 20:57:41 +0200 From: Mario Moder <clay-man(a)freenet.de> To: jcid(a)softhome.net Subject: dillo - my favorite browser Hi! Congratulations! I have been using Mozilla Firebird 0.6.1 and it is very disappointing that it takes half a minute to start on my good old P1-200 with 64 MB RAM. Even when it is running, it runs really slow and uses a lot of memory. I tried the new dillo 0.7.3 and, wow, it started in one second and it renders 90% of the pages I visit everyday quite good and fast. I really look forward to watch this browser grow more mature but stay small and fast. Best wishes, Mario

1 0

Dillo hang as use chinese font
by Siu Kam Ming Aug. 18, 2003

Aug. 18, 2003

Hi , I was using the attached patch to show chinese character in dillo. It worked fine. I can use a very fast browser to browse my favorite chinese webpage. Since 0.7.2, dillo become unstable after applying this patch. As I selected any text in the rendered area, dillo hang. Does someone know the reason or the solution? Best Regards, Ming

2 1

Version 8 of tab+frame patch: smallproblemwithwhitespaceinmultilength
by Frank de Lange Aug. 13, 2003

Aug. 13, 2003

Hi'all, I just noticed that the current version of my tab+frames patch does not handle multilengths with whitespace correctly. I already fixed it in my version, will include the fix in the next version of the patch. If you encounter framesets which look weird, it may be caused by this problem. As the multilength parser was based on the one found in html.c, this problem is also present in anything which uses Html_parse_multi_length_list(), or which feeds length strings with leading whitespace to Html_parse_length_or_multi_length(). So, a multilength like "*,*" works fine, as does "* ,*", but "*. *" does not work correctly. As far as I can see, these functions are used in tables and embedded widgets. My fix in gtkframeset.c was to rewrite the parser into something slightly more robust (using g_strsplit and g_strstrip). Something similar is needed for html.c. I might include this in my patch, but unpatched Dillo also needs a fix so it is better solved in CVS. Cheers//Frank -- WWWWW ________________________ ## o o\ / Frank de Lange \ }# \| / +46-734352015 \ \ `--| _/ <Hacker for Hire> \ `---' \ +31-640037120 / \ frank(a)unternet.org / `------------------------' [ "Omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est." ]

1 0

RPM now available for SuSE 8.1!
by Kelson Vibber Aug. 13, 2003

Aug. 13, 2003

With the assistance of Patrick Shanahan, a Dillo 0.7.3 RPM suitable for SuSE 8.1 is now available! http://www.hyperborea.org/software/dillo/ Kelson Vibber www.hyperborea.org

1 0

[PATCH] Version 8 of tab+frame patch is out
by Frank de Lange Aug. 13, 2003

Aug. 13, 2003

Hi'all, The next version (8) of my tab patch is ready (see bottom of this message for download instructions). Frameset support has been upgraded to an 'almost usable' level. The major missing feature is manual frame resize support, which will come in a later version. Frameset (and iframe) support is enabled by default, it can be disabled using the configure option --enable-xhtml-dtd=strict. You can also use --enable-xhtml-dtd=transitional, which will give you iframe but no frameset support. The default option, --enable-xhtml-dtd=frameset, gives both iframe and frameset support. As usual, tab support can be disabled using --disable-tabs. Other new features (since the last version): - CTRL-R will reload a page (or frameset) Of course, the patch also fixes some bugs (and possibly introduces new ones). Please, test this patch and tell me what breaks. Or what works. Or whatever. More on frame/iframe support ============================ Frameset support is based on a new widget (GtkFrameset) which implements <frameset> as specified in xhtml-dtd-frameset. This widget is not Dillo-specific, so it can in theory be used in any GTK program. It is a pure GTK widget, not rooted in Dw. It does not provide any special support for Dw-specific features (eg. the full-screen-off-button is not displayed in the widget). I might make a special version for Dillo, which does implement these features. What works: - loading (nested) framesets, displaying frames at their appropriate location and appropriate size, including margin width/height. - loading (nested) iframes, displaying iframes at their appropriate location and with (mostly) appropriate size and alignment. - all frame-related menu options (in the popup menu). Frame menus can be accessed using the 'This Frame...' button in the popup menu, nested framesets have their own 'This Frameset...' menu. This way, Dillo allows you to limit actions to a subset of a framed document, eg. you can search for text in just one frame/iframe/frameset. Top-level menu options usually refer to the whole document (the frameset). An exception to this is the 'Jump To' option, which refers to the frame in which the menu was opened (question: should this contain the anchors for all frames?). Use the 'Show only this Frame' option in the Frame Options menu to get a full-page view of a frame. Same for 'Show only this Frameset'. What does not work yet: - the full_screen_off_button is not shown in frame documents. Use double-click instead to toggle the full screen setting... - manual frame resize is not supported yet. - the noresize and scrolling attributes are not supported yet Tabs, Frames and browser bloat ============================== 0.7.3 unpatched: 285152 bytes Patched, no tabs, no frames: 289356 bytes Patched, no tabs, with frames: 303052 bytes Patched, with tabs, no frames: 297100 bytes Patched, with tabs and frames: 310828 bytes The frameset implementation has grown about 10 Kb since the last patch. This is mainly due to the implementation of frame size support and the shift to a new widget (which has to be included in the binary, as it is not part of GTK (yet?)). Anyway, tabs alone cost about 12 Kb, frameset alone approximately 18 Kb, both together about 25 Kb. Bear in mind that the code has not been optimized for size yet. Download instructions ===================== There are now two versions of the patch, one for Dillo 0.7.3 (release) and one for Dillo 0.8.0-pre (CVS). Make sure to download the correct version for the Dillo source you want to patch! If you patch the CVS source, do not forget to run ./autogen.sh after applying the patch. The CVS patch is made against 20030813 CVS, while the 0.7.3 patch is made against Dillo-0.7.3.tar.bz2 (the source tarball on the main Dillo website). Both versions of the patch can be downloaded from the usual (ad-happy, but...) Geocities site: http://www.geocities.com/ikbenfrank/ More specifically, the CVS version can be found here: http://www.geocities.com/ikbenfrank/dillo-20030813-tabs.patch.gz size: 64875 bytes (gzipped, as downloaded) 272722 bytes (uncompressed) md5sum: a1315001bff22326b33a96b3828341b3 (gzipped, as downloaded) eed9681876272e81f161e254b684fd85 (uncompressed) The 0.7.3 version is here: http://www.geocities.com/ikbenfrank/dillo-0.7.3-tabs.patch.gz size: 117970 bytes (gzipped, as downloaded) 544402 bytes (uncompressed) md5sum: 93b55564ad4b18367f9632c4616e5890 (gzipped, as downloaded) c84a87687950d0e6670e01ea377d3ee7 (uncompressed) As always, check the site for the most recent version: That's all for now, Cheers//Frank -- WWWWW ________________________ ## o o\ / Frank de Lange \ }# \| / +46-734352015 \ \ `--| _/ <Hacker for Hire> \ `---' \ +31-640037120 / \ frank(a)unternet.org / `------------------------' [ "Omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est." ] -- WWWWW ________________________ ## o o\ / Frank de Lange \ }# \| / +46-734352015 \ \ `--| _/ <Hacker for Hire> \ `---' \ +31-640037120 / \ frank(a)unternet.org / `------------------------' [ "Omnis enim res, quae dando non deficit, dum habetur et non datur, nondum habetur, quomodo habenda est." ]

5 6

[patch] Remote Control
by Melvin Hadasht Aug. 13, 2003

Aug. 13, 2003

Hi, I just made a patch to add remote control for Dillo. 1. Summary: 1.1 Why? To be able to command Dillo from another process. E.g.: if Dillo is used as a help doc browser of a program, the program can implement context sensitive help and associate to each widget a help URL. The same Dillo instance can then be used to show the differents help docs. Other example, which also was the motivation of this patch: Sylpheed-Claws uses Dillo to render HTML mail, but currently it has to run a new instance each time the user views an HTML mail. Having only one instance is more efficient and improve responsiveness. 1.2. HowTo To be remote controlled, a Dillo instance must be explicitly started with a special command line option (-s, --server NAME), e.g.: dillo -s server A client instance sends commands to this server by referencing it and by using the -c, --command options, e.g.: a. dillo -s server -c "open(http://www.google.com)" -c "new-window(www.av.com)" b. dillo -s server http://www.google.com amounts to: dillo -s server -c "new-window(http://www.google.com)" Hence, running dillo -s server -c "open(URL)" mutliple times will launch only one instance of dillo at the first time, and the subsequent calls will just hand the arguments to the server. This is more practical than Mozilla's current behaviour where the user has to use two different commands to start and to control a server. The list of currently known commands is provided by the -H, --help-commands option: dillo -H Commands that can be used with the -c, --command option: open(URL) Open URL in current window. new-window(URL) Open URL in a new window. Exit() Terminate browsing session and exit. Command names are case insensitive. It is allowed to have more than one command option. New commands can be added, the syntax being: command(.*) 2. Ressources In dillorc: server_socket_path by default, it is "/tmp". It is used to create the socket name which is: $server_socket_path/dillo-$USER-$NAME. where $NAME is the server name from the -s, --server NAME option. E.g.: /tmp/dillo-john-help 3. Caveat If gtk specific command lines arguments are given, arguments are messed around, due to the fact that the patch delays gtk_init() after private arguments parsing, This will be fixed in the next version If a Dillo server does not quit cleanly, the socket corresponding to the server is not removed. Subsequent attempts to run a server with the same name will fail until the socket is removed (by default: /tmp/dillo-$USER-$SERVER_NAME). A solution is currently searched. The -c "exit()" works only when used from the client side The -c commands are currently ignored if no -s option is given. Any comments are welcome. Especially use cases for the remote control. 4. Details 4.1. Becoming a server Communication between server and client is done via UNIX domain socket. The advantage between this and FIFO being that a separate connection is made for each client, so communcation are guaranteed to not mixup. FIFO, on the other hand, can mixup communication of multiple clients if the messages are larger than some value (PIPE_BUF). This patch uses dynamic allocation for the messages received by the server. When the -s, --server NAME option is given, dillo checks wether a socket correspondig to that name is available and responding. If yes, it will hand any -c, --command CMD and any URL|FILE it finds in its command line options to that running server. URL|FILE are interpreted as "-c new-window(URL|FILE)". If no server socket is responding, it turns it self into a server by creating that socket. After that, it handles the -c, --command CMD and the URL|FILE options itself. 4.2. GUI initialisation The gtk_init() function is now called after dillo options parsing to speed up client execution. Indeed, a client instance does not need to initialise the GUI. This introduces the problem of messed up args if gtk args are given. This could affect the URL|FILE arguments. A solution would be to save them in a separate list for later processing. Another solution would be to interprete the URL|FILE arguments as -c open(URL|FILE) at an early stage. This will also help fixing the problem of the -c commands being ignored if no -s option is given. 4.3. Commands Commands are simple of the form: command(arg..). Arguments are parsed by the shell, so the user must take care that the command is interpreted as a single argument by the shell. Arguments should not contain unescaped parenthesis or spaces. Therefore, the following example is incorrect: dillo -s server -c open(file:///home/user/My documents) because: 1. the shell interprets the parenthesis, and 2. there is a space in the argument. This should be escaped in the following manner by using quotes: dillo -s server -c "open(file:///home/user/My documents)" It is easy to add new commands. -- Melvin Hadasht

3 4

Recovering health.
by Jorge Arellano Cid Aug. 12, 2003

Aug. 12, 2003

Hi there, Sorry about being absent from the list the past few days. I caught what here's called "rotavirus", it starts with stomach problems and when you recover, it knocks you with a heavy cold! That's several days, staying in bed, etc... (I'm now recovering from the cold). What we were doing? Ah yes, trying to make the 0.7.3 release. Well, as the bug discovered requires an extensive revision of the incremental resize mechanism, we decided to tag current CVS as 0.7.3-pre and upload 0.8.0-pre there and to continue as if it was released. Once there's a patch that fixes the bug, 0.7.3 will be released to the public, but in the meanwhile we'll be working on 0.8.0-pre. I updated the dpi1 spec (to be commited after this email), and hope to upload my 0.8.0-pre tarball before Sunday. Thorben: I sent a couple of emails to your auriga address, but got no answer. Maybe they were filtered as SPAM (some servers filter softhome, and I don't blame them!). That's why I'm writing to you here. Did you receive them? How is the email accounts enabling process at auriga going on? Ah, and I received a couple of SPAM rejection notes from external servers for SPAM delivered by our list. Is it possible to hook SpamAssassin to dillo-dev so it filters out SPAM? Best Jorge.-

2 1