January 2005 - Dillo-dev - lists.mailman3.com

[Dillo-dev]RPM status

by Kelson Vibber

Over the weekend I managed to get User-Mode Linux running again (under an old Red Hat 7.3 partition that I hadn't gotten around to deleting), so I was able to fill in RPMs for the rest of the releases I had used for Dillo 0.8.3. I also installed Mandrake 10.1, and was able to add an RPM for that release. I installed SuSE 9.2, though the combination of slow hardware, average DSL, slow mirror, the occasional missing file on the mirror and little free time to work with meant that it took several days to get a working install. Last night I finally built Dillo on it, but there's a font problem. When I get a chance I'll try installing some other GTK-1.2 apps and see if it's just a system config problem, and if it is, I'll post the RPM. At that point I'll have RPMs for current releases of Fedora Core, Mandrake, and SuSE, and a number of older releases. I'm thinking of dropping some of the older releases next time, just to simplify things. -- Kelson Vibber www.hyperborea.org

Jan. 20, 2005

Re: [Dillo-dev]Patch review: Frank de Lange patch (20031124)

by Roberto A. Foglietta

Bastiaan Jacques wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Hello, > > > I am dissecting and reviewing Frank de Lange's 14,000 line patch that > was rejected by the core dillo developers due to its size. I will email > the list from time to time to gather input as to whether a particular > portion of the patch is useful or not, beginning with the patch a little > further down this message. Is this patch 0.8.4 ready? Is this patch available? May try it and check the code it modified? Cheers, -- Roberto A. Foglietta Analista Programmatore GNU/Linux SAD Trasporto Locale S.p.a. Corso Italia 13/N 39100 BOLZANO (I) Tel. +39/0471-450.261 Fax +39/0471-450.253

Jan. 20, 2005

[Dillo-dev]Patch review: Frank de Lange patch (20031124)

by Bastiaan Jacques

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello, I am dissecting and reviewing Frank de Lange's 14,000 line patch that was rejected by the core dillo developers due to its size. I will email the list from time to time to gather input as to whether a particular portion of the patch is useful or not, beginning with the patch a little further down this message. The following patch does does two things in selection.c: - - make a_Selection_set_dclick_callback replace the callback instead of returning when the callback was already set. * This function is called only once in the code (in interface.c); the callback pointer always points to the same function. So I don't see a point in merging this first hunk, although it won't do any harm. Thoughts? - - Replace three comparison operators (==) with bitwise AND.. * This only works as expected when either lvalue and rvalue have exactly one or zero bits set to 1. Although this currently is true, this may change in the future; thus, I don't think these three hunks should be merged. Regards, Bastiaan Jacques. diff -Nur --exclude-from /home/andi/freeciv/diff_ignore dillo-bak/src/selection.c dillo/src/selection.c - --- dillo-bak/src/selection.c 2004-02-20 14:28:59.000000000 +0100 +++ dillo/src/selection.c 2004-02-20 14:30:17.000000000 +0100 @@ -76,7 +76,10 @@ ~ void (*fn) (gpointer data), ~ gpointer data) ~ { - - g_return_if_fail (selection->dclick_callback == NULL); + /* this used to return when the callback was already set, + * changed function to replace callback instead. + * g_return_if_fail (selection->dclick_callback == NULL); + */ ~ selection->dclick_callback = fn; ~ selection->callback_data = data; ~ } @@ -426,7 +429,7 @@ ~ DwIterator *top = it->stack[it->stack_top]; ~ gint len; - - if (top->content.type == DW_CONTENT_TEXT) + if (top->content.type & DW_CONTENT_TEXT) ~ len = strlen(top->content.data.text); ~ else ~ len = 1; @@ -486,7 +489,7 @@ ~ for (i = a_Dw_ext_iterator_clone (a), start = TRUE; ~ (cmp = a_Dw_ext_iterator_compare (i, b)) <= 0; ~ a_Dw_ext_iterator_next (i), start = FALSE) { - - if (i->content.type == DW_CONTENT_TEXT) { + if (i->content.type & DW_CONTENT_TEXT) { ~ if (fl) { ~ if (start) { ~ DEBUG_MSG (2, " highlighting %s from %d to %d\n", @@ -537,7 +540,7 @@ ~ cmp = a_Dw_ext_iterator_compare (selection->from, selection->to); ~ if (cmp == 0) { - - if (selection->from->content.type == DW_CONTENT_TEXT) { + if (selection->from->content.type & DW_CONTENT_TEXT) { ~ si = selection->from->stack[selection->from->stack_top]; ~ if (selection->from_char < selection->to_char) ~ tmp = g_strndup(si->content.data.text + selection->from_char, -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.0 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFB77I5wDeUwhiOHwURAkYeAKCzu03fggXw2DvdEiBWAP4GicLw5wCggUid Kg67vhiAUHc2dK+jGGVtE9o= =AGq9 -----END PGP SIGNATURE-----

Jan. 20, 2005

Re: [Dillo-dev]https, ssl & certificates

by robert w hall

In article <nNw1CoADb66BFwrn(a)n-cantrell.demon.co.uk>, robert w hall <bobh(a)n-cantrell.demon.co.uk> writes >But the situation over certificates appears confusing: - >as I (cursorily) read the code, it uses standard SSL functions to parse >entries in the .dillo/certs directory. OK, Garrett has been trying to help me understand root certificates. The problem I now appear to have is that openssl-0.9.6 had a fairly comprehensive block of certs but they're all out of date. Latest openssl (0.9.7) is more up-to-date but rather restricted (eg no RSA etc). Did anyone find a decent set of root certs hiding anywhere or find a workaround via their own self-signing route?? Bob -- robert w hall

Jan. 18, 2005

[Dillo-dev]https, ssl & certificates

by robert w hall

Where are we at and where are we going with the issue of https, ssl and certificates? I usually run with https enabled and find it is a VERY useful adjunct to the main program - it has quirky behaviour on sites such as playstation2-linux (where the answer is to log in via ssl then leave ssl), but at least it enables posting there. Other sites now accessible include amazon.co.uk and groups.yahoo.com - a significant gain. But the situation over certificates appears confusing: - as I (cursorily) read the code, it uses standard SSL functions to parse entries in the .dillo/certs directory. I have a block of certificates stolen from another browser but they don't seem to change the need to force by the 'remote certificate cannot be verified20' messages. (Are these imported certs in a non-standard format perhaps?) And the code also appears to have a function to store accepted certificates, but I've never been aware of this happening. The code is reaching a level of maturity and acceptance where these remaining blemishes are puzzling 'us users'. Bob -- robert w hall

Jan. 17, 2005

[Dillo-dev]Gtk2 port, GtkArg{Get,Set}Func

by Keenan Pepper

Hi, I'm working on a Gtk2 port of Dillo to use on my Zaurus (with the GPE rom from OpenZaurus). Before everyone starts saying "That's stupid, Gtk2 is so bloated and slow, why would you want to do that?", let me say that right now I'm using the Gtk2 port of Dillo 0.6.6, which as you all know is ancient and renders html badly compared to newer versions of Dillo, but it's still the best and fastest browser I can get to work on GPE. So periodically I'll be asking "What does this thing do?"-type questions on this list as I try to make current Dillo compile and run with Gtk2. For now I just want to request that all the deprecated (GtkArgSetFunc) NULL and (GtkArgGetFunc) NULL casts be removed and changed to simply NULL as recommended here: http://developer.gnome.org/doc/API/2.2/gtk/gtk-changes-1-2.html This has nothing to do with Gtk2, in fact eventually I should change all the GtkObject stuff to GObject stuff anyway, but when I first blindly tried to make it compile that was the error that came up the most, and it can and should be changed upstream. Cheers, Keenan Pepper

Jan. 15, 2005

[Dillo-dev]why can dillo be small and fast

by Yu Zenghui

Relative to other browsers, why can dillo be small and fast? Can anybody tell me the reason from the viewpoint of design and implementation. The code of dillo is fewer than other larger desktop browser. But fewer code means fewer functions. How to keep balance between the two conflicts? Thank you!

Jan. 14, 2005

Re: [Dillo-dev]cache replace

by Brian Hechinger

On Thu, Jan 13, 2005 at 12:20:26AM +0000, higuita wrote: > > also, at least now, you can only have one cookie enable > dillo active, so we have to "hunt" for that dillo when > we need a cookie enable > so cache replace is still needed 8) i'm currently working on the configuration dpi, and moving the cookie stuff to a dpi is my next task after i'm finished here (the cookie dpi is going to depend on the config dpi) so that at least solves the cookie issue. cache is the next thing on my list, we'll talk when i get there. > 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 i *LOVE* this quote, if bush wasn't such an idiot i could swear that he studied Goering. ;) -brian -- "Now you know why I got the everliving hell OUT of Windows administration. Knowing it doesn't make it any easier. It's just broken-as-designed."

Jan. 13, 2005

[Dillo-dev]what is the event for rendering

by Yu Zenghui

i try to look for the function which is used to begin rendering, but i failed. who can tell me what the function is? if there is no such function, what event triggers the rendering? does dillo begin to render after the whole html document is parsed, or once after one element is parsed? thank you!

Jan. 13, 2005

Re: [Dillo-dev]cache replace

by higuita

On Fri, 7 Jan 2005 00:11:41 -0700, Scott Scriven <dillo-dev(a)toykeeper.net> wrote: > 'net access. It also speeds things up compared to browsers which > check the network for updates before (or instead of) using cache. yep, i use this also maybe this could be one option (command line or run time)!? > Instead, I generally start a new instance for each site or task, > and close it when I'm done. I think I have about 20-30 open this doesnt work for low end machines, where we have little memory and/or little CPU also, at least now, you can only have one cookie enable dillo active, so we have to "hunt" for that dillo when we need a cookie enable so cache replace is still needed 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

Jan. 13, 2005