- Dillo-dev - lists.mailman3.com

revision control and other project infrastructure

by corvid＠lavabit.com

Jeremy wrote: > On Wed, Dec 17, 2008 at 05:37:44PM +0000, corvid wrote: > > > And, as you alluded to above, there is a > > http://freehg.org/u/dillo/main/ which might replace the cvs. > > AFAICT freehg.org provides nothing except repository hosting. If we > wanted forums etc. the devs would have to set that up somewhere else. > Is there a mercurial hosting service with a more "batteries-included" > approach? Right, I didn't mean to imply a forum at the same place as the repository. Scott mentioned user forums earlier in the msg, and it caught my eye. I don't really know whether a forum would get any activity, but I imagine the casual user's options right now look like 1) ask questions on some other forum that none of us frequent 2) ask Jorge directly -- and of course he's busy. Then again, I see there was a dillo-users list which apparently just got spam.

Dec. 17, 2008

revision control and other project infrastructure

by Johannes.Hofmann＠gmx.de

On Wed, Dec 17, 2008 at 06:27:15PM +0000, Jeremy Henty wrote: > On Wed, Dec 17, 2008 at 05:37:44PM +0000, corvid wrote: > > > And, as you alluded to above, there is a > > http://freehg.org/u/dillo/main/ which might replace the cvs. > > AFAICT freehg.org provides nothing except repository hosting. If we > wanted forums etc. the devs would have to set that up somewhere else. > Is there a mercurial hosting service with a more "batteries-included" > approach? Right, freehg.org is just a temporary place for the css-prototype branch and for some experimenting with mercurial in general. Regards, Johannes

Dec. 17, 2008

[patch]: use $(srcdir) in Make rules

by onepoint＠starurchin.org

I was revising the autoconf docs and noticed that make rules should prefix all sources with $(srcdir) to support separate builds[1]. We don't do this, so here's a patch. (Patch generated with Mercurial Queues, BTW.) Jeremy Henty [1]: (autoconf.info)Installation Directory Variables

Dec. 17, 2008

revision control and other project infrastructure

by onepoint＠starurchin.org

On Wed, Dec 17, 2008 at 05:37:44PM +0000, corvid wrote: > And, as you alluded to above, there is a > http://freehg.org/u/dillo/main/ which might replace the cvs. AFAICT freehg.org provides nothing except repository hosting. If we wanted forums etc. the devs would have to set that up somewhere else. Is there a mercurial hosting service with a more "batteries-included" approach? Regards, Jeremy Henty

Dec. 17, 2008

revision control and other project infrastructure

by corvid＠lavabit.com

Scott wrote: > I noticed some recent messages about revision control and > possibly switching to a DVCS. This makes me happy. :) > > So, I was wondering about general project infrastructure... > things like bug trackers, todo lists, mailing lists, help forums, > release management, news, patch queues and code review, code > repositories, etc. > > What I've found so far is the following: > > dillo.org: > - general project info > - release tarballs > - bug list > - FAQ > > auriga.wearlab.de: > - mailing lists > - cvs > > freshmeat.net: release announcements > > freenode: IRC (inactive?) > > http://misc.andi.de1.cc/dillo/ : Packaging for Debian, Ubuntu, Openmoko > > http://www.hyperborea.org/software/dillo/ : Packaging for Red Hat > > http://freehg.org/u/dillo/css-prototype : Prototype CSS branch > > Is that about right? Did I miss anything? Sounds about right to me. I don't know anything about the irc, though (In my day, irc was something that you turned to if you wanted to fail out of university, so I steered clear.) And, as you alluded to above, there is a http://freehg.org/u/dillo/main/ which might replace the cvs. I have wondered whether a help forum might do some good.

Dec. 17, 2008

revision control and other project infrastructure

by dillo-dev＠toykeeper.net

Hello. I noticed some recent messages about revision control and possibly switching to a DVCS. This makes me happy. :) So, I was wondering about general project infrastructure... things like bug trackers, todo lists, mailing lists, help forums, release management, news, patch queues and code review, code repositories, etc. What I've found so far is the following: dillo.org: - general project info - release tarballs - bug list - FAQ auriga.wearlab.de: - mailing lists - cvs freshmeat.net: release announcements freenode: IRC (inactive?) http://misc.andi.de1.cc/dillo/ : Packaging for Debian, Ubuntu, Openmoko http://www.hyperborea.org/software/dillo/ : Packaging for Red Hat http://freehg.org/u/dillo/css-prototype : Prototype CSS branch Is that about right? Did I miss anything? -- Scott

Dec. 17, 2008

dillo_dbg_rendering

by Johannes.Hofmann＠gmx.de

On Tue, Dec 16, 2008 at 12:11:47AM +0000, corvid wrote: > I'm finally bored enough to copy the cmdline code from dillo1. > > Does anyone care about the dillo_dbg_rendering option? > It forces borders in table.cc. I would remove it - at least once we add css support. table, td {border: 1px solid black !important} in the user stylesheet has the same effect. Cheers, Johannes

Dec. 16, 2008

dillo_dbg_rendering

by corvid＠lavabit.com

I'm finally bored enough to copy the cmdline code from dillo1. Does anyone care about the dillo_dbg_rendering option? It forces borders in table.cc.

Dec. 16, 2008

shaded colors and css

by Johannes.Hofmann＠gmx.de

On Sun, Dec 14, 2008 at 08:00:48PM +0000, corvid wrote: > Johannes wrote: > > On Sun, Nov 23, 2008 at 04:13:56AM +0000, corvid wrote: > > > Johannes wrote: > > > > mapping css properties to dw::core::style::Style members went very > > > > smooth, because Style was obviously implemented with css in mind. > > > > > > > > Just for the shaded color thing I can't find a corresponding concept > > > > in css. > > > > What exactly is the definition of a shaded color? Is it just a bit > > > > darker than the normal one? > > > > Does anyone know the corresponding concept in css? > > > > > > I had figured the shaded colors were just for giving > > > tables and rulers a three-dimensional appearance. > > > > Right, it seems to be used in drawBorder() in style.cc. > > Only now I noticed that the only difference between > > shaded colors and normal colors is that for shaded ones two more > > variations (SHADING_DARK and SHADING_LIGHT) are computed in addition > > to SHADING_NORMAL and SHADING_INVERSE. > > So it seems that it's an optimization to avoid calls to > > Color::shadeColor(). > > > > I'd like to get rid of the two types of colors and use always shaded > > colors. > > The current scheme adds complexity and can lead to errors > > (e.g. a normal color is passed to drawBorder(). > > > > In addition we can get more sharing of colors if they don't have a > > type member. > > > > If needed we could call Color::shadeColor() only on demand, but hide > > that optimization in the Color class. > > > > Any opinions? > > No objections here. > Ok, I pushed a corresponding change to css-prototype. Cheers, Johannes

Dec. 15, 2008

shaded colors and css

by corvid＠lavabit.com

Johannes wrote: > On Sun, Nov 23, 2008 at 04:13:56AM +0000, corvid wrote: > > Johannes wrote: > > > mapping css properties to dw::core::style::Style members went very > > > smooth, because Style was obviously implemented with css in mind. > > > > > > Just for the shaded color thing I can't find a corresponding concept > > > in css. > > > What exactly is the definition of a shaded color? Is it just a bit > > > darker than the normal one? > > > Does anyone know the corresponding concept in css? > > > > I had figured the shaded colors were just for giving > > tables and rulers a three-dimensional appearance. > > Right, it seems to be used in drawBorder() in style.cc. > Only now I noticed that the only difference between > shaded colors and normal colors is that for shaded ones two more > variations (SHADING_DARK and SHADING_LIGHT) are computed in addition > to SHADING_NORMAL and SHADING_INVERSE. > So it seems that it's an optimization to avoid calls to > Color::shadeColor(). > > I'd like to get rid of the two types of colors and use always shaded > colors. > The current scheme adds complexity and can lead to errors > (e.g. a normal color is passed to drawBorder(). > > In addition we can get more sharing of colors if they don't have a > type member. > > If needed we could call Color::shadeColor() only on demand, but hide > that optimization in the Color class. > > Any opinions? No objections here.

Dec. 14, 2008

shaded colors and css

by Johannes.Hofmann＠gmx.de

On Sun, Nov 23, 2008 at 04:13:56AM +0000, corvid wrote: > Johannes wrote: > > mapping css properties to dw::core::style::Style members went very > > smooth, because Style was obviously implemented with css in mind. > > > > Just for the shaded color thing I can't find a corresponding concept > > in css. > > What exactly is the definition of a shaded color? Is it just a bit > > darker than the normal one? > > Does anyone know the corresponding concept in css? > > I had figured the shaded colors were just for giving > tables and rulers a three-dimensional appearance. Right, it seems to be used in drawBorder() in style.cc. Only now I noticed that the only difference between shaded colors and normal colors is that for shaded ones two more variations (SHADING_DARK and SHADING_LIGHT) are computed in addition to SHADING_NORMAL and SHADING_INVERSE. So it seems that it's an optimization to avoid calls to Color::shadeColor(). I'd like to get rid of the two types of colors and use always shaded colors. The current scheme adds complexity and can lead to errors (e.g. a normal color is passed to drawBorder(). In addition we can get more sharing of colors if they don't have a type member. If needed we could call Color::shadeColor() only on demand, but hide that optimization in the Color class. Any opinions? Cheers, Johannes

Dec. 14, 2008

Plans for the near future

by francis＠daoine.org

On Thu, Dec 11, 2008 at 10:24:38AM +0000, Jeremy Henty wrote: > On Wed, Dec 10, 2008 at 04:17:02PM -0300, Jorge Arellano Cid wrote: > > 2.- Basic authentication > > a) I have to review this patch after working on point 1.- > > b) This is a very requested feature after CSS. > > It'd be great to have it working. > > It works locally on various web servers (Apache and others) and also > on the LinuxFromScratch wiki. So (subject to any further comments I > get) I think it is working. My test machine has the wrong autotools, so I'm afraid I can't built and test this myself, but: does it work as expected if an img on a page requires authentication? As far as I recall (hazy memory) the unofficial patches for dillo1 didn't. If this one does, then great! Cheers, f -- Francis Daly francis(a)daoine.org

Dec. 11, 2008

Plans for the near future

by jcid＠dillo.org

On Thu, Dec 11, 2008 at 10:24:38AM +0000, Jeremy Henty wrote: > On Wed, Dec 10, 2008 at 04:17:02PM -0300, Jorge Arellano Cid wrote: > > > 2.- Basic authentication > > a) I have to review this patch after working on point 1.- > > b) This is a very requested feature after CSS. > > It'd be great to have it working. > > It works locally on various web servers (Apache and others) and also > on the LinuxFromScratch wiki. So (subject to any further comments I > get) I think it is working. If anyone has an account they'd like me > to test then I'll try it out. Good. > > 5.- Close icon in Tab label > > a) This is much needed as anyone using Tabs has experienced. > > b) Not too hard to implement, but I have to work on 1 and 2. > > I'd like to do this but I don't see how to get a button into the tab > label. It just displays the widget's label(). I wondered if there > was some Symbol magic that might do it but Symbols don't seem to > respond to events. Am I missing something? Maybe the close button > will have to go in the tab body. Perhaps to the right of the web > search bar? (Top-right seems to be the traditional location.) My original idea was to use a '+' symbol rotated 45 degrees, then to override the virtual handle method to react when the click was in its "zone". Just an idea, I haven't coded anything. There was an aqua-like tab widget somewhere. I was suggested to look at it for insight. I don't have the reference now. :( -- Cheers Jorge.-

Dec. 11, 2008

Plans for the near future

by onepoint＠starurchin.org

On Wed, Dec 10, 2008 at 04:17:02PM -0300, Jorge Arellano Cid wrote: > 2.- Basic authentication > a) I have to review this patch after working on point 1.- > b) This is a very requested feature after CSS. > It'd be great to have it working. It works locally on various web servers (Apache and others) and also on the LinuxFromScratch wiki. So (subject to any further comments I get) I think it is working. If anyone has an account they'd like me to test then I'll try it out. > 5.- Close icon in Tab label > a) This is much needed as anyone using Tabs has experienced. > b) Not too hard to implement, but I have to work on 1 and 2. I'd like to do this but I don't see how to get a button into the tab label. It just displays the widget's label(). I wondered if there was some Symbol magic that might do it but Symbols don't seem to respond to events. Am I missing something? Maybe the close button will have to go in the tab body. Perhaps to the right of the web search bar? (Top-right seems to be the traditional location.) > 6.- Switch to mercurial > a) From comments it looks like switching to Hg is a good idea. > b) Yes, I'm only updating Mercurial these days as an experiment. > c) Switching as soon as I have some time to do it. Probably after > 1.- and 2.- Cool! Jeremy Henty

Dec. 11, 2008

Dillo and Gmail

by daniel.victoria＠gmail.com

Wow, that's a lot of things for a newbie to do... I guess I'll just keep firefox along with dillo. One for gmail and the other for quick things Thanks for your help anyway Daniel On Fri, Dec 5, 2008 at 11:56 AM, Francis Daly <francis(a)daoine.org> wrote: > On Thu, Dec 04, 2008 at 10:17:00PM -0200, Daniel Victoria wrote: > > Hi there, > >> The only thing I'm missing is Gmail. Does Dillo work with gmail? Do I >> need to do any special tricks? > > I have it "working" with dillo1, and don't see why it shouldn't work in > the same way with dillo2. > > "Working" means "I can read and write mail, http only" -- but only after > I've got the authentication cookie with a different browser. (My dillo1 > doesn't do https.) > > If that's good enough for you, then you can try the same with dillo2 > and report on whether it still works. > > My cookiesrc says: > > DEFAULT DENY > mail.google.com ACCEPT > > I use a mozilla browser to sign in, and then I don't sign out so as > not to invalidate the cookie on the server. It seems to remain valid > for a couple of weeks, so I occasionally have to redo this part. Find > the cookie called GX, and write it to a file in the correct format -- > 7 tab-separated fields. My current one is a file ~/gmcookie which looks > like this (apart from the value of the cookie, which has been modified): > > mail.google.com TRUE /mail FALSE 2000000000 GX DQAAAIIAAACY9ABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFHELLOKITTYEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEF__p_abcdef > > Then, to load or reload the cookie into dillo, I do > > pkill -HUP cookie > pkill cookie > cat $HOME/gmcookie >> $HOME/.dillo/cookies.txt > > And then I go to my bookmarked gmail page at > > http://mail.google.com/mail/?ui=html&zy=n > > and it all Just Works. If I get redirected to a https page, that indicates > that the cookie is absent or invalid, so I put it back or get a fresh one. > > No guarantees that it will keep working, of course. The gmail folk can > change their authentication requirements at any stage. And you miss out on > the javascript goodness that is a large part of the gmail experience. And > there was a stage when the html returned was broken in a way that dillo > couldn't usefully handle, although that has since been changed back. But for > straightforward mail tasks, it works well enough for me. > > Couple it with a lightweight gmail-notifier script, and you don't have > to hit reload to see the new news. > > Good luck, > > f > -- > Francis Daly francis(a)daoine.org >

Dec. 10, 2008

Plans for the near future

by akemnade＠tzi.de

Hi, On Wed, 10 Dec 2008 16:17:02 -0300 Jorge Arellano Cid <jcid(a)dillo.org> wrote: > Hi there, > > Sorry for the unresponsiveness. I'm too busy in my regular job > these days... > > Besides that, I've read the mailing list and gave some thought > to these interesting topics. > > 1.- CSS plan (basic) > 2.- Basic authentication > 3.- Toolbox Icon > 4.- Gmail compatibility (at least read/send) > 5.- Close icon in Tab label > 6.- Switch to mercurial 7. (or better earlier?) long left mouse button press = right mouse button press -> should make things easier for people without a real right mouse button. with touchscreens you have pratically only one button, if you have luck there is mouse button rotation applet. popup menus are hardly accessible If desired, I guess I can make the patch for that. Greetings Andreas Kemnade

Dec. 10, 2008

Plans for the near future

by jcid＠dillo.org

Hi there, Sorry for the unresponsiveness. I'm too busy in my regular job these days... Besides that, I've read the mailing list and gave some thought to these interesting topics. 1.- CSS plan (basic) 2.- Basic authentication 3.- Toolbox Icon 4.- Gmail compatibility (at least read/send) 5.- Close icon in Tab label 6.- Switch to mercurial I believe our next release would attract more people if it had, at least, implemented a couple in the {1,2,3,4,5} set. ----------------- A bit more detail: ----------------- 1.- CSS plan (basic) a) I have to implement the network retrieval of CSS stylesheets, so Johannes can continue working from there. b) Deciding a basic CSS set (e.g. just colors) is the next step before including it in the main tree c) Hooking it to a dillorc preference and to point 3.- (Read below) 2.- Basic authentication a) I have to review this patch after working on point 1.- b) This is a very requested feature after CSS. It'd be great to have it working. 3.- Toolbox Icon a) Since a long time ago I'd like to have a Toolbox icon to the right of bookmarks with a drop down menu. Something like: [Tb Icon] .-------------------. .---------. | Change charset >|--------------------------------->| UTF-8 | | Force my colors >|-------------------->.-----. | Latin1 | | CSS >|---------->.-----. | On | | Cp 1250 | | Free memory >|--. | On | | Off | | ... | '-------------------' | | Off | '-----' '---------' | '-----' | .------------------. '->| Images | | Local Files | | Everything unused | '-------------------' 4.- Gmail compatibility (at least read/send) a) From what I read it's clear that at least it can be made to work with some tweaks. What do we need to have it working out of the box? b) It'd make a nice headline! ("Fast Read/Send from gmail"). 5.- Close icon in Tab label a) This is much needed as anyone using Tabs has experienced. b) Not too hard to implement, but I have to work on 1 and 2. 6.- Switch to mercurial a) From comments it looks like switching to Hg is a good idea. b) Yes, I'm only updating Mercurial these days as an experiment. c) Switching as soon as I have some time to do it. Probably after 1.- and 2.- -- Cheers Jorge.-

Dec. 10, 2008

Thoughts on revision control

by onepoint＠starurchin.org

On Tue, Dec 09, 2008 at 09:17:05PM +0100, Johannes Hofmann wrote: > Given the current patch-rate we could also easily stick with CVS for > the main repository I think. It's just for the longer living > css-prototype branch where I wanted something more comfortable than > CVS. So what will happen to non-css patches (*cough* HTTP auth *cough*)? Will they go into CVS or the dillo-main Mercurial repository? Didn't Jorge say something about committing only to Mercurial for a while? (Or did I misunderstand?) To be honest, why not migrate to Mercurial now? After all, CVS *is* obsolete and clunky, Mercurial *can* do the job, and some key Dillo developers are *already* maintaining and using Mercurial repositories of the Dillo source. It would also save me the trouble of layering git over CVS just to manage my local patches. Why not just do it? Regards, Jeremy Henty -- And so concludes the verse "Jabberwocky" by Lewis Carroll. A piece of nonsense verse embedded in a nonsense work, it resembles the "real world" about as much as a computer program does. Which is perhaps why Lewis Carroll is so popular among computer programmers. -- Craig A. Finseth.

Dec. 10, 2008

empty html elements ... and Shift+{Left, Right} change

by bf2006a＠yahoo.com

--- On Tue, 12/9/08, corvid <corvid(a)lavabit.com> wrote: > From: corvid <corvid(a)lavabit.com> > Subject: Re: [Dillo-dev] empty html elements ... and Shift+{Left, Right} change > To: dillo-dev(a)dillo.org > Date: Tuesday, December 9, 2008, 1:51 AM > b. wrote: > > Another note: I've noticed that with dillo cvs > snapshots dated after > > Nov. 9 there is a regression (or at least a change) in > the behavior of > > Shift+{Left,Right}: in the later snapshots I lose the > ability to shift > > to other tabs with these keys after right-clicking on > a link in the current > > tab with the mouse, and can only regain this ability > to shift tabs after > > opening a new tab with a middle-click on a link in the > current tab, or by > > navigating to an existing tab with the mouse and > closing it, or by quickly > > opening and closing a blank tab with Ctrl+T,Ctrl+Q > combination. Any of > > these methods significantly slows browsing with many > tabs open. Was this > > intended? Is there a new dillorc option to control > this behavior, or some > > new key/mouse controls that I missed? > > You could try this: > Your patch seems to restore the lost functionality. Thank you for the quick response. Regards, b.

Dec. 10, 2008

how about a form menu?

by corvid＠lavabit.com

Johannes wrote: > On Mon, Dec 08, 2008 at 11:10:59PM +0000, corvid wrote: > > I was playing around with making a form menu. > > So far: > > 1) Submit. Nicer from the user's perspective than prefs.generate_submit. > > Yes, this seems cleaner. Is it possible to get the menu on a form with > just a single text input? I assume one has to right-click on the > form area. Yeah, right click on any form control. > > 2) Reset. Not really needed when you can just press back and forward. > > 3) Show/Hide Hiddens. Lets you edit them, which I believe the user should > > have the power to do. > > That would be really nice! > > > ..and thinking of making an Info option to show submission charset and, oh, > > whatever else. > > Maybe some sort of save/load form data to/from file? Oh, that's a good idea...

Dec. 9, 2008

Thoughts on revision control

by Johannes.Hofmann＠gmx.de

On Mon, Dec 08, 2008 at 11:57:57AM +0000, Jeremy Henty wrote: > > I've been using git to manage the Dillo source locally. I just > installed mercurial to experiment with. Things I like about git are > the ability to rebase, ie. reorder patches. This is neat because I > can develop a patch on top of my local modifications and then push it > down the stack so that I submit a patch against the pristine source > rather than my customised version. The ability to stash your current > changes in a temporary branch is useful too. You could do all this > with explicit branches, but it would be more involved and git gives > you nice, quick and easy ways to do the same thing. > > Mercurial seems to be the same sort of beast, though perhaps it is a > bit less of a "Swiss Army Chainsaw" than git. I'd be surprised if > there were any real issues using either to manage Dillo since I doubt > we'd be using anything more than the basic facilities that both offer. > I'd prefer git but only out of laziness as it's the one I already know > (which means I should probably investigate mercurial just to broaden > my mind if nothing else). It's the same for me with git. I wouldn't mind learning to use it. Given the current patch-rate we could also easily stick with CVS for the main repository I think. It's just for the longer living css-prototype branch where I wanted something more comfortable than CVS. Regards, Johannes

Dec. 9, 2008

how about a form menu?

by Johannes.Hofmann＠gmx.de

On Mon, Dec 08, 2008 at 11:10:59PM +0000, corvid wrote: > I was playing around with making a form menu. > So far: > 1) Submit. Nicer from the user's perspective than prefs.generate_submit. Yes, this seems cleaner. Is it possible to get the menu on a form with just a single text input? I assume one has to right-click on the form area. > 2) Reset. Not really needed when you can just press back and forward. > 3) Show/Hide Hiddens. Lets you edit them, which I believe the user should > have the power to do. That would be really nice! > ..and thinking of making an Info option to show submission charset and, oh, > whatever else. Maybe some sort of save/load form data to/from file? Cheers, Johannes

Dec. 9, 2008

empty html elements ... and Shift+{Left, Right} change

by thomas-forum＠orgis.org

Am Mon, 8 Dec 2008 18:54:18 -0800 (PST) schrieb bf <bf2006a(a)yahoo.com>: > Leanness and speed are always nice, but are they and the ability to reduce > the size of an existing instance necessarily mutually exclusive? Unless > the introduction of a way to empty caches/buffers would require a huge > amount of new code, it they don't seem to be. I second that... being able to free RAM should be a good thing. I usually keep apps open for a long time, especially at least one browser window. Alrighty then, Thomas.

Dec. 9, 2008

empty html elements ... and Shift+{Left, Right} change

by corvid＠lavabit.com

b. wrote: > Another note: I've noticed that with dillo cvs snapshots dated after > Nov. 9 there is a regression (or at least a change) in the behavior of > Shift+{Left,Right}: in the later snapshots I lose the ability to shift > to other tabs with these keys after right-clicking on a link in the current > tab with the mouse, and can only regain this ability to shift tabs after > opening a new tab with a middle-click on a link in the current tab, or by > navigating to an existing tab with the mouse and closing it, or by quickly > opening and closing a blank tab with Ctrl+T,Ctrl+Q combination. Any of > these methods significantly slows browsing with many tabs open. Was this > intended? Is there a new dillorc option to control this behavior, or some > new key/mouse controls that I missed? You could try this:

Dec. 9, 2008

empty html elements ... and Shift+{Left, Right} change

by bf2006a＠yahoo.com

--- On Mon, 12/8/08, corvid <corvid(a)lavabit.com> wrote: > From: corvid <corvid(a)lavabit.com> > Subject: Re: [Dillo-dev] empty html elements ... and Shift+{Left, Right} change > To: dillo-dev(a)dillo.org > Date: Monday, December 8, 2008, 7:39 PM > bf wrote: > > I'm surprised > > we don't have a feature to clear image > caches/buffers without restart -- or > > am I missing something?). > > I think a good thing is to keep dillo nimble enough that it > feels more like a utility than an application. I can open > and use instances whenever I need them, much like rxvt > (and unlike xterm, which I have to wait for). > Leanness and speed are always nice, but are they and the ability to reduce the size of an existing instance necessarily mutually exclusive? Unless the introduction of a way to empty caches/buffers would require a huge amount of new code, it they don't seem to be. Consider the following scenario: you've opened a number of tabs in one instance, containing content you'd still like to look over. Then you open a few more, and find that the new pages occupy a large amount of memory, which you need for other uses. Right now, you can't close these tabs and clear the cache/buffer to bring the memory footprint back down to a lower level. Sure, you could kill the entire instance and start over, but what about the first set of tabs that you're not done with? You'd have to navigate back to each of them in a new instance, which could take a lot of time. You could mitigate that by bookmarking them, but that still takes time, and then you may have to clean your bookmarks up, and unless you're using a cache other than that of the browser, you'd have to download them again, too. This doesn't seem desirable. Am I missing something? > > Another note: I've noticed that with dillo cvs > snapshots dated after > > Nov. 9 there is a regression (or at least a change) in > the behavior of > > Shift+{Left,Right}: in the later snapshots I lose the > ability to shift > > to other tabs with these keys after right-clicking on > a link in the current > > tab with the mouse, and can only regain this ability > to shift tabs after > > opening a new tab with a middle-click on a link in the > current tab, or by > > navigating to an existing tab with the mouse and > closing it, or by quickly > > opening and closing a blank tab with Ctrl+T,Ctrl+Q > combination. Any of > > these methods significantly slows browsing with many > tabs open. Was this > > intended? Is there a new dillorc option to control > this behavior, or some > > new key/mouse controls that I missed? > > Hmmm. I must have broken it with > http://cvs.auriga.wearlab.de/cgi-bin/cvsweb.cgi/dillo2/dw/fltkviewbase.cc.d… > A voice offstage right: "Shame! Shame!"

Dec. 9, 2008

empty html elements ... and Shift+{Left, Right} change

by corvid＠lavabit.com

bf wrote: > I'm surprised > we don't have a feature to clear image caches/buffers without restart -- or > am I missing something?). I think a good thing is to keep dillo nimble enough that it feels more like a utility than an application. I can open and use instances whenever I need them, much like rxvt (and unlike xterm, which I have to wait for). > Another note: I've noticed that with dillo cvs snapshots dated after > Nov. 9 there is a regression (or at least a change) in the behavior of > Shift+{Left,Right}: in the later snapshots I lose the ability to shift > to other tabs with these keys after right-clicking on a link in the current > tab with the mouse, and can only regain this ability to shift tabs after > opening a new tab with a middle-click on a link in the current tab, or by > navigating to an existing tab with the mouse and closing it, or by quickly > opening and closing a blank tab with Ctrl+T,Ctrl+Q combination. Any of > these methods significantly slows browsing with many tabs open. Was this > intended? Is there a new dillorc option to control this behavior, or some > new key/mouse controls that I missed? Hmmm. I must have broken it with http://cvs.auriga.wearlab.de/cgi-bin/cvsweb.cgi/dillo2/dw/fltkviewbase.cc.d…

Dec. 9, 2008

empty html elements ... and Shift+{Left, Right} change

by bf2006a＠yahoo.com

> > Oh, I committed a new version allowing [ "'] > before />, > and also a self closing element: <x/>. > > This because the former one failed on things like: > > <a href=/html/k_links/Kayden_Kross/> > Kayden Kross > </a> > > Example taken from: > > http://www.freeones.com/html/d_links/Denise_Milani/ > > look at the non-working bottom links, > or elsewhere if you like. :-) Yeah, and of course you really don't want to fail when it comes to Denise or Kayden. (Given the line this "testing" is taken, I'm surprised we don't have a feature to clear image caches/buffers without restart -- or am I missing something?). Another note: I've noticed that with dillo cvs snapshots dated after Nov. 9 there is a regression (or at least a change) in the behavior of Shift+{Left,Right}: in the later snapshots I lose the ability to shift to other tabs with these keys after right-clicking on a link in the current tab with the mouse, and can only regain this ability to shift tabs after opening a new tab with a middle-click on a link in the current tab, or by navigating to an existing tab with the mouse and closing it, or by quickly opening and closing a blank tab with Ctrl+T,Ctrl+Q combination. Any of these methods significantly slows browsing with many tabs open. Was this intended? Is there a new dillorc option to control this behavior, or some new key/mouse controls that I missed? Regards, b.

Dec. 9, 2008

how about a form menu?

by corvid＠lavabit.com

I was playing around with making a form menu. So far: 1) Submit. Nicer from the user's perspective than prefs.generate_submit. 2) Reset. Not really needed when you can just press back and forward. 3) Show/Hide Hiddens. Lets you edit them, which I believe the user should have the power to do. ..and thinking of making an Info option to show submission charset and, oh, whatever else. What I have now is icky proof-of-concept code (although it's been working fine), so I want to see how well the idea goes over before trying to clean it up.

Dec. 9, 2008

empty html elements

by jcid＠dillo.org

On Mon, Dec 01, 2008 at 06:56:12PM +0100, Johannes Hofmann wrote: > Hi, > > while working on css I came across the following: > > <a class="d" name="INCR32"/> (found at [1]) > > dillo currently doesn't recognize that the element is empty. > The relevant code is at html.cc:3530: > > /* Test for </x>, ReqTagClose, <x /> and <x/> */ > if (*start == '/' || /* </x> */ > html->ReqTagClose || /* request */ > (tag[tagsize - 2] == '/' && /* XML: */ > (isspace(tag[tagsize - 3]) || /* <x /> */ > (size_t)tagsize == strlen(Tags[ni].name) + 3))) { /* <x/> */ > > > We could simply drop the requirement for the blank before '/' which > would lead to > > /* Test for </x>, ReqTagClose, <x /> and <x/> */ > if (*start == '/' || /* </x> */ > html->ReqTagClose || /* request */ > (tag[tagsize - 2] == '/')) { /* XML */ > > > or allow a '"' instead of the blank or keep things as they are, if > XML requires a blank before the '/'. > >>From http://www.w3.org/TR/2006/REC-xml-20060816/#sec-starttags > I'd say the blank is optional. > > What do you think? > > Cheers, > Johannes > > [1] http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/lib/libc/port… Oh, I committed a new version allowing [ "'] before />, and also a self closing element: <x/>. This because the former one failed on things like: <a href=/html/k_links/Kayden_Kross/> Kayden Kross </a> Example taken from: http://www.freeones.com/html/d_links/Denise_Milani/ look at the non-working bottom links, or elsewhere if you like. :-) -- Cheers Jorge.-

Dec. 8, 2008

Thoughts on revision control

by onepoint＠starurchin.org

I've been using git to manage the Dillo source locally. I just installed mercurial to experiment with. Things I like about git are the ability to rebase, ie. reorder patches. This is neat because I can develop a patch on top of my local modifications and then push it down the stack so that I submit a patch against the pristine source rather than my customised version. The ability to stash your current changes in a temporary branch is useful too. You could do all this with explicit branches, but it would be more involved and git gives you nice, quick and easy ways to do the same thing. Mercurial seems to be the same sort of beast, though perhaps it is a bit less of a "Swiss Army Chainsaw" than git. I'd be surprised if there were any real issues using either to manage Dillo since I doubt we'd be using anything more than the basic facilities that both offer. I'd prefer git but only out of laziness as it's the one I already know (which means I should probably investigate mercurial just to broaden my mind if nothing else). Regards, Jeremy Henty

Dec. 8, 2008

Html_tag_close_* called twice

by jcid＠dillo.org

On Sun, Dec 07, 2008 at 08:50:50PM +0100, Johannes Hofmann wrote: > Hi, > > corvid pointed out that Html_tag_close_* functions are getting > called twice. This is causing problems when parsing > <style>...</style> contents in Html_tag_close_style() > as done in css-prototype now. > We could work around this somehow, but wouldn't it be cleaner to > make sure Html_tag_close_* is called just once, when an element > is closed? Oh, true... Patch commited to Hg repository. Note: I left nice MSG output on, so you can see the parsed HTML structure on the fly. -- Cheers Jorge.-

Dec. 8, 2008

empty html elements

by corvid＠lavabit.com

Johannes wrote: > On Sun, Dec 07, 2008 at 04:38:06PM -0300, Jorge Arellano Cid wrote: > > BTW, I want to give mercurial a more serious try so will keep > > updating only this repo for some time as a test: > > > > (http://freehg.org/u/dillo/main) > > > > In a nutshell: > > To get the repo into your HD: > > hg clone http://freehg.org/u/dillo/main dillo > > To update: > > cd dillo > > hg pull http://freehg.org/u/dillo/main > > > > Yes, I'm yet to update the web pages in our site... > > I personally like mercurial, but what do the other developers think? Hmm...looks like I'll have to make some room for python...

Dec. 7, 2008

empty html elements

by Johannes.Hofmann＠gmx.de

On Sun, Dec 07, 2008 at 04:38:06PM -0300, Jorge Arellano Cid wrote: > On Mon, Dec 01, 2008 at 06:56:41PM +0000, corvid wrote: > > Johannes wrote: > > > Hi, > > > > > > while working on css I came across the following: > > > > > > <a class="d" name="INCR32"/> (found at [1]) > > > > > > dillo currently doesn't recognize that the element is empty. > > > The relevant code is at html.cc:3530: > > > > > > /* Test for </x>, ReqTagClose, <x /> and <x/> */ > > > if (*start == '/' || /* </x> */ > > > html->ReqTagClose || /* request */ > > > (tag[tagsize - 2] == '/' && /* XML: */ > > > (isspace(tag[tagsize - 3]) || /* <x /> */ > > > (size_t)tagsize == strlen(Tags[ni].name) + 3))) { /* <x/> */ > > > > > > > > > We could simply drop the requirement for the blank before '/' which > > > would lead to > > > > > > /* Test for </x>, ReqTagClose, <x /> and <x/> */ > > > if (*start == '/' || /* </x> */ > > > html->ReqTagClose || /* request */ > > > (tag[tagsize - 2] == '/')) { /* XML */ > > > > > > > > > or allow a '"' instead of the blank or keep things as they are, if > > > XML requires a blank before the '/'. > > > >From http://www.w3.org/TR/2006/REC-xml-20060816/#sec-starttags > > > I'd say the blank is optional. > > > > > > What do you think? > > > > Yes, it looks like whitespace is optional there. > > > > EmptyElemTag ::= '<' Name (S Attribute)* S? '/>' > > S ::= (#x20 | #x9 | #xD | #xA)+ > > > Committed to the mercurial repository. > > BTW, I want to give mercurial a more serious try so will keep > updating only this repo for some time as a test: > > (http://freehg.org/u/dillo/main) > > In a nutshell: > To get the repo into your HD: > hg clone http://freehg.org/u/dillo/main dillo > To update: > cd dillo > hg pull http://freehg.org/u/dillo/main > > Yes, I'm yet to update the web pages in our site... I personally like mercurial, but what do the other developers think? Cheers, Johannes

Dec. 7, 2008

Html_tag_close_* called twice

by Johannes.Hofmann＠gmx.de

Hi, corvid pointed out that Html_tag_close_* functions are getting called twice. This is causing problems when parsing <style>...</style> contents in Html_tag_close_style() as done in css-prototype now. We could work around this somehow, but wouldn't it be cleaner to make sure Html_tag_close_* is called just once, when an element is closed? Cheers, Johannes

Dec. 7, 2008

empty html elements

by jcid＠dillo.org

On Mon, Dec 01, 2008 at 06:56:41PM +0000, corvid wrote: > Johannes wrote: > > Hi, > > > > while working on css I came across the following: > > > > <a class="d" name="INCR32"/> (found at [1]) > > > > dillo currently doesn't recognize that the element is empty. > > The relevant code is at html.cc:3530: > > > > /* Test for </x>, ReqTagClose, <x /> and <x/> */ > > if (*start == '/' || /* </x> */ > > html->ReqTagClose || /* request */ > > (tag[tagsize - 2] == '/' && /* XML: */ > > (isspace(tag[tagsize - 3]) || /* <x /> */ > > (size_t)tagsize == strlen(Tags[ni].name) + 3))) { /* <x/> */ > > > > > > We could simply drop the requirement for the blank before '/' which > > would lead to > > > > /* Test for </x>, ReqTagClose, <x /> and <x/> */ > > if (*start == '/' || /* </x> */ > > html->ReqTagClose || /* request */ > > (tag[tagsize - 2] == '/')) { /* XML */ > > > > > > or allow a '"' instead of the blank or keep things as they are, if > > XML requires a blank before the '/'. > > >From http://www.w3.org/TR/2006/REC-xml-20060816/#sec-starttags > > I'd say the blank is optional. > > > > What do you think? > > Yes, it looks like whitespace is optional there. > > EmptyElemTag ::= '<' Name (S Attribute)* S? '/>' > S ::= (#x20 | #x9 | #xD | #xA)+ Committed to the mercurial repository. BTW, I want to give mercurial a more serious try so will keep updating only this repo for some time as a test: (http://freehg.org/u/dillo/main) In a nutshell: To get the repo into your HD: hg clone http://freehg.org/u/dillo/main dillo To update: cd dillo hg pull http://freehg.org/u/dillo/main Yes, I'm yet to update the web pages in our site... -- Cheers Jorge.- PS: I've been quite busy, trying to catch up. Please be patient...

Dec. 7, 2008

[ladislav@distrowatch.com: Re: Donation for Dillo]

by jcid＠dillo.org

FYI, -- Cheers Jorge.-

Dec. 7, 2008

[patch]: Basic Authentication

by onepoint＠starurchin.org

Next iteration: one bugfix. * In Auth_parse_auth_basic(), fix the test for the "realm" token. Jeremy Henty

Dec. 6, 2008

Dillo and Gmail

by furaisanjin＠gmail.com

Hi Have you tried http://gmail.com? I can see mails from there. I modified dillo HTTP request header so some modification might be needed. Regards, furaisanjin

Dec. 6, 2008

[patch]: Basic Authentication (updated)

by onepoint＠starurchin.org

On Fri, Dec 05, 2008 at 05:20:38PM +0000, corvid wrote: > Jeremy wrote: > > + > > + token_value_pairs_found = 1; > > + > > + if (strncasecmp("realm",token,token_size) == 0 && realm == NULL) { > > Token could be a prefix of "realm". <head-desk/> Will fix. You're good at this! ;-) Jeremy Henty

Dec. 6, 2008

[patch]: Basic Authentication (updated)

by corvid＠lavabit.com

Jeremy wrote: > + > + token_value_pairs_found = 1; > + > + if (strncasecmp("realm",token,token_size) == 0 && realm == NULL) { Token could be a prefix of "realm".

Dec. 5, 2008

Dillo and Gmail

by francis＠daoine.org

On Thu, Dec 04, 2008 at 10:17:00PM -0200, Daniel Victoria wrote: Hi there, > The only thing I'm missing is Gmail. Does Dillo work with gmail? Do I > need to do any special tricks? I have it "working" with dillo1, and don't see why it shouldn't work in the same way with dillo2. "Working" means "I can read and write mail, http only" -- but only after I've got the authentication cookie with a different browser. (My dillo1 doesn't do https.) If that's good enough for you, then you can try the same with dillo2 and report on whether it still works. My cookiesrc says: DEFAULT DENY mail.google.com ACCEPT I use a mozilla browser to sign in, and then I don't sign out so as not to invalidate the cookie on the server. It seems to remain valid for a couple of weeks, so I occasionally have to redo this part. Find the cookie called GX, and write it to a file in the correct format -- 7 tab-separated fields. My current one is a file ~/gmcookie which looks like this (apart from the value of the cookie, which has been modified): mail.google.com TRUE /mail FALSE 2000000000 GX DQAAAIIAAACY9ABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFHELLOKITTYEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEFABCDEF__p_abcdef Then, to load or reload the cookie into dillo, I do pkill -HUP cookie pkill cookie cat $HOME/gmcookie >> $HOME/.dillo/cookies.txt And then I go to my bookmarked gmail page at http://mail.google.com/mail/?ui=html&zy=n and it all Just Works. If I get redirected to a https page, that indicates that the cookie is absent or invalid, so I put it back or get a fresh one. No guarantees that it will keep working, of course. The gmail folk can change their authentication requirements at any stage. And you miss out on the javascript goodness that is a large part of the gmail experience. And there was a stage when the html returned was broken in a way that dillo couldn't usefully handle, although that has since been changed back. But for straightforward mail tasks, it works well enough for me. Couple it with a lightweight gmail-notifier script, and you don't have to hit reload to see the new news. Good luck, f -- Francis Daly francis(a)daoine.org

Dec. 5, 2008

[patch]: Basic Authentication (updated)

by onepoint＠starurchin.org

New version, slightly amended as per previous comments. Changes: * Fixed a memory leak in Realm_add_path(). * Case-insensitive comparison in Host_from_url(). * Moved the Auth header to between Accept-Encoding and Host in a_Http_make_query_str(). Please comment, Jeremy Henty

Dec. 5, 2008

Dillo and Gmail

by corvid＠lavabit.com

Daniel wrote: > Does Dillo work with gmail? I haven't heard of it working for anyone.

Dec. 5, 2008

[patch]: Basic Authentication

by corvid＠lavabit.com

Jeremy wrote: > On Thu, Dec 04, 2008 at 05:59:21PM +0000, corvid wrote: > > > +/* > > > + * Return the host that contains a URL, or NULL if there is no such host. > > > + */ > > > +static AuthHost *Host_from_url(const DilloUrl *url) > > > +{ > > > + AuthHost *host; > > > + int i; > > > + > > > + for (i = 0; (host = dList_nth_data(auth_hosts, i)); i++) > > > + if (((strcmp(URL_SCHEME(url), host->scheme) == 0) && > > > + (strcmp(URL_AUTHORITY(url), host->authority) == 0))) > > > + return host; > > > + > > > + return NULL; > > > +} > > > > dStrcasecmp > > When comparing the scheme, the authority, or both? I think both. *pokes through rfc2396 for a bit* "When a scheme uses elements of the common syntax, it will also use the common syntax equivalence rules, namely that the scheme and hostname are case insensitive..."

Dec. 5, 2008

Dillo and Gmail

by daniel.victoria＠gmail.com

Hi all, First off, I got to say, congrats on the great browser. I have an old laptop and firefox just takes ages to load. Dillo, on the other hand, is lightning fast! Great! The only thing I'm missing is Gmail. Does Dillo work with gmail? Do I need to do any special tricks? I compiled the latest DIllo2 from svn and I did enabled https so I can get past the login page but then I'm stuck at a warning page about a non-standard redirect and, the link on that page does not work. The link is: https://www.google.com/accounts/CheckCookie?continue=http%3A%2F%2Fmail.goog… Thanks for the great work Daniel

Dec. 5, 2008

[patch]: Basic Authentication

by onepoint＠starurchin.org

On Thu, Dec 04, 2008 at 05:59:21PM +0000, corvid wrote: > Remember that there's a header ordering convention. D'oh, I saved the link to the discussion of this and then forget to review it! Will fix. > > +/* > > + * Return the host that contains a URL, or NULL if there is no such host. > > + */ > > +static AuthHost *Host_from_url(const DilloUrl *url) > > +{ > > + AuthHost *host; > > + int i; > > + > > + for (i = 0; (host = dList_nth_data(auth_hosts, i)); i++) > > + if (((strcmp(URL_SCHEME(url), host->scheme) == 0) && > > + (strcmp(URL_AUTHORITY(url), host->authority) == 0))) > > + return host; > > + > > + return NULL; > > +} > > dStrcasecmp When comparing the scheme, the authority, or both? > > +static void Realm_add_path(AuthRealm *realm, const char *path) > > +{ > > + int len, i; > > + char *realm_path; > > + > > + len = strlen(path); > > + /* ignore a trailing '/' */ > > + if (len && path[len - 1] == '/') > > + len--; > > + > > + /* delete existing paths that are inside the new one */ > > + for (i = 0; (realm_path = dList_nth_data(realm->paths, i)); i++) { > > + if (path_is_inside(realm_path, path, len)) { > > + dList_remove_fast(realm->paths, realm_path); > > + i--; /* reconsider this slot */ > > + } > > + } > > + > > + dList_append(realm->paths, dStrndup(path, len)); > > +} > > I think you need a dFree in there. For the realm_path that gets passed to dList_remove_fast()? Will fix. > PS I took the liberty of adding an entry for Basic auth to Plans.html No problem. Thanks for looking at this. Jeremy Henty

Dec. 4, 2008

[patch]: Basic Authentication

by corvid＠lavabit.com

Jeremy wrote: > ... I haven't looked into how anything fits together yet at all, but skimming through: > @@ -234,12 +236,13 @@ Dstr *a_Http_make_query_str(const DilloUrl *url, bool_t use_proxy) > "User-Agent: Dillo/%s\r\n" > "Content-Length: %ld\r\n" > "Content-Type: %s\r\n" > - "%s" > + "%s" /* cookies */ > + "%s" /* auth */ > "\r\n", > full_path->str, URL_AUTHORITY(url), > proxy_auth->str, referer, VERSION, > URL_DATA(url)->len, content_type->str, > - cookies); > + cookies, auth ? auth : ""); > dStr_append_l(query, URL_DATA(url)->str, URL_DATA(url)->len); > dStr_free(content_type, TRUE); > } else { Remember that there's a header ordering convention. > +/* > + * Return the host that contains a URL, or NULL if there is no such host. > + */ > +static AuthHost *Host_from_url(const DilloUrl *url) > +{ > + AuthHost *host; > + int i; > + > + for (i = 0; (host = dList_nth_data(auth_hosts, i)); i++) > + if (((strcmp(URL_SCHEME(url), host->scheme) == 0) && > + (strcmp(URL_AUTHORITY(url), host->authority) == 0))) > + return host; > + > + return NULL; > +} dStrcasecmp > +static void Realm_add_path(AuthRealm *realm, const char *path) > +{ > + int len, i; > + char *realm_path; > + > + len = strlen(path); > + /* ignore a trailing '/' */ > + if (len && path[len - 1] == '/') > + len--; > + > + /* delete existing paths that are inside the new one */ > + for (i = 0; (realm_path = dList_nth_data(realm->paths, i)); i++) { > + if (path_is_inside(realm_path, path, len)) { > + dList_remove_fast(realm->paths, realm_path); > + i--; /* reconsider this slot */ > + } > + } > + > + dList_append(realm->paths, dStrndup(path, len)); > +} I think you need a dFree in there. PS I took the liberty of adding an entry for Basic auth to Plans.html

Dec. 4, 2008

[patch]: Basic Authentication

by onepoint＠starurchin.org

On Thu, Dec 04, 2008 at 01:36:39PM +0100, Joerg Sonnenberger wrote: > Have you thought about how other protocols can be integrated with > this? No, except that I've deliberately made the User/Password dialog generic so that any other code can reuse it with a different callback. Other than that I think that all of my code is entirely specific to HTTP Basic authentication. I can't see how other protocols could be integrated with it in any useful sense. But my opinion doesn't count for much since I know next to nothing about how those other protocols work! > Examples are: > - TLS with client side authentication or verification for the server > certificate I have no idea how this works. Every time I try to study certificates and authentication my head starts spinning and I have to lie down. Does anyone know of some gentle introductions? > - non-anonymous FTP The FTP DPI could check the output of wget. If wget complained that the FTP server needed authentication than it could pop up the user/password dialog with a callback that retries the wget. The download DPI would have to be modified to receive the user and password. (Or we could stuff them into the download URL, but that's insecure since the password would appear in the process listing.) We would have to extend the DPI API to let DPIs use the user/password dialog but that is straightforward (I already implemented it when I thought authentication was going to be a DPI). Regards, Jeremy Henty

Dec. 4, 2008

css progress

by Johannes.Hofmann＠gmx.de

On Thu, Dec 04, 2008 at 08:13:49AM +0900, furaisanjin wrote: > Hi, > > > For now you could set force_my_colors=YES in dillorc. > > That wold be one way to avoid the problem but another issue is we > can't change dillo config on the fly, so I have to start dillo up if I > want to do so. If I keep force_my_colors=YES, dillo appearance looks > bad. That's true. I just checked your example page and it seems to look better with the current CSS prototype. The letters are visible - but still not readable for me ;-) Regards, Johannes

Dec. 4, 2008

[patch]: Basic Authentication

by joerg.sonnenberger＠web.de

On Thu, Dec 04, 2008 at 11:11:03AM +0000, Jeremy Henty wrote: > Opening the authentication dialog runs the FLTK event loop, which > means that incoming data can call the cache again, triggering CCC > warnings. We work around it by deferring the dialog to a timeout so > that the cache finishes iterating over the clients before the dialog > appears. This is annoying but I don't see a better way at the moment. Have you thought about how other protocols can be integrated with this? Examples are: - TLS with client side authentication or verification for the server certificate - non-anonymous FTP Joerg

Dec. 4, 2008