- Dillo-dev - lists.mailman3.com

No subject

by bogus＠does.not.exist.com

worked around the Window's issues with mouse copy & pasting. --- Roger http://rogerx.freeshell.org/

Oct. 21, 2011

use C locale for LC_CTYPE

by corvid＠lavabit.com

If you've been following along on fltk.general, you'll know we've been in discussion over some unpleasantness regarding Turkic locales and strcasecmp/toupper/tolower where 'i' and 'I' are different letters. ( http://en.wikipedia.org/wiki/Dotted_and_dotless_I ) The typical solution is to make sort of ASCIIfied versions where you force i<->I. I think that, for our purposes, we can just use the ordinary POSIX locale and save ourselves the trouble. To see this behaviour: export LC_ALL=tr_TR and then look at: <span style="font-size: 0.5IN">these letters should be big</span> <p> <I>This text should be italic</I>

Oct. 21, 2011

Address bar keys (bug #1012)

by obeythepenguin＠gmail.com

Rogutes Sparnuotos wrote: > Benjamin Johnson (2011-10-20 15:56):How should dillo behave when a user > binds Ctrl-A? It should perform the action bound to Ctrl-A, of course. Just to clarify, the only thing I want to change is the key bindings when the user is typing in the URL input box. Specifically, I'm talking about the CustInput::handle() function in ui.cc: http://hg.dillo.org/dillo/file/e82663b3c477/src/ui.cc#l96 Dillo intercepts several key bindings normally handled by the input box itself. The problem is, some of these key bindings conventionally have different meanings than Dillo assigns (see below). Not all the bindings are bad. I'd keep the binding for Ctrl-L, for example, which rehighlights the input text when you're typing in the box. That's pretty intuitive, and matches the behavior of other browsers. On the other hand, most users aren't familiar with Unix command line bindings, and even those who are usually don't expect that behavior in a graphical application such as Dillo. It might be convenient, but usability isn't always about convenience; it's about doing what users *expect*. Basically, a good interface is one an unwitting mere mortal could figure out without ever reading the manual, or even thinking too hard. Joel Spolsky explains it a lot better than I do: http://www.joelonsoftware.com/uibook/fog0000000249.html > > >Having it consistent over all text inputs and configurable via> > >keysrc sounds reasonable to me. I don't know the keysrc details> > >though, so I might miss something.> >Cheers,> >JohannesHaving > input/insert mode and macros in keysrc would be interesting, if nottoo > complex. It would be interesting, but it certainly would be complex, and I'm not sure the benefits outweigh the costs. I suspect most users are probably fine with just the basic key bindings FLTK already provides. There's also the matter of consistency, not just within Dillo, but between Dillo and other FLTK applications. Part of the advantage of using a toolkit *is* the consistency; that was the original motivation behind KDE, for example. > > I was thinking just a simple boolean option in dillorc -- and turn> it > OFF by default (i.e., no Unix keys), because it's not typical GUI> > behavior, and anyone who wants it on can quite easily set it in his> or > her own dillorc.Would be a strange dillorc option. A real patch would > implement this inkeysrc :) Hard-coding is much stranger than either of those, and much uglier, too. The dillorc option would actually make sense here, because these particular bindings usually go together as a group (it seems weird to do the Unix Ctrl-A, but not the others). Though it would make more sense still just to leave well enough alone, since there's no real need to change the input box's behavior. > Somehow I failed to notice that Ctrl-A for select all is "typical GUI > behavior", perhaps I am not using GUIs enough... > [...snip...] >Here, Ctrl-A goes to the beginning of the line in all the input boxes > ("find text", "web search", <input> in forms). Ctrl-A is conventionally used for select all, and at least on Windows, FLTK binds it to select all in input boxes. It may be that FLTK uses the Unix bindings (Ctrl-A for beginning of line, Ctrl-E for end, etc.) on those platforms, in which case Dillo's hard-coding that behavior is pointless and redundant. I haven't booted into Unix in a while, so offhand I haven't checked. Now, I know there will likely never be Windows support in mainline Dillo, and so you're probably wondering why Windows' key bindings should matter to you. While Unix doesn't have any formal standards for graphical user interface design -- indeed, the X Window System was deliberately designed that way -- most toolkits tend to follow what Windows does. Windows has explicit formal standards for command names, menu layouts, shortcut keys, etc., and not only are they standardized (at least on that platform), they're also well known since roughly 97% of desktop computer users run a Windows operating system. (Microsoft also helped design Motif, if anyone remembers that -- which is kind of weird for me to say, since I was roughly two years old at the time. Moving on...) For reference, here's FLTK's human interface guidelines, which borrow "common elements from the CDE/Motif, GNOME, KDE, MacOS, and Windows environments": http://www.fltk.org/hig.php And here's a list of standard accelerator keys from Qt's documentation, which explicitly recommends following Microsoft's recommendations: http://doc.qt.nokia.com/5.0-snapshot/accelerators.html To get back on topic, what does all this mean for Dillo? My own opinion is, the toolkit designers put a lot of thought into this stuff, and probably know more about how their widgets should behave than you and I do. So personally, I don't think we should override that behavior at all unless there's a really good reason to do so. > > Although I will add, the Shift-{Left,Right} bindings in the address> > bar *have* to go. They're fine elsewhere, if a bit quirky, but in> the > address bar they interfere with text selection. Ctrl-Tab and> > Ctrl-Shift-Tab are much more logical and convenient, anyway.> ~Benjamin > Are we using different versions of dillo? Shift-{Left,Right} doesn't do > anything here. I'm talking about this block of code right here: http://hg.dillo.org/dillo/file/e82663b3c477/src/ui.cc#l117 Pressing Shift-{Left,Right} in the address bar moves between tabs, rather than selecting text as expected. (Maybe something left over from dillo2?) > For the arguments sake, how would having to press Ctrl-Shift-Tab be more > "logical and convenient" than Shift-Left? It's a matter of taste (habit). > In my opinion, both bindings have a disadvantage: Ctrl-Shift-Tab is > difficult for the fingers, Shift-Left is unusual to most people. That's > why it is configurable, isn't it?---- Rogut s Sparnuotos Well, Ctrl-Tab's nice because of the obvious mnemonic; because it's really easy to type; and (you can probably guess what's coming) because it's familiar, since lots of other browsers use it. In contrast, the Shift-{Left,Right} keys are much harder to type, especially on a cramped laptop keyboard where your fingers are right on top of each other. And of course, it's a bit weird because I (and probably most users) associate those keys with "select text," not "change tabs." Basically, what all this boils down to is that consistency is key. Yes, it may be the hobgoblin of little minds, but when faced with an unfamiliar interface, *every* user has a little mind. So anyway, there's my rant for tonight. :-) Regards & cheers, ~Benjamin

Oct. 21, 2011

Address bar keys (bug #1012)

by rogutes＠googlemail.com

Benjamin Johnson (2011-10-20 15:56): > Johannes wrote: > > >On Wed, Oct 19, 2011 at 11:16:34AM -0400, Benjamin Johnson wrote: > >>Hi again, > >> > >>I was just wondering, could we add an option to control the address > >>bar key bindings? I know they're set to behave like a Unix command > >>line, which can be convenient, but it also interferes with some custom > >>key bindings (as one of my users mentioned in bug #1012). > >> > >>I've also noticed that the address bar bindings confuse a lot of > >>people, since they're not used to the Unix behavior. Conventionally, > >>Ctrl-A is "select all," and other input boxes in Dillo follow that, so > >>they aren't expecting it to go to the beginning of the line instead. How should dillo behave when a user binds Ctrl-A? > >Having it consistent over all text inputs and configurable via > >keysrc sounds reasonable to me. I don't know the keysrc details > >though, so I might miss something. > > > >Cheers, > >Johannes Having input/insert mode and macros in keysrc would be interesting, if not too complex. > I was thinking just a simple boolean option in dillorc -- and turn > it OFF by default (i.e., no Unix keys), because it's not typical GUI > behavior, and anyone who wants it on can quite easily set it in his > or her own dillorc. Would be a strange dillorc option. A real patch would implement this in keysrc :) Somehow I failed to notice that Ctrl-A for select all is "typical GUI behavior", perhaps I am not using GUIs enough... > Making it consistent globally would be a bit tricky, since we'd have > to replace all Fl_Inputs with our custom widget (including in FLTK's > stock dialogs). Besides that, there's the matter of other FLTK > applications -- would users expect them to behave our way, too? > > Personally, I'd just as soon ditch the Unix keys altogether, but if > we must keep them, let's let the address bar remain the exception. > That way, we're not breaking the behavior either way -- new users > aren't surprised, and old users can keep it as it was. Here, Ctrl-A goes to the beginning of the line in all the input boxes ("find text", "web search", <input> in forms). > Although I will add, the Shift-{Left,Right} bindings in the address > bar *have* to go. They're fine elsewhere, if a bit quirky, but in > the address bar they interfere with text selection. Ctrl-Tab and > Ctrl-Shift-Tab are much more logical and convenient, anyway. > > ~Benjamin Are we using different versions of dillo? Shift-{Left,Right} doesn't do anything here. For the arguments sake, how would having to press Ctrl-Shift-Tab be more "logical and convenient" than Shift-Left? It's a matter of taste (habit). In my opinion, both bindings have a disadvantage: Ctrl-Shift-Tab is difficult for the fingers, Shift-Left is unusual to most people. That's why it is configurable, isn't it? -- -- Rogut?s Sparnuotos

Oct. 20, 2011

view src

by obeythepenguin＠gmail.com

Jeremy wrote: > corvid wrote: > >> I viewed source, edited the page, reloaded, and view source keeps >> showing me the old source. >> >> I thought this used to work. > > It may have worked once but my impression is that it has been like > this for quite a while. I have certainly had a note about it in my > TO-DO list for a long time. > > Regards, > Jeremy Henty Funny, the old view source dialog doesn't have that problem at all... ;-) ~Benjamin

Oct. 20, 2011

Address bar keys (bug #1012)

by obeythepenguin＠gmail.com

Johannes wrote: > On Wed, Oct 19, 2011 at 11:16:34AM -0400, Benjamin Johnson wrote: >> Hi again, >> >> I was just wondering, could we add an option to control the address >> bar key bindings? I know they're set to behave like a Unix command >> line, which can be convenient, but it also interferes with some custom >> key bindings (as one of my users mentioned in bug #1012). >> >> I've also noticed that the address bar bindings confuse a lot of >> people, since they're not used to the Unix behavior. Conventionally, >> Ctrl-A is "select all," and other input boxes in Dillo follow that, so >> they aren't expecting it to go to the beginning of the line instead. > > Having it consistent over all text inputs and configurable via > keysrc sounds reasonable to me. I don't know the keysrc details > though, so I might miss something. > > Cheers, > Johannes I was thinking just a simple boolean option in dillorc -- and turn it OFF by default (i.e., no Unix keys), because it's not typical GUI behavior, and anyone who wants it on can quite easily set it in his or her own dillorc. Making it consistent globally would be a bit tricky, since we'd have to replace all Fl_Inputs with our custom widget (including in FLTK's stock dialogs). Besides that, there's the matter of other FLTK applications -- would users expect them to behave our way, too? Personally, I'd just as soon ditch the Unix keys altogether, but if we must keep them, let's let the address bar remain the exception. That way, we're not breaking the behavior either way -- new users aren't surprised, and old users can keep it as it was. Although I will add, the Shift-{Left,Right} bindings in the address bar *have* to go. They're fine elsewhere, if a bit quirky, but in the address bar they interfere with text selection. Ctrl-Tab and Ctrl-Shift-Tab are much more logical and convenient, anyway. ~Benjamin

Oct. 20, 2011

No subject

by bogus＠does.not.exist.com

worked around the Window's issues with mouse copy & pasting. --- Roger http://rogerx.freeshell.org/

Oct. 20, 2011

view src

by corvid＠lavabit.com

Jorge wrote: > A quick trick to avoid viewing source of the "view source" page. > A proper solution would be appreciatted. Does that break things, or do you just dislike it?

Oct. 20, 2011

Dillo 3 on OpenBSD

by obeythepenguin＠gmail.com

Just wanted to let you know that dillo-3.0 runs quite nicely on OpenBSD (tested with both mainline and Dillo-Win32, on OpenBSD 4.9/i386). Here are some quick build instructions for mainline if anyone wants to try it out. Note that I like to install things to my home directory, but you can remove the --prefix= and --bindir= options if you want to install it system-wide. 1. Download the Dillo and FLTK 1.3 sources to a convenient location -- I use ~/download. Note that OpenBSD doesn't currently have an FLTK 1.3 package. 2. Install some dependencies from OpenBSD's packages: $ sudo pkg_add -i bzip2 jpeg png 3. Build and install FLTK: $ cd /tmp $ tar -xzvf ~/download/fltk-1.3.0-source.tar.gz $ cd fltk-1.3.0 $ CFLAGS='-Os' CXXFLAGS='-Os' ./configure \ --prefix=$HOME/download/fltk --bindir=$HOME/bin --enable-xft $ make && make install 4. Build and install Dillo: $ cd /tmp $ tar -xjvf ~/download/dillo-*.tar.bz2 $ cd dillo-* $ CFLAGS='-Os' CXXFLAGS='-Os' ./configure \ --prefix=$HOME/download/dillo --bindir=$HOME/bin \ --disable-threaded-dns $ make && make install && strip $HOME/bin/dillo Dillo should be installed and ready to go. By the way, I don't know if you *need* --disable-threaded-dns in step 4. I usually add it, and I'm sure there's a reason, but I don't remember exactly why. So if you try it, let me know if Dillo runs without that option. :-) Cheers, ~Benjamin

Oct. 20, 2011

view src

by jcid＠dillo.org

On Wed, Oct 19, 2011 at 08:47:28PM +0000, corvid wrote: > Jeremy wrote: > > corvid wrote: > > > > > I viewed source, edited the page, reloaded, and view source keeps > > > showing me the old source. > > > > > > I thought this used to work. > > > > It may have worked once but my impression is that it has been like > > this for quite a while. I have certainly had a note about it in my > > TO-DO list for a long time. > > It mustn't have ever worked, since a_Capi_open_url() says > > } else if (Capi_url_uses_dpi(web->url, &server)) { > /* dpi request */ > if ((safe = a_Capi_dpi_verify_request(web->bw, web->url))) { > if (dStrcasecmp(scheme, "dpi") == 0) { > if (strcmp(server, "vsource") == 0) { > /* don't reload the "view source" page */ > } else { > > for some reason. A quick trick to avoid viewing source of the "view source" page. A proper solution would be appreciatted. -- Cheers Jorge.-

Oct. 20, 2011