- Dillo-dev - lists.mailman3.com

css and font specification [Re: Dillo2 + FLTK2 + NxLib running on mips linux]

by Johannes.Hofmann＠gmx.de

On Fri, May 08, 2009 at 07:59:43PM +0000, corvid wrote: > Johannes wrote: > > Hm, this seems to work fine for me. Maybe you have other stuff in > > your style.css that confuses the CSS parser? > > Yes, it turns out to have been an error in my style.css. > The previous rule was commented out, but I used html > comment delimiters instead of css comment delimiters. Sigh... > Sorry about that. All is well now. Not quite.... Html comment delimiters should be completely ignored by the CSS scanner/parser. With the re2c implementation that did work, but it is missing from the current scanner.

May 9, 2009

image memory leak

by jcid＠dillo.org

On Thu, Apr 23, 2009 at 06:01:56PM +0200, Hofmann Johannes wrote: > Hi, > > I did some more tests regarding that image related memory leak, > Jeremy reported some time ago. > It turns out that memory related to aborted images is not released. > With CSS and the related repushs this seems to happen frequently. > But I can also trigger it without CSS by hitting Ctrl-R while > loading a large image. > Attached diagnostic patch shows the issue. Load a page with a single > large JPEG image and hit Ctrl-R while it is loading. > In that case only the second instance of DilloJpeg is deleted. > > I tried to figure out, why the first cache client is not getting > aborted properly but the cache related code is quite complex... > Does anyone know how this should work? The repository has a patch for this bug. It has the debug messages for easy testing. I've given it some stress but would like to see how it works in the outside world. Please give it some testing. -- Cheers Jorge.-

May 8, 2009

css and font specification [Re: Dillo2 + FLTK2 + NxLib running on mips linux]

by corvid＠lavabit.com

Johannes wrote: > Hm, this seems to work fine for me. Maybe you have other stuff in > your style.css that confuses the CSS parser? Yes, it turns out to have been an error in my style.css. The previous rule was commented out, but I used html comment delimiters instead of css comment delimiters. Sigh... Sorry about that. All is well now.

May 8, 2009

whitespace entities inside words

by corvid＠lavabit.com

Jorge wrote: > On Fri, May 08, 2009 at 03:53:03AM +0000, corvid wrote: > > Is a whitespace entity inside a word supposed to be > > treated like normal whitespace? > > > > I thought it would be, but first I saw that > > adjacent   ... aren't collapsed, > > then I realized that we are leaving the > > space inside the word instead of calling > > addSpace, and then I saw that PrevWasSPC > > isn't set if the space was at the end of the > > word. > > > > If it's just something that no one ever got > > around to dealing with, I'll make a patch. > > That hits into a deeper problem: the treatment of > whitespace is not well defined inside the textblock yet. > The same problem arises when handling NBSP. There's no > clear definition on how to handle/flag it. > If you're willing to study the textblock word wrapping, > and the whitespace problem, great! I think the next big > step for dillo is to handle DIV as a floating word (so it > can be rendered side by side, as intended), and AFAIS that > needs changes in the textblock. > This is a task for 2.2. I have some Textblock code around to fix "white-space: nowrap" (which shows up a lot on wikipedia, where the category stuff at the bottoms of pages is displayed badly) that I've been saving for discussion after this release. I have absolutely no clue about floating, though, and have been hoping that Johannes will develop an interest in it :)

May 8, 2009

whitespace entities inside words

by jcid＠dillo.org

On Fri, May 08, 2009 at 03:53:03AM +0000, corvid wrote: > Is a whitespace entity inside a word supposed to be > treated like normal whitespace? > > I thought it would be, but first I saw that > adjacent   ... aren't collapsed, > then I realized that we are leaving the > space inside the word instead of calling > addSpace, and then I saw that PrevWasSPC > isn't set if the space was at the end of the > word. > > If it's just something that no one ever got > around to dealing with, I'll make a patch. That hits into a deeper problem: the treatment of whitespace is not well defined inside the textblock yet. The same problem arises when handling NBSP. There's no clear definition on how to handle/flag it. If you're willing to study the textblock word wrapping, and the whitespace problem, great! I think the next big step for dillo is to handle DIV as a floating word (so it can be rendered side by side, as intended), and AFAIS that needs changes in the textblock. This is a task for 2.2. -- Cheers Jorge.-

May 8, 2009

css and font specification [Re: Dillo2 + FLTK2 + NxLib running on mips linux]

by Johannes.Hofmann＠gmx.de

On Fri, May 08, 2009 at 03:27:56PM +0000, corvid wrote: > ----- Forwarded message from ?? ? ----- > > I have try current version dillo + fltk base on X11 in X86 PC (Ubuntu) > > with LC_LANG=zh_TW.UTF-8. > > > > Using default font_xxxx I can see chinese text and write chinese characters to text editor also works. > > > > But, I cannot see any chinese text after google search ??? > > The same URL works on dillo + fltk base on nxlib. > > > > What I think is the FLTK already supports multi bytes characters well. > > The problem is caused by others (I mean fonts / X11 configuration etc.) > > Am I right ? > > (This came with screenshots showing that www.google.com.tw doesn't work.) > > If dillo doesn't parse embedded css, it works for me. > The "body,td,a,p,.h{font-family:arial,sans-serif}" in the page overrides > my font_* prefs. > So I tried to add a !important rule to set font-family in style.css, e.g. > * {font-family: "Bitstream Cyberbit" !important} Hm, this seems to work fine for me. Maybe you have other stuff in your style.css that confuses the CSS parser? > but either it's impossible or I was getting the syntax wrong somehow. Looks ok to me and works here. What does xlsfonts | grep -i cyberbit say? Cheers, Johannes

May 8, 2009

css and font specification [Re: Dillo2 + FLTK2 + NxLib running on mips linux]

by corvid＠lavabit.com

----- Forwarded message from ?? ? ----- > I have try current version dillo + fltk base on X11 in X86 PC (Ubuntu) > with LC_LANG=zh_TW.UTF-8. > > Using default font_xxxx I can see chinese text and write chinese characters to text editor also works. > > But, I cannot see any chinese text after google search ??? > The same URL works on dillo + fltk base on nxlib. > > What I think is the FLTK already supports multi bytes characters well. > The problem is caused by others (I mean fonts / X11 configuration etc.) > Am I right ? (This came with screenshots showing that www.google.com.tw doesn't work.) If dillo doesn't parse embedded css, it works for me. The "body,td,a,p,.h{font-family:arial,sans-serif}" in the page overrides my font_* prefs. So I tried to add a !important rule to set font-family in style.css, e.g. * {font-family: "Bitstream Cyberbit" !important} but either it's impossible or I was getting the syntax wrong somehow. (This is not necessarily the same problem he's having, but...)

May 8, 2009

Image invalid access problems

by onepoint＠starurchin.org

It's too early to be absolutely certain, but from the logging data I've been collecting it seems that invalid DilloImage accesses always come from this call sequence: a_Bw_stop_clients() -> a_Capi_stop_client() -> a_Cache_stop_client() -> Cache_client_dequeue() -> a_Web_free() -> a_Image_unref() -> Image_free() -> dFree() So the question is: why isn't this call sequence stopping the image callbacks from running before it dFree()s the image? When Cache_process_queue() sees that all the data has arrived for a client it executes "(Client->Callback)(CA_Close, Client);" before calling Cache_client_dequeue(). Perhaps the fix is for a_Cache_stop_client() to close the client too? Regards, Jeremy Henty PS: Interestingly, a_Bw_stop_clients() has a comment: /* * Stop the active clients of this bw's top page. * Note: rendering stops, but the cache continues to be fed. */ This seems to be talking about our problem: the cache is still running and therefore accessing the client, but the client has been freed.

May 8, 2009

[patch]: replace DW2TB with HT2TB

by onepoint＠starurchin.org

I noticed that the convenience macro DW2TB is always used as DW2TB(html->dw) so it makes sense to use an even more convenient macro HT2TB to reduce DW2TB(html->dw) to HT2TB(html). Patch attached. Regards, Jeremy Henty

May 8, 2009

whitespace entities inside words

by corvid＠lavabit.com

Is a whitespace entity inside a word supposed to be treated like normal whitespace? I thought it would be, but first I saw that adjacent   ... aren't collapsed, then I realized that we are leaving the space inside the word instead of calling addSpace, and then I saw that PrevWasSPC isn't set if the space was at the end of the word. If it's just something that no one ever got around to dealing with, I'll make a patch.

May 8, 2009

change FLTK scheme for in dillo2

by akhilman＠gmail.com

?? Fri, 8 May 2009 00:37:50 +0200 Johannes Hofmann <Johannes.Hofmann(a)gmx.de> ????????: > What do people think about reading X resources as in attached patch > (mainly copied from commented fltk2 code)? > It allows to specify e.g. > > dillo.background:red > dillo.foreground:white > > in your .Xdefaults Will be great for me. I'm using .Xdefaults to tweak colour scheme for urxvt and other programs such as xpdf and puredata. -- ?????? ??????????? aka AkhIL ??? ??? 8 09:24:47 KRAST 2009 Fri May 8 01:24:47 UTC 2009 ---------------------------------- Visit my home page http://akhilman.blogspot.com/ jabber: akhil(a)jabber.ru ---------------------------------- ??? ????????? ????????? ????????? ??, ??? ?? ??????? ?????????, ?????? ?? ???? ????, ??? ?? ???????????? ??? ?????????. (?.?????????) ---------------------------------- Linux artstation 2.6.28-gentoo-r5 #1 SMP Sat May 2 18:40:24 KRAST 2009 x86_64 AMD Phenom(tm) 9550 Quad-Core Processor AuthenticAMD GNU/Linux up 5 days, 14:19, 4 users, load average: 0.43, 0.50, 0.52

May 8, 2009

change FLTK scheme for in dillo2

by Johannes.Hofmann＠gmx.de

On Thu, May 07, 2009 at 03:44:19PM +0000, corvid wrote: > IL'dar wrote: > > Can I change colours of dialogues and scrollbar in dillo2? > > I would like to gets something less bright. > > Even something like fluid2's -bg option will be ok. > > Not currently. > > I would not necessarily be opposed to making ui_fg and ui_bg prefs and > adding a bunch of [class]::default_style->[whatever]color() > to dillo.cc, although I wouldn't be surprised if we ran into our > share of places where fltk hardcodes things to, say, GRAY75. What do people think about reading X resources as in attached patch (mainly copied from commented fltk2 code)? It allows to specify e.g. dillo.background:red dillo.foreground:white in your .Xdefaults Cheers, Johannes

May 8, 2009

Applause: image resizing fix

by onepoint＠starurchin.org

I spend a lot of time surfing the DoomWorld picture viewer[1] and so I'm used to the thumbnails at the bottom displaying random small chunks of the image instead of anything useful. But now I'm seeing proper previews, just as the site designers intended. It's *really* nice to see this working properly! Dillo just keeps getting better! Regards, Jeremy Henty [1]: eg. http://www.doomworld.com/php/screenie.php?dir=/images/newstuff/341/&filenam…

May 8, 2009

[patch]: more cleanups

by jcid＠dillo.org

On Mon, May 04, 2009 at 06:59:34AM +0100, Jeremy Henty wrote: > > Three more cleanup patches: > > 1: Inline Cache_client_make_key() into Cache_client_enqueue(). Why > have a separate function just to increment a variable? It's just > as clear (and less noisy) to increment a static variable inside > Cache_client_enqueue(). Committed. > 2: The width and height arguments of a_Image_new() do nothing. The > function itself ignores them and its callers always set them to 0. > Since neither the function or its callers seem to care what these > arguments are, why not remove them? Committed. > 3: a_Web_dispatch_by_type() has a strange convention for its return > type, it returns -1 to indicate failure. It's less confusing to > use the common convention of a boolean: 1 indicates success and 0 > indicates failure. It's not strange. Just like libc: 0 success, < 0 error, > 0 success with return code. -- Cheers Jorge.-

May 7, 2009

[Patch 3/4] Key bindings

by rogutes＠googlemail.com

Johannes Hofmann (2009-05-07 21:17): > On Thu, May 07, 2009 at 06:02:08PM +0000, corvid wrote: > > Rogut?s wrote: > > > And one more suggestion: wouldn't it be more intuitive to have > > > "command = keybinding" instead of "keybinding = command"? > > > > *agrees* > > Well in muttrc you have e.g: > > bind pager <Down> next-line > > So it's rather "keybinding = command". > Perhaps the '=' is a bit confusing. Looks like you're right. And shell's bindkey expects a key first (as well as vim's :map). Intuitive is the notion that the shorter part should be supplied first. Currently, dillo's commands look shorter/cleaner than keybindings. And it is the other way around with mutt's and vim's complex (more evolved?) commands... -- -- Rogut?s Sparnuotos

May 7, 2009

[Patch 3/4] Key bindings

by Johannes.Hofmann＠gmx.de

On Thu, May 07, 2009 at 06:02:08PM +0000, corvid wrote: > Rogut?s wrote: > > And one more suggestion: wouldn't it be more intuitive to have > > "command = keybinding" instead of "keybinding = command"? > > *agrees* Well in muttrc you have e.g: bind pager <Down> next-line So it's rather "keybinding = command". Perhaps the '=' is a bit confusing. Cheers, Johannes

May 7, 2009

Re: Dillo2 + FLTK2 + NxLib running on mips linux

by stevegigijoe＠yahoo.com.tw

>On Thu, May 07, 2009 at 02:07:22AM +0800, ???? ?i wrote:=0A>> Hello All=0A= >>=0A>> One years ago I had tried dillo + fltk + nxlib and unsuccessful.=0A= >> Today, the dillo2 works. I'm so happy.=0A>>=0A>> Also I have record my b= uilding progress in my blog=0A>>http://tw.myblog.yahoo.com/stevegigijoe/art= icle?mid=3D138&prev=3D-1&next=3D137=0A>>=0A>> Thank you all.=0A=0A> Good! = =0A> Do you have a screenshot?=0A=0AScreenshot=0A=0A=0Ahttp://f23.yahoofs.= com/myper/824eoFyBBBKXGuU1zjFOFp3xJAY-/blog/ap_F23_20090506105626866.jpg?TT= AeMBKBHJ2Wr2zc=0A=0Ahttp://f23.yahoofs.com/myper/824eoFyBBBKXGuU1zjFOFp3xJA= Y-/blog/ap_F23_20090506092752201.jpg?TTAeMBKBlaX8P12A=0A=0Ahttp://f23.yahoo= fs.com/myper/824eoFyBBBKXGuU1zjFOFp3xJAY-/blog/ap_F23_20090506111231234.jpg= ?TTAeMBKB7IcMpZy_=0A=0ASteve=0A=0A=0A ________________________________= ______________________________________________________________________=0A= =A5I=B6O=A4~=AEe=B6q=B5L=A4W=AD=AD=A1HYahoo!=A9_=BC=AF=B9q=A4l=ABH=BDc2.0= =A7K=B6O=B5=B9=A7A=A1A=ABH=A5=F3=A5=C3=BB=B7=A4=A3=A5=B2=A7R=A1I http://tw.= mg0.mail.yahoo.com/dc/landing

May 7, 2009

[Patch 3/4] Key bindings

by rogutes＠googlemail.com

Re-sending, because the patch was missing some spaces. corvid (2009-05-03 23:01): > Jorge wrote: > > The patch is not 100% complete. For instance there's > > a need to provide a default keysrc file in a similar way > > that cookiesrc or dillorc (with useful comments). > > Started looking into making a default keysrc. > > Some of the bindings in keys.cc are not the same as those in > dillo2-help.html. Is that intentional? > > What is the purpose of KEYS_HIDE_PANELS when we have KEYS_FULLSCREEN? Hello, I'm also wondering why KEYS_HIDE_PANELS is there, since KEYS_FULLSCREEN works fine. It seems to have been merged by Jorge together with the configurable keybindings patch, but I think the functionality was actually proposed by Tim. So I'm bumping this thread and also attaching a small patch, which fixes an omission during the configurable keys merge. The following code was removed from ui.cc, and I think it should be added to keys.cc (currently, no keybindings work if numlock is on): - // We're only interested in some flags - unsigned modifier = event_state() & (SHIFT | CTRL | ALT); And one more question: why are the Button1, Button2, Button3 modifierNames there? And one more suggestion: wouldn't it be more intuitive to have "command = keybinding" instead of "keybinding = command"? Now the keysrc file will look like this: Esc = hide-panels <Ctrl>a = hide-panels , = back Backspace = back . = forward <Shift>Backspace = forward while it could be: hide-panels = Esc hide-panels = <Ctrl>a back = , back = Backspace forward = . forward = <Shift>Backspace -- -- Rogut?s Sparnuotos

May 7, 2009

BUMP: Re: [Dillo-dev] [patch]: more cleanups

by onepoint＠starurchin.org

What is the verdict on these? My stack of local patches is getting uncomfortably long, so decisions either way are welcome so I can shrink it. Regards, Jeremy Henty On Mon, May 04, 2009 at 06:59:34AM +0100, Jeremy Henty wrote: > > Three more cleanup patches: > > 1: Inline Cache_client_make_key() into Cache_client_enqueue(). Why > have a separate function just to increment a variable? It's just > as clear (and less noisy) to increment a static variable inside > Cache_client_enqueue(). > > 2: The width and height arguments of a_Image_new() do nothing. The > function itself ignores them and its callers always set them to 0. > Since neither the function or its callers seem to care what these > arguments are, why not remove them? > > 3: a_Web_dispatch_by_type() has a strange convention for its return > type, it returns -1 to indicate failure. It's less confusing to > use the common convention of a boolean: 1 indicates success and 0 > indicates failure. > > Regards, > > Jeremy Henty > diff -r 817ab0ff08a1 -r c671a697420f src/cache.c > --- a/src/cache.c Sat May 02 22:57:36 2009 +0100 > +++ b/src/cache.c Sat May 02 23:27:14 2009 +0100 > @@ -136,18 +136,6 @@ > /* Client operations ------------------------------------------------------ */ > > /* > - * Make a unique primary-key for cache clients > - */ > -static int Cache_client_make_key(void) > -{ > - static int ClientKey = 0; /* Provide a primary key for each client */ > - > - if (++ClientKey <= 0) > - ClientKey = 1; > - return ClientKey; > -} > - > -/* > * Add a client to ClientQueue. > * - Every client-field is just a reference (except 'Web'). > * - Return a unique number for identifying the client. > @@ -155,11 +143,13 @@ > static int Cache_client_enqueue(const DilloUrl *Url, DilloWeb *Web, > CA_Callback_t Callback, void *CbData) > { > - int ClientKey; > + static int ClientKey = 0; /* Provide a primary key for each client */ > CacheClient_t *NewClient; > > + if (++ClientKey <= 0) > + ClientKey = 1; > + > NewClient = dNew(CacheClient_t, 1); > - ClientKey = Cache_client_make_key(); > NewClient->Key = ClientKey; > NewClient->Url = Url; > NewClient->Version = 0; > diff -r 4ab60f74c46b src/dicache.c > --- a/src/dicache.c Sun May 03 15:29:42 2009 +0100 > +++ b/src/dicache.c Sun May 03 15:32:34 2009 +0100 > @@ -398,7 +398,7 @@ > dReturn_val_if_fail(MimeType && Ptr, NULL); > > if (!web->Image) > - web->Image = a_Image_new(0, 0, NULL, web->bgColor); > + web->Image = a_Image_new(NULL, web->bgColor); > > /* Add an extra reference to the Image (for dicache usage) */ > a_Image_ref(web->Image); > diff -r 4ab60f74c46b src/html.cc > --- a/src/html.cc Sun May 03 15:29:42 2009 +0100 > +++ b/src/html.cc Sun May 03 15:32:34 2009 +0100 > @@ -2062,7 +2062,7 @@ > html->styleEngine->setNonCssHints(&props); > > /* Add a new image widget to this page */ > - Image = a_Image_new(0, 0, alt_ptr, 0); > + Image = a_Image_new(alt_ptr, 0); > if (DW2TB(html->dw)->getBgColor()) > Image->bg_color = DW2TB(html->dw)->getBgColor()->getColor(); > > diff -r 4ab60f74c46b src/image.cc > --- a/src/image.cc Sun May 03 15:29:42 2009 +0100 > +++ b/src/image.cc Sun May 03 15:32:34 2009 +0100 > @@ -33,10 +33,7 @@ > /* > * Create and initialize a new image structure. > */ > -DilloImage *a_Image_new(int width, > - int height, > - const char *alt_text, > - int32_t bg_color) > +DilloImage *a_Image_new(const char *alt_text, int32_t bg_color) > { > DilloImage *Image; > > diff -r 4ab60f74c46b src/image.hh > --- a/src/image.hh Sun May 03 15:29:42 2009 +0100 > +++ b/src/image.hh Sun May 03 15:32:34 2009 +0100 > @@ -50,8 +50,7 @@ > /* > * Function prototypes > */ > -DilloImage *a_Image_new(int width, int height, > - const char *alt_text, int32_t bg_color); > +DilloImage *a_Image_new(const char *alt_text, int32_t bg_color); > void a_Image_ref(DilloImage *Image); > void a_Image_unref(DilloImage *Image); > > diff -r 02d822ae435a src/cache.c > --- a/src/cache.c Sat May 02 23:38:41 2009 +0100 > +++ b/src/cache.c Sat May 02 23:42:40 2009 +0100 > @@ -1112,7 +1112,7 @@ > const char *curr_type = Cache_current_content_type(entry); > st = a_Web_dispatch_by_type(curr_type, ClientWeb, > &Client->Callback, &Client->CbData); > - if (st == -1) { > + if (!st) { > /* MIME type is not viewable */ > if (ClientWeb->flags & WEB_RootUrl) { > MSG("Content-Type '%s' not viewable.\n", curr_type); > diff -r 02d822ae435a src/web.cc > --- a/src/web.cc Sat May 02 23:38:41 2009 +0100 > +++ b/src/web.cc Sat May 02 23:42:40 2009 +0100 > @@ -47,8 +47,8 @@ > * Given the MIME content type, and a fd to read it from, > * this function connects the proper MIME viewer to it. > * Return value: > - * 1 on success (and Call and Data properly set). > - * -1 for unhandled MIME types (and Call and Data untouched). > + * 1 on success (and Call and Data properly set). > + * 0 for unhandled MIME types (and Call and Data untouched). > */ > int a_Web_dispatch_by_type (const char *Type, DilloWeb *Web, > CA_Callback_t *Call, void **Data) > @@ -57,7 +57,7 @@ > > _MSG("a_Web_dispatch_by_type\n"); > > - dReturn_val_if_fail(Web->bw != NULL, -1); > + dReturn_val_if_fail(Web->bw != NULL, 0); > > // get the Layout object from the bw structure. > Layout *layout = (Layout*)Web->bw->render_layout; > @@ -71,8 +71,8 @@ > Web->bgColor= styleEngine.backgroundStyle()->backgroundColor->getColor(); > > dw = (Widget*) a_Mime_set_viewer(Type, Web, Call, Data); > - if (dw == NULL) > - return -1; > + if (!dw) > + return 0; > > dw->setStyle (styleEngine.style ()); > > @@ -100,7 +100,7 @@ > if (!dw) { > MSG_HTTP("unhandled MIME type: \"%s\"\n", Type); > } > - return (dw ? 1 : -1); > + return dw ? 1 : 0; > } > > > diff -r 02d822ae435a src/web.hh > --- a/src/web.hh Sat May 02 23:38:41 2009 +0100 > +++ b/src/web.hh Sat May 02 23:42:40 2009 +0100 > @@ -38,7 +38,7 @@ > int a_Web_valid(DilloWeb *web); > void a_Web_free (DilloWeb*); > int a_Web_dispatch_by_type (const char *Type, DilloWeb *web, > - CA_Callback_t *Call, void **Data); > + CA_Callback_t *Call, void **Data); > > #ifdef __cplusplus > } > _______________________________________________ > Dillo-dev mailing list > Dillo-dev(a)dillo.org > http://lists.auriga.wearlab.de/cgi-bin/mailman/listinfo/dillo-dev

May 7, 2009

[Patch 3/4] Key bindings

by corvid＠lavabit.com

Rogut?s wrote: > And one more suggestion: wouldn't it be more intuitive to have > "command = keybinding" instead of "keybinding = command"? *agrees*

May 7, 2009

[Patch 3/4] Key bindings

by rogutes＠googlemail.com

corvid (2009-05-03 23:01): > Jorge wrote: > > The patch is not 100% complete. For instance there's > > a need to provide a default keysrc file in a similar way > > that cookiesrc or dillorc (with useful comments). > > Started looking into making a default keysrc. > > Some of the bindings in keys.cc are not the same as those in > dillo2-help.html. Is that intentional? > > What is the purpose of KEYS_HIDE_PANELS when we have KEYS_FULLSCREEN? Hello, I'm also wondering why KEYS_HIDE_PANELS is there, since KEYS_FULLSCREEN works fine. It seems to have been merged by Jorge together with the configurable keybindings patch, but I think the functionality was actually proposed by Tim. So I'm bumping this thread and also attaching a small patch, which fixes an omission during the configurable keys merge. The following code was removed from ui.cc, and I think it should be added to keys.cc (currently, no keybindings work if numlock is on): - // We're only interested in some flags - unsigned modifier = event_state() & (SHIFT | CTRL | ALT); And one more question: why are the Button1, Button2, Button3 modifierNames there? And one more suggestion: wouldn't it be more intuitive to have "command = keybinding" instead of "keybinding = command"? Now the keysrc file will look like this: Esc = hide-panels <Ctrl>a = hide-panels , = back Backspace = back . = forward <Shift>Backspace = forward while it could be: hide-panels = Esc hide-panels = <Ctrl>a back = , back = Backspace forward = . forward = <Shift>Backspace -- -- Rogut?s Sparnuotos

May 7, 2009

image memory leak

by onepoint＠starurchin.org

You should note that I have applied this patch: http://starurchin.org/dillo/patch/img_ref.diff.txt I list all my local patches at http://starurchin.org/dillo/valgrind.html . Only img_ref.diff.txt affects how DilloImages are managed. Regards, Jeremy Henty

May 7, 2009

image memory leak

by onepoint＠starurchin.org

On Thu, May 07, 2009 at 11:43:36AM -0400, Jorge Arellano Cid wrote: > Have you found a simplified test-case to reproduce it? Unfortunately no. Here's what I know from good-old printf() debugging. It's definitely the case that sometimes an image callback is called with a freed DilloImage. In fact this happens a lot, much more often than the valgrind logs suggest, so I think it doesn't always cause a bad memory access. Mostly the callback is only called two extra times after the free, but sometimes it is called many more times than that. The only way this could happen is if the DilloImage is put in a DilloWeb and then Cache_client_dequeue() is called without the CCC being shut down. Cache_client_dequeue() is called at various places in cache.c but the whenever the problem appears it turns out that it was called from a_Cache_stop_client(). I have not yet traced the stack further back. I'll do that next and let you know what it shows. Regards, Jeremy Henty

May 7, 2009

Dillo2 + FLTK2 + NxLib running on mips linux

by corvid＠lavabit.com

furaisanjin wrote: > 2009/5/7 corvid <corvid(a)lavabit.com>: > > Wouldn't it be wrong to wrap after most multibyte chars? > > (e.g., if you go to uk.wikipedia.org) > > No, but I thoght some people might concern the cost of utf8decode and > more memory allocation/deallocation due to this patch. Even in English pages it's very common to use some multibyte chars. Spanish, German, etc., of course have accented chars that use two bytes in UTF8. As for whether the code should be optimized or a pref needed, I never tried profiling it. Anyway, it should be moved into Html_process_word...

May 7, 2009

change FLTK scheme for in dillo2

by corvid＠lavabit.com

IL'dar wrote: > Can I change colours of dialogues and scrollbar in dillo2? > I would like to gets something less bright. > Even something like fluid2's -bg option will be ok. Not currently. I would not necessarily be opposed to making ui_fg and ui_bg prefs and adding a bunch of [class]::default_style->[whatever]color() to dillo.cc, although I wouldn't be surprised if we ran into our share of places where fltk hardcodes things to, say, GRAY75.

May 7, 2009

image memory leak

by jcid＠dillo.org

On Sat, Apr 25, 2009 at 01:18:19PM +0100, Jeremy Henty wrote: > On Fri, Apr 24, 2009 at 06:20:20PM +0200, Hofmann Johannes wrote: > > On Thu, Apr 23, 2009 at 09:36:23PM +0100, Jeremy Henty wrote: > > > > > > On Thu, Apr 23, 2009 at 06:01:56PM +0200, Hofmann Johannes wrote: > > > > > > > It turns out that memory related to aborted images is not released. > > > > > > What about the other way round, ie. memory being released and > > > then accessed? If you check the invalid read errors at the > > > valgrind logs page[1] for the past few days you will see they are > > > all due to reading freed image memory. Is this another aspect of > > > the same thing, or something else? > > > > I'd say it's something different. It's always a read of size 4, > > isn't it? I guess that's just one uninitialized pointer. > > I don't think that it's uninitialized pointers. If it were then you > would expect valgrind to complain that the address had never been > malloced or freed. That's not what we see. Eg. look at the first > instance of 'Jpeg_write: Invalid read of size 4'[1]. The freed memory > is a DilloImage structure. The line that does the invalid read is > reading ... a member of a DilloImage structure! The same is true for > the errors from dw::core::Widget::getLayout() , Png_datarow_callback > and Png_datainfo_callback . In every case the pointer refers to a > malloced structure of the right type. The problem is that the > structure has since been freed. I think it's pretty clear that the > problem is not uninitialised pointers, it's that the code is using > pointers after freeing them. > > It's probably significant that these errors all come from within image > callbacks (a_Jpeg_callback and a_Png_callback). It looks like a race > condition to me: callbacks are running after the image has been freed. > Does that make sense? Yes, it makes sense to me. Have you found a simplified test-case to reproduce it? Finally now I have some time to investigate this bug. I'll start with the problem Johannes described (large jpeg) because of the test-case. Let's start... -- Cheers Jorge.-

May 7, 2009

Dillo2 + FLTK2 + NxLib running on mips linux

by furaisanjin＠gmail.com

2009/5/7 corvid <corvid(a)lavabit.com>: > Wouldn't it be wrong to wrap after most multibyte chars? > (e.g., if you go to uk.wikipedia.org) No, but I thoght some people might concern the cost of utf8decode and more memory allocation/deallocation due to this patch. Regards, furaisanjin

May 7, 2009

Dillo2 + FLTK2 + NxLib running on mips linux

by corvid＠lavabit.com

furaisanjin wrote: > > ?BTW, ?AFAIR ?the ?patch ?for ?wrapping chinese is "basic". As I > > don't ?know ?anything ?about ?chinese ?it's ?very ?hard for me to > > evaluate. > > I think the patch for Html_write_raw is rather simple than I thought. > If I can give one comment, it's better to treat all multi byte UTF8 > characters in the same way (not only Chinese and Japanese) so that > more people would be happy. If some people don't want to add this > patch for pure ASCII only language, would you able to add prefs > control for multi byte UTF8 character languages? Wouldn't it be wrong to wrap after most multibyte chars? (e.g., if you go to uk.wikipedia.org)

May 7, 2009

Dillo2 + FLTK2 + NxLib running on mips linux

by corvid＠lavabit.com

Jorge wrote: > BTW, AFAIR the patch for wrapping chinese is "basic". As I > don't know anything about chinese it's very hard for me to > evaluate. > > Corvid: is it worth to push this patch to the repo? Hard for me to evaluate, too. To me, it seems like a good thing on the whole... furaisanjin said that doing this makes text in tables too narrow, but I'd imagine that's a smaller matter than the lack of wrapping would be. I think the current patch would fail to work if numeric entities are used to specify characters.

May 7, 2009

Dillo2 + FLTK2 + NxLib running on mips linux

by furaisanjin＠gmail.com

> ?BTW, ?AFAIR ?the ?patch ?for ?wrapping chinese is "basic". As I > don't ?know ?anything ?about ?chinese ?it's ?very ?hard for me to > evaluate. I think the patch for Html_write_raw is rather simple than I thought. If I can give one comment, it's better to treat all multi byte UTF8 characters in the same way (not only Chinese and Japanese) so that more people would be happy. If some people don't want to add this patch for pure ASCII only language, would you able to add prefs control for multi byte UTF8 character languages? Regards, furaisanjin

May 7, 2009

Code for Wikipedia searches

by jcid＠dillo.org

On Mon, May 04, 2009 at 05:06:26PM -0700, Freya wrote: > > > Hiya! > A big thanks to Corvid who helped me loads to get my new dillo compiled and working! > > I wanted to give a little back, so I did a load of reasearch and finally managed to work out the code to get dillo to do direct wikipedia searches. > Just add this to the dillorc file: > > search_url="http://www.wikipedia.org/wiki/Special:Search?search=%s" > > It would be good to include it with future dillo releases as it's not easy to find out how to do this! It's lots of fun tho, just enter your keywords and it takes you right to the wikipedia page! :) > > I'm a bit of a wikipedia fan as you may have noticed. > > Anyway I know it's not loads but I think it's groovy and adds something to dillo. Thanks for the URL. It was just committed in dillorc, so 2.1 will have it hinted there. -- Cheers Jorge.-

May 7, 2009

Dillo2 + FLTK2 + NxLib running on mips linux

by jcid＠dillo.org

On Thu, May 07, 2009 at 02:07:22AM +0800, ???? ?i wrote: > Hello All > > One years ago I had tried dillo + fltk + nxlib and unsuccessful. > Today, the dillo2 works. I'm so happy. > > Also I have record my building progress in my blog > http://tw.myblog.yahoo.com/stevegigijoe/article?mid=138&prev=-1&next=137 > > Thank you all. Good! Do you have a screenshot? BTW, AFAIR the patch for wrapping chinese is "basic". As I don't know anything about chinese it's very hard for me to evaluate. Corvid: is it worth to push this patch to the repo? Steve: what's your opinion on the patch? Can you improve it? and finally, how are the "sung" fonts installed? (to be able to render chinese here). -- Cheers Jorge.-

May 7, 2009

Dillo2 + FLTK2 + NxLib running on mips linux

by stevegigijoe＠yahoo.com.tw

>On Thu, May 07, 2009 at 02:07:22AM +0800, ?? ? wrote: >> Hello All >> >> One years ago I had tried dillo + fltk + nxlib and unsuccessful. >> Today, the dillo2 works. I'm so happy. >> >> Also I have record my building progress in my blog >>http://tw.myblog.yahoo.com/stevegigijoe/article?mid=138&prev=-1&next=137 >Nice! On what hardware do you run dillo? >Cheers, >Johannes Running on MIPS 4KEc CPU clock 162.5MHz 16MBytes flash, 32MBytes memory Steve ______________________________________________________________________________________________________ ¥I¶O€~®e¶qµL€W¡HYahoo!©_Œ¯¹q€l«Hœc2.0§K¶Oµ¹§A¡A«H¥ó¥Ã»·€£¥²§R¡I http://tw.mg0.mail.yahoo.com/dc/landing

May 7, 2009

change FLTK scheme for in dillo2

by akhilman＠gmail.com

Greetings. Can I change colours of dialogues and scrollbar in dillo2? I would like to gets something less bright. Even something like fluid2's -bg option will be ok. Thank you for great lightweight browser. -- ?????? ??????????? aka AkhIL ??? ??? 7 13:39:04 KRAST 2009 Thu May 7 05:39:04 UTC 2009 ---------------------------------- Visit my home page http://akhilman.blogspot.com/ jabber: akhil(a)jabber.ru ---------------------------------- ??? ????????? ????????? ????????? ??, ??? ?? ??????? ?????????, ?????? ?? ???? ????, ??? ?? ???????????? ??? ?????????. (?.?????????) ---------------------------------- Linux artstation 2.6.28-gentoo-r5 #1 SMP Sat May 2 18:40:24 KRAST 2009 x86_64 AMD Phenom(tm) 9550 Quad-Core Processor AuthenticAMD GNU/Linux up 4 days, 18:33, 4 users, load average: 0.48, 0.40, 0.37

May 7, 2009

image resize

by furaisanjin＠gmail.com

Hi, 2009/5/7 Johannes Hofmann <Johannes.Hofmann(a)gmx.de>: > Hi, > > On Wed, May 06, 2009 at 12:25:03PM +0900, furaisanjin wrote: >> Hi. >> >> I studied this issue a bit further and notice that style causes the >> problem. For example, >> >> <img width="120" height="100" src="http://img.news.goo.ne.jp/photo/gooeditor/typ >> hoon.jpg"/ style="border:1px solid #ccc;"> >> >> This resizing is something strange. However if I remove style >> definition from img tag, resizing works. > > I think I've fixed it in tip. Please give it a try. Thank you, the fix works very well. Regards, furaisanjin

May 7, 2009

patch: a binding for stop

by corvid＠lavabit.com

Jeremy wrote: > That clashes with "Cut". Select some text in the URL window and > <ctrl>+x will delete it. I think it's not a good idea to grab > something that (a) already has a conventional action, and (b) already > performs that conventional action in the GUI. > > And while we're talking about conventional use: Dillo is highly > unconventional when it uses <ctrl>+q to close a tab. Everyone else[1] > uses <ctrl>+w to close a tab (<ctrl>+<shift>+q to close a window) and > <ctrl>+q to quit. Which can be annoying when switching between the > two. > > Firefox uses Escape to stop, and arora[2] uses <ctrl>+. . You know, if dillo just changed all of its default bindings to whatever firefox does, that would probably be a relief to every single casual user, and probably would only force a little relearning or keysrc editing on the subset of the population interested enough to be subscribed to dillo-dev.

May 6, 2009

Dillo2 + FLTK2 + NxLib running on mips linux

by Johannes.Hofmann＠gmx.de

On Thu, May 07, 2009 at 02:07:22AM +0800, ?? ? wrote: > Hello All > > One years ago I had tried dillo + fltk + nxlib and unsuccessful. > Today, the dillo2 works. I'm so happy. > > Also I have record my building progress in my blog > http://tw.myblog.yahoo.com/stevegigijoe/article?mid=138&prev=-1&next=137 Nice! On what hardware do you run dillo? Cheers, Johannes

May 6, 2009

image resize

by Johannes.Hofmann＠gmx.de

Hi, On Wed, May 06, 2009 at 12:25:03PM +0900, furaisanjin wrote: > Hi. > > I studied this issue a bit further and notice that style causes the > problem. For example, > > <img width="120" height="100" src="http://img.news.goo.ne.jp/photo/gooeditor/typ > hoon.jpg"/ style="border:1px solid #ccc;"> > > This resizing is something strange. However if I remove style > definition from img tag, resizing works. I think I've fixed it in tip. Please give it a try. Regards, Johannes

May 6, 2009

Dillo2 + FLTK2 + NxLib running on mips linux

by stevegigijoe＠yahoo.com.tw

Hello All One years ago I had tried dillo + fltk + nxlib and unsuccessful. Today, the dillo2 works. I'm so happy. Also I have record my building progress in my blog http://tw.myblog.yahoo.com/stevegigijoe/article?mid=138&prev=-1&next=137 Thank you all. ______________________________________________________________________________________________________ ¥I¶O€~®e¶qµL€W¡HYahoo!©_Œ¯¹q€l«Hœc2.0§K¶Oµ¹§A¡A«H¥ó¥Ã»·€£¥²§R¡I http://tw.mg0.mail.yahoo.com/dc/landing

May 6, 2009

patch: a binding for stop

by onepoint＠starurchin.org

On Wed, May 06, 2009 at 11:32:31AM -0400, Jorge Arellano Cid wrote: > > The other one I can think of as mnemonic is <ctrl>x. What do > people think? That clashes with "Cut". Select some text in the URL window and <ctrl>+x will delete it. I think it's not a good idea to grab something that (a) already has a conventional action, and (b) already performs that conventional action in the GUI. And while we're talking about conventional use: Dillo is highly unconventional when it uses <ctrl>+q to close a tab. Everyone else[1] uses <ctrl>+w to close a tab (<ctrl>+<shift>+q to close a window) and <ctrl>+q to quit. Which can be annoying when switching between the two. Firefox uses Escape to stop, and arora[2] uses <ctrl>+. . Regards, Jeremy Henty [1] ie. Firefox :-) [2] http://code.google.com/p/arora/ - WebKit+Qt-based browser. It looks very nice and may supplant Firefox as "the browser to fall back to when Dillo can't handle the page".

May 6, 2009

patch: a binding for stop

by jcid＠dillo.org

On Wed, May 06, 2009 at 02:46:45PM +0000, corvid wrote: > I can't find the old discussion, but long ago I asked > Jorge about adding a binding for stop, and I remember > his response being something like "Maybe. Perhaps we > could use <ctrl>P." The other one I can think of as mnemonic is <ctrl>x. What do people think? > As for me, I'm going to change it to use Escape in my > keysrc because I can press it in a hurry after I > accidentally mistype/misclick. I found it quite useful > back in my netscape/mozilla-using pre-dillo days. > Obviously that can't be the default here, though, > because Esc is already taken. That's the good part of it being configurable, I'm just trying to define a good default. -- Cheers Jorge.-

May 6, 2009

patch: a binding for stop

by corvid＠lavabit.com

I can't find the old discussion, but long ago I asked Jorge about adding a binding for stop, and I remember his response being something like "Maybe. Perhaps we could use <ctrl>P." As for me, I'm going to change it to use Escape in my keysrc because I can press it in a hurry after I accidentally mistype/misclick. I found it quite useful back in my netscape/mozilla-using pre-dillo days. Obviously that can't be the default here, though, because Esc is already taken.

May 6, 2009

image resize

by Johannes.Hofmann＠gmx.de

Hi, On Wed, May 06, 2009 at 12:25:03PM +0900, furaisanjin wrote: > Hi. > > I studied this issue a bit further and notice that style causes the > problem. For example, > > <img width="120" height="100" src="http://img.news.goo.ne.jp/photo/gooeditor/typ > hoon.jpg"/ style="border:1px solid #ccc;"> > > This resizing is something strange. However if I remove style > definition from img tag, resizing works. Ah that's interesting. It seems the problem occurs if border-width, margin, or padding are non-zero and even then only after a reload. Regards, Johannes

May 6, 2009

image resize

by furaisanjin＠gmail.com

Hi. I studied this issue a bit further and notice that style causes the problem. For example, <img width="120" height="100" src="http://img.news.goo.ne.jp/photo/gooeditor/typ hoon.jpg"/ style="border:1px solid #ccc;"> This resizing is something strange. However if I remove style definition from img tag, resizing works. Regards, furaisanjin

May 6, 2009

patch: punctuation keys

by corvid＠lavabit.com

Jorge wrote: > On Tue, May 05, 2009 at 02:13:34PM +0000, corvid wrote: > > Jorge wrote: > > > On Mon, May 04, 2009 at 03:20:14AM +0000, corvid wrote: > > > > PS <shift>TabKey doesn't work for me, but it wasn't working for me before, > > > > either. > > > > > > It works here. Maybe the window manager or terminal is eating it? > > > > It highlights the location and gives it focus. > > Yes, and it also serves to go backwards through controls > inside a form. > > > Does xev show ISO_Left_Tab for the keysym? > > Yes. And (just making sure here) it is also doing what you'd expect as a result of { "left-tab" , KEYS_LEFT_TAB , fltk::SHIFT , fltk::TabKey } Or, oh, wait, KEYS_LEFT_TAB and KEYS_RIGHT_TAB aren't hooked up to anything yet. Now I understand.

May 5, 2009

patch: punctuation keys

by jcid＠dillo.org

On Tue, May 05, 2009 at 02:13:34PM +0000, corvid wrote: > Jorge wrote: > > On Mon, May 04, 2009 at 03:20:14AM +0000, corvid wrote: > > > PS <shift>TabKey doesn't work for me, but it wasn't working for me before, > > > either. > > > > It works here. Maybe the window manager or terminal is eating it? > > It highlights the location and gives it focus. Yes, and it also serves to go backwards through controls inside a form. > Does xev show ISO_Left_Tab for the keysym? Yes. -- Cheers Jorge.-

May 5, 2009

patch: punctuation keys

by corvid＠lavabit.com

Jorge wrote: > On Mon, May 04, 2009 at 03:20:14AM +0000, corvid wrote: > > PS <shift>TabKey doesn't work for me, but it wasn't working for me before, > > either. > > It works here. Maybe the window manager or terminal is eating it? It highlights the location and gives it focus. Does xev show ISO_Left_Tab for the keysym?

May 5, 2009

patch: punctuation keys

by jcid＠dillo.org

On Mon, May 04, 2009 at 03:20:14AM +0000, corvid wrote: > I'm thinking it might be more natural to be able to put, e.g., > "*" instead of "<shift>8" (my keyboard, anyway). > This patch is working for the cases I've happened to try, but > I am not sure that event_text() will necessarily work for everything. > > (Last bit of the patch allows binding '<', which is a shifted key for me.) Committed. > PS <shift>TabKey doesn't work for me, but it wasn't working for me before, > either. It works here. Maybe the window manager or terminal is eating it? -- Cheers Jorge.-

May 5, 2009

image resize

by furaisanjin＠gmail.com

2009/5/5 Johannes Hofmann <Johannes.Hofmann(a)gmx.de>: > On Tue, May 05, 2009 at 02:06:43PM +0900, furaisanjin wrote: >> Hi all. >> >> I think image resize doesn't work when image was already fetched. >> >> This is one example page. http://news.goo.ne.jp >> This page contains CSS so that repush happens. Images have been >> fetched already when repush is made. > > Is this a recent regression? Could you try to use "hg bisect" to > find the revision that introduced this problem? I'm not 100% sure but I don't think this is recent regression. I'm wondering where I should start. Regards, furaisanjin

May 5, 2009

image resize

by Johannes.Hofmann＠gmx.de

On Tue, May 05, 2009 at 02:06:43PM +0900, furaisanjin wrote: > Hi all. > > I think image resize doesn't work when image was already fetched. > > This is one example page. http://news.goo.ne.jp > This page contains CSS so that repush happens. Images have been > fetched already when repush is made. Is this a recent regression? Could you try to use "hg bisect" to find the revision that introduced this problem? Regards, Johannes

May 5, 2009