RE: [Dillo-dev]Re: About bug #60 (transparency in PNG's) From: Eric GAUDET - 2000-11-29 10:46 -- En reponse de "[Dillo-dev]Re: About bug #60 (transparency in PNG's)" de Livio Baldini Soares, le 29-Nov-2000 : > Does anybody know a better/faster solution to turn an guint32 into > to three int's representing red, green and blue parts of the color? > The solution I've implemented doesn't look too good :-( > Tomorrow, after getting some sleep I'll look up my algebra and see if > I can do some decent bit manipulation instead of calling strtol(). If > anyone has any ideas please send them in. > IMHO there's several problem with your calculations : 1) the strol part is very inefficient, but that's not a real problem because this code is _outside_ the loop. Anyway, here's how to do it : (strip all the sprintf) bg_blue = (prefs.bg_color0 & 0xFF; bg_green = (prefs.bg_color>>8) & 0xFF; bg_red = (prefs.bg_color>>16) & 0xFF; 2) The loop is _always_ the sensible part ! You _never_ want to allocate local variables like you did : for (...) { gint p,a; (some compiler may optimize this, but you never know) As a rule of thumb, it's better to allocate local variables at the begining of a function. 3) If you really want to improve speed, do all calculation only once in local variables : i*3 is computed three times, it would be better to have calculated i3 = i*3 and use i3. (same with i*4 used twice). You could also have done i3 += 3 each loop, but on pentium-generation cpu (whatever the brand, ppc, arm, ...) multiplications and additions cost the same so it doesn't matter, and using i*3 is often more readable and more robust (not sensible to initiallisation). 4) Even better : look into a table is a calculation. png->linebuf[3*i] is actually *(png + linebuf + 3*i). You could have declared guchar *pl = png->linebuf; and in the loop : if (!a){ *(pl++) = bg_red; *(pl++) = bg_green; *(pl++) = bg_blue; } The else part is basically the same, if a little more trickier, and you can get rid of the costy for() : I let it to you as an exercise ;-) (hint : row_num*png->rowbytes is ugly) 5) Last : if(!a) is more complex than if(a) : invert the if/else. Hope this helps. So long Livio :-) ----------------------------------- Eric GAUDET Le 29-Nov-2000 a 19:19:52 "Le verbe inexister ... inexiste !" ----------------------------------- [Dillo-dev]Re: About bug #60 (transparency in PNG's) (sending patch) From: Livio Baldini Soares - 2000-11-29 03:36 Attachments: png_transparent.diff Sebastian Geerken writes: > Livio Baldini Soares writes: > > Hi all, > > > > I've been looking into the PNG transparency bug (#60). Ok, so I got a working patch that gets the user preference bg_color and sets that as transparent. Since it might take a while before `a_Dw_widget_get_bg_color` or mask clipping gets done, this might be a good thing to put it for now. The only thing is that this patch didn't turn out as good as I'd like it to be... Does anybody know a better/faster solution to turn an guint32 into to three int's representing red, green and blue parts of the color? The solution I've implemented doesn't look too good :-( Tomorrow, after getting some sleep I'll look up my algebra and see if I can do some decent bit manipulation instead of calling strtol(). If anyone has any ideas please send them in. I'm sending the patch as an attachment and should be applyied against src/png.c from CVS. bye all, best regards, -- Livio Re: [Dillo-dev]Still some problems with the bugtrack engine From: Eric GAUDET - 2000-11-29 01:37 -- En reponse de "Re: [Dillo-dev]Still some problems with the bugtrack engine" de Livio Baldini Soares, le 28-Nov-2000 : > Hi Eric, how you doing? > Hi Livio, nice to see you again in the list (I was afraid you left ;-) > Eric GAUDET writes: >> I have Internal Server errors every time I try. >> > > I added the bug for you. It's bug number *102*, check out the > bug-track. I think that the problem appears when you leave out one of > the fields in blank. I was getting an Internal Error too, but had > left > in blank the HTReproduce. When I filled it in with `Nothing`, the bug > was submitted. > Hmmm... I just tried it and I think you're right. Thanks ----------------------------------- Eric GAUDET Le 29-Nov-2000 a 10:34:35 "Le verbe inexister ... inexiste !" ----------------------------------- Re: [Dillo-dev]About bug #60 (transparency in PNG's) From: Sebastian Geerken - 2000-11-28 22:11 Livio Baldini Soares writes: > Hi all, > > I've been looking into the PNG transparency bug (#60). After an > hour of searches along src/png.c and libpng-dev thinking that it was > an initilization issue I found out it had nothing to do with > that. Later I found out that the problem was in > `Png_datarow_callback`, in the `case 4:` the alpha channel wasn't > being activated (there was even a `todo`, which I missed!). I've made > a sketch that works, but I've come to a problem: > > Where do I find the current window (page) background color? Better: the background color of the parent of the image widget. This will be different from the page background color, when tables are implemented. > I guess > this is the same problem with GIF transparency. When you have a > transparent GIF and open it with the standard dillorc file, it'll look > alright, but if you change the bg_color in your dillorc then the GIF > will have in the transparent pixels NOT your background color, but the > color for the standard bg_color that was in the original dillorc... > > So I guess that's a problem with GIF's and PNG's. Sebastian, is this > bg_color integrated to the new Dw your making up? Should I wait for it > to sent in the patch for PNG transparency, or should I make it > already? (Actually, it's already done and working, but I didn't want > to sent it in with this background color issue). I haven't done much on colors yet, currently there is only a function a_Dw_widget_set_bg_color (which does nothing, but will soon work for top-level widgets), and there will probably (when needed) a function a_Dw_widget_get_bg_color (I'd prefer a function instead of a direct access on structure members). Meanwhile, you may use dummy function with that name (returning DW_PAINT_DEFAULT_BGND or so). However, I already suggested to Jorge to use clipping masks instead of the parent's background color, since the latter would cause problems with background images (this could be solved, but only complicated and slow). Real alpha (more than 1 bit depth), as supported by pgn (afaik) is not supported by X, but may be implemented by dithered clipping masks (will look a bit ugly, but I did never see a png in a web page which uses this feature). Sebastian Re: [Dillo-dev]Still some problems with the bugtrack engine From: Livio Baldini Soares - 2000-11-28 21:10 Hi Eric, how you doing? Eric GAUDET writes: > I have Internal Server errors every time I try. > > I wanted to make this entry : > - There's no "Copy Link Location" being worked by "egaudet@fr... > 100%" > > :-/ I added the bug for you. It's bug number *102*, check out the bug-track. I think that the problem appears when you leave out one of the fields in blank. I was getting an Internal Error too, but had left in blank the HTReproduce. When I filled it in with `Nothing`, the bug was submitted. Jorge are you the one who take cares of the dillo's bug-track CGI's? Could it be that in Dillo_insert.cgi a blank field could cause an error? Maybe it should be tuned for errors and such, and give a warning like: "X field left in blank, please fill it in..."... Just a guess. best regards to all, -- Livio [Dillo-dev]Still some problems with the bugtrack engine From: Eric GAUDET - 2000-11-28 16:40 I have Internal Server errors every time I try. I wanted to make this entry : - There's no "Copy Link Location" being worked by "egaudet@fr... 100%" :-/ ----------------------------------- Eric GAUDET Le 29-Nov-2000 a 01:38:01 "Le verbe inexister ... inexiste !" ----------------------------------- [Dillo-dev]About bug #60 (transparency in PNG's) From: Livio Baldini Soares - 2000-11-28 15:59 Hi all, I've been looking into the PNG transparency bug (#60). After an hour of searches along src/png.c and libpng-dev thinking that it was an initilization issue I found out it had nothing to do with that. Later I found out that the problem was in `Png_datarow_callback`, in the `case 4:` the alpha channel wasn't being activated (there was even a `todo`, which I missed!). I've made a sketch that works, but I've come to a problem: Where do I find the current window (page) background color? I guess this is the same problem with GIF transparency. When you have a transparent GIF and open it with the standard dillorc file, it'll look alright, but if you change the bg_color in your dillorc then the GIF will have in the transparent pixels NOT your background color, but the color for the standard bg_color that was in the original dillorc... So I guess that's a problem with GIF's and PNG's. Sebastian, is this bg_color integrated to the new Dw your making up? Should I wait for it to sent in the patch for PNG transparency, or should I make it already? (Actually, it's already done and working, but I didn't want to sent it in with this background color issue). best regards to all, -- Livio Re: [Dillo-dev]Misc. From: Livio Baldini Soares - 2000-11-28 13:34 Hi again, I forgot something important, read below. Livio Baldini Soares writes: > Hi again, > > think I fixed bug 100 already, read below. > > Livio Baldini Soares writes: > > > > > > - There's a small bug within a_Url_squeeze. When there're too > > > much "/.." sequences in the URL, it removes the server! > > > I got to it, and I think I fixed it ok. I thought the original code > kind of confusing, but I've done little changes (3 lines). Basically > here is my diff (against src/IO/Url.c from CVS, but I think 0.3.0 will > be ok too since this hasn't changed, I think): The a_Url_squeeze seems to work fine, but that never gets called for user typed URL's. Im my opinion this is wrong, so I have another patch which changes a_Nav_push to change the user specified URL to the squeezed one. Then you'll be able to see how the a_Url_squeeze changes the URL (using the previous patch I sent). Here is the patch against src/nav.c, I changed only the a_Nav_push funtion: ***************************************************** --- dillo/src/nav.c Mon Nov 13 10:01:55 2000 +++ dillo.new/src/nav.c Tue Nov 28 10:38:20 2000 @@ -203,20 +203,22 @@ void a_Nav_remove_top_url(BrowserWindow */ void a_Nav_push(BrowserWindow *bw, const char *Url) { - char *p, *url; + char *p, *url, *sUrl; g_return_if_fail (bw != NULL); - url = g_strdup(Url); + sUrl = a_Url_squeeze(g_strdup(Url)); /* get the squeezed Url */ + url = g_strdup(sUrl); if ( (p = strstr(url, "?!POST")) != NULL ) *p = '\0'; a_Nav_cancel_expect(bw); - bw->nav_expect.url = g_strdup(url); + bw->nav_expect.url = g_strdup(sUrl); bw->nav_expect.title = NULL; bw->nav_expecting = TRUE; - Nav_open_url(bw, Url, FALSE); + Nav_open_url(bw, sUrl, FALSE); g_free(url); + g_free(sUrl); } /* **************************************** that's all, bye, -- Livio Re: [Dillo-dev]Misc. From: Livio Baldini Soares - 2000-11-28 11:13 Hi again, think I fixed bug 100 already, read below. Livio Baldini Soares writes: > > > - There's a small bug within a_Url_squeeze. When there're too > > much "/.." sequences in the URL, it removes the server! I got to it, and I think I fixed it ok. I thought the original code kind of confusing, but I've done little changes (3 lines). Basically here is my diff (against src/IO/Url.c from CVS, but I think 0.3.0 will be ok too since this hasn't changed, I think): ************************************************************ diff -pru dillo/src/IO/Url.c dillo.new/src/IO/Url.c --- dillo/src/IO/Url.c Sat Nov 11 01:54:45 2000 +++ dillo.new/src/IO/Url.c Mon Nov 27 23:43:33 2000 @@ -252,7 +252,10 @@ char *a_Url_squeeze(char *str) str[ni++] = s[i]; if ( nc && ni ) --ni; + nc = ni; while ( ni && str[--ni] != '/'); + if (!ni || (!strncmp(str, ") && ni == 6)) + ni = nc+1; /* if we couldn't find a parent direcory, restore value */ s = p = p + 3; } else if ( p[2] == '/' || !p[2] ) { /* "/./" or "/." */ nc = p - s; @@ -264,6 +267,8 @@ char *a_Url_squeeze(char *str) p += 2; } } + + if (ni && str[ni-1] == '/') ni--; /* Append the rest of 'str' */ if ( ni == 0 && !*s ) str[ni++] = '/'; while ( (str[ni++] = *s++) ); ***************************************************** What I do is when trying to look for a parent directory, prevent it from eating up the hostname. That is done in the `if` I inserted. If the variable `ni` got back to the beginning of the hostname (==0) or found a '/' expect it's the one from " then restore it's value... Umm, guess I'm not very good at explaining this.. is any body understanding what the hell am I saying? If anybody wants me to explain this again, just ask, I won't complain! I'll change the status to 100% in the bug-track when I get back net connection. Oh, and I'm NOT suppossed to put `done`, this is done by Jorge when he inserts the patch in the code (IF we agree the code is worth putting in), is this correct Jorge? Another thing, there was a comment in the header of a_Url_squeeze made by Jorge, saying the he wasn't sure that it was necessary. Well I checked APACHE and it's NOT necessary to squeeze the URL. I sent in a request with a telnet to port 80, and sent an `GET /./FAQ/../ HTTP/1.0`, and got the /index.html. Seems to work fine with both /../ and /./, but I'm not sure if this is true to all HTTP servers. The only time APACHE complains is when we send in a request for /FAQ/../../, APACHE claims this is a BAD REQUEST, but the new a_Url_squeeze prevents this. that's enough for now!, any doubts or flames, don't hesitate to send, bye yall, -- Livio RE: [Dillo-dev]Re: From: Eric GAUDET - 2000-11-28 08:48 -- En reponse de "[Dillo-dev]Re:" de Jorge Arellano Cid, le 27-Nov-2000 : >> A friend of mine as access to several sparc stations (both 32 and 64 >> bits, under linux and solaris 8). I asked him to test Dillo. > > Thanks. > Actually dillo works perfectly on a Sparc 32, but the Sparc 64 > and Solaris architectures haven't been tested. I'll be expecting > that info. > Dillo has been successfully tested under : - sparc 5 (32 bits) linux, glibc 2.1.3, gcc 2.95.3 - sparc ultra 1 (64 bits kernel, 32 bits userspace) linux, glibc 2.1.3, gcc 2.95.3 - sparc ultra 10 (64 bits) solaris 8, gcc 2.95.2 No problem at all : just compile and run :-)))) ----------------------------------- Eric GAUDET Le 28-Nov-2000 a 17:44:57 "Le verbe inexister ... inexiste !" ----------------------------------- RE: [Dillo-dev]Re: From: Jörgen Viksell - 2000-11-28 07:12 >GtkHSeparator doesn't seem to have all the features needed for painting >a HR tag with attributes : no line thickness (-> no SIZE attribute), >only 3d-rendering (-> no NOSHADE attribute). >Will it be a problem ? If the new Dw handles GTK widget as GTK does itself (resizing and stuff) then it should only be the matter of writing a new function for widget->style->klass->draw_hline I did a quick experiment with that and there were no problems with those attributes. I guess reality will prove me to be wrong though... ;-) // Jörgen Viksell _____________________________________________________________________________________ Get more from the Web. FREE MSN Explorer download : http://explorer.msn.com RE: [Dillo-dev]Re: From: Eric GAUDET - 2000-11-28 02:22 -- En reponse de "[Dillo-dev]Re:" de Sebastian Geerken, le 27-Nov-2000 : > Jorge Arellano Cid writes: > > Hrulers > > > > I considered patches from Allan, Eric and some progress by > > Sebastian. > > > > Hrulers will have to be reimplemented for the new Dw, and as > > both patches implemented a resize mechanism that's not the one > > that current-Dw uses, and because both showed several problems > > (incompletness or fails) I'd prefer to wait for the new Dw and > > try to construct the hruler based upon the good ideas that both > > patches showed. > > Both patches will be obsolete, this is already working (very > different) in the new Dw with embedded GtkHSeparator's (but for > support of all attributes of
there is still the need for a new > DwHRuler). > > Sebastian GtkHSeparator doesn't seem to have all the features needed for painting a HR tag with attributes : no line thickness (-> no SIZE attribute), only 3d-rendering (-> no NOSHADE attribute). Will it be a problem ? ----------------------------------- Eric GAUDET Le 28-Nov-2000 a 11:17:54 "Le verbe inexister ... inexiste !" ----------------------------------- RE: [Dillo-dev]Re: From: Eric GAUDET - 2000-11-28 01:58 -- En reponse de "[Dillo-dev]Re:" de Jorge Arellano Cid, le 27-Nov-2000 : > Pagemarks: This must be moved to the rightmost-button menu and > set insensitive when there're no headings in the page. > I'll do it soon. ----------------------------------- Eric GAUDET Le 28-Nov-2000 a 10:58:23 "Le verbe inexister ... inexiste !" ----------------------------------- [Dillo-dev]Re: From: Sebastian Geerken - 2000-11-27 23:34 Jorge Arellano Cid writes: > Hrulers > > I considered patches from Allan, Eric and some progress by > Sebastian. > > Hrulers will have to be reimplemented for the new Dw, and as > both patches implemented a resize mechanism that's not the one > that current-Dw uses, and because both showed several problems > (incompletness or fails) I'd prefer to wait for the new Dw and > try to construct the hruler based upon the good ideas that both > patches showed. Both patches will be obsolete, this is already working (very different) in the new Dw with embedded GtkHSeparator's (but for support of all attributes of
there is still the need for a new DwHRuler). Sebastian Re: [Dillo-dev]Misc. From: Livio Baldini Soares - 2000-11-27 18:33 Livio Baldini Soares writes: > I put in a patch for this, it's bug number is 100. I put in that I > was going to work on it.... but I guess I did something wrong and only > my name appeared, is it possible for me to change the `Being worked:` > field? If not, Jorge could you change it to livio@li..., I > think it's much more informative than just `Livio`. Nevermind, just changed it to show my e-mail, sorry about that... bye yall, -- Livio Re: [Dillo-dev]Misc. From: Livio Baldini Soares - 2000-11-27 18:17 Jorge Arellano Cid writes: > > Hi there! > Hi Jorge, I liked the new version (0.3.0) a lot. Nice work. By the way, if by any chance I get the chance to patch anything, then I should probably patch against CVS, right? Another thing, do you (or anybody else) have anything against sending patches to the list, so everyone can "pitch" in, and publically(?) discuss them. If not do you want me to CC: them to you, Jorge? > - There's a small bug within a_Url_squeeze. When there're too > much "/.." sequences in the URL, it removes the server! > > Patches for these BUGs, and their bug-track-entries, would be > appreciated. > I put in a patch for this, it's bug number is 100. I put in that I was going to work on it.... but I guess I did something wrong and only my name appeared, is it possible for me to change the `Being worked:` field? If not, Jorge could you change it to livio@li..., I think it's much more informative than just `Livio`. best regards to all, -- Livio [Dillo-dev]Re: From: Jorge Arellano Cid - 2000-11-27 16:25 Hi! The code with the Findtext and Pagemarks is in the CVS. The server reported no problems when uplading, so I hope its working (there were several BUG reports at sourceforge for this; Ah, and sourceforge's download counters had gone carzy too!!! Dillo shows a negative download rate the day with the highest peak of page visits!!! :-) Back to Findtext: this implementation is a basic one, there's still work to do. Most notably stripping of '\n\r\t' and spaces when adding to the Text buffer, feedback messages and finally a port to the new Dw when it's ready. Pagemarks: This must be moved to the rightmost-button menu and set insensitive when there're no headings in the page. ----- Dillo in a 64bit machine: > Actually it does work on 64-bit systems. I currently run it on an Alpha > It isn't my main browser (yet), but it is slowly getting there. FYI I > also maintain the Debian dillo package. I'm very happy to know it works!!! (my efforts haven't been wasted!). In regard to the slow progress, I agree with you, and its mostly due to an effort to clean and implement design solutions in the code that serve as framework for future enhancement: networking has been rewritten, Dw is getting a new design and plugins are being developed to simplify several tasks. If we succeed with them, I hope we can see faster improvement in the future. bear with us and let's get there! > A friend of mine as access to several sparc stations (both 32 and 64 > bits, under linux and solaris 8). I asked him to test Dillo. Thanks. Actually dillo works perfectly on a Sparc 32, but the Sparc 64 and Solaris architectures haven't been tested. I'll be expecting that info. --- Hrulers I considered patches from Allan, Eric and some progress by Sebastian. Hrulers will have to be reimplemented for the new Dw, and as both patches implemented a resize mechanism that's not the one that current-Dw uses, and because both showed several problems (incompletness or fails) I'd prefer to wait for the new Dw and try to construct the hruler based upon the good ideas that both patches showed. slangley: I saw your entry (BUG#62) and I'd prefer you to work it with the new Dw (just not to waste your efforts). --- >> - There a small bug within a_Url_squeeze. When there're too >> much "/.." sequences in the URL, it removes the server! > > How-to reproduce it ... ? Try something like: http://some.site.org/../main.html or http://some.site.org/../main.html they should resolve to http://some.site.org/main.html --- > I'm having the hardest time making bug-track-entries : 99% of the time > I'm getting : > > Internal Server Error ... I just made an entry from other network, for Find Text (BUG#99). If someone else has problems please tell me. Jorge.- RE: [Dillo-dev]Misc. From: Eric GAUDET - 2000-11-27 14:44 -- En reponse de "[Dillo-dev]Misc." de Jorge Arellano Cid, le 25-Nov-2000 : > Patches for these BUGs, and their bug-track-entries, would be > appreciated. > I'm having the hardest time making bug-track-entries : 99% of the time I'm getting : << Internal Server Error The server encountered an internal error or misconfiguration and was unable to complete your request. Please contact the server administrator, root@localhost and inform them of the time the error occurred, and anything you might have done that may have caused the error. More information about this error may be available in the server error log. Apache/1.3.12 Server at huallepen.inf.utfsm.cl Port 80 >> BTW: if comment and howto-reproduce lines are limited to 256 chars, it would be easier for everbody to ask for : ROWS=4 COLS=64 instead of the stupid : ROWS=5 COLS=80 ('see what I mean ? ;-) ... And since I'm talking about the bug-track engine, it would be nice to have a "Wish insertion page" : all unimplemented features are not "bugs", and I'm not sure I want to see "I'd like to have SSL", "I'd like a single download window" pollute the bug list. I don't want them forgotten when we'll be looking for new features implementation and long term projects. (would a poll be hard to implement, for most-wanted features ?) ----------------------------------- Eric GAUDET Le 27-Nov-2000 a 23:18:12 "Le verbe inexister ... inexiste !" ----------------------------------- RE: [Dillo-dev]Misc. From: Eric GAUDET - 2000-11-27 01:36 -- En reponse de "[Dillo-dev]Misc." de Jorge Arellano Cid, le 25-Nov-2000 : > - A "Copy link location" item (to allow pasting) I'm going to try that, but I can't promise anything. > - There's a small bug within a_Url_squeeze. When there're too > much "/.." sequences in the URL, it removes the server! How-to reproduce it ... ? ----------------------------------- Eric GAUDET Le 27-Nov-2000 a 10:35:44 "Le verbe inexister ... inexiste !" ----------------------------------- Re: [Dillo-dev]Misc. From: Ron Farrer - 2000-11-26 16:30 Attachments: Message as HTML Eric GAUDET (egaudet@in...) wrote: > If you know some C, would you mind investigating and trying to fix > alpha-specific bug#89 and 90 ? (since you're the only alpha-guy around > ;-) Is there a way of contacting whoever submitted bug#89? I am unable to get dillo to segfault on lwn.net. Also, bug#90 is using a non-free compiler.. I can't help with that. Ron --=20 Email: , Home: Alpha Linux Powered: Re: [Dillo-dev]Misc. From: Eric GAUDET - 2000-11-26 12:30 -- En reponse de "Re: [Dillo-dev]Misc." de Ron Farrer, le 26-Nov-2000 : > Actually it does work on 64-bit systems. I currently run it on an > Alpha > It isn't my main browser (yet), but it is slowly getting there. FYI I > also maintain the Debian dillo package. > If you know some C, would you mind investigating and trying to fix alpha-specific bug#89 and 90 ? (since you're the only alpha-guy around ;-) ----------------------------------- Eric GAUDET Le 26-Nov-2000 a 21:27:05 "Le verbe inexister ... inexiste !" ----------------------------------- [Dillo-dev][ANN]Preferences plugin ! From: Eric GAUDET - 2000-11-26 12:24 Hi, I just published my Preferences Dillo-plugin at my now usual address (http://www.rti-zone.org/dillo/) Fully working : it writes dillorc ! (provided your dillo's patched for "dpi" ;-) With a screenshot ! ----------------------------------- Eric GAUDET Le 26-Nov-2000 a 21:19:18 "Le verbe inexister ... inexiste !" ----------------------------------- Re: [Dillo-dev]Misc. From: Ron Farrer - 2000-11-26 05:24 Attachments: Message as HTML Eric GAUDET (egaudet@in...) wrote: > -- En reponse de "[Dillo-dev]Misc." de Jorge Arellano Cid, le > 25-Nov-2000 : > > - Make Dillo work in a 64bit machine! (I worked on this a lot, > > but I no longer have access to the machine I was using). > >=20 >=20 > A friend of mine as access to several sparc stations (both 32 and 64 > bits, under linux and solaris 8). I asked him to test Dillo. Actually it does work on 64-bit systems. I currently run it on an Alpha It isn't my main browser (yet), but it is slowly getting there. FYI I=20 also maintain the Debian dillo package.=20 Ron --=20 Email: , Home: Debian on Alpha: [Dillo-dev]Tiny bug in FORM checkboxes From: Eric GAUDET - 2000-11-26 03:04 Hi, There's a tiny bug in the CHECKBOX handling : if no VALUE attribute is specified, the checked-box value is an empty string ; the default value should be "on" so one can differentiate a checked box with an empty TEXT input. If it's not a bug, it's definitely annoying for POST parsing. Here's how to fix it : (dillo v0.3.0) in html.c : line 1645 - init_str = ""; + init_str = "on"; ----------------------------------- Eric GAUDET Le 26-Nov-2000 a 11:57:59 "Le verbe inexister ... inexiste !" ----------------------------------- RE: [Dillo-dev]Misc. From: Eric GAUDET - 2000-11-26 01:29 -- En reponse de "[Dillo-dev]Misc." de Jorge Arellano Cid, le 25-Nov-2000 : > - Make Dillo work in a 64bit machine! (I worked on this a lot, > but I no longer have access to the machine I was using). > A friend of mine as access to several sparc stations (both 32 and 64 bits, under linux and solaris 8). I asked him to test Dillo. ----------------------------------- Eric GAUDET Le 26-Nov-2000 a 10:28:01 "Le verbe inexister ... inexiste !" ----------------------------------- Re: [Dillo-dev]Misc. From: Will Newton - 2000-11-25 14:57 On Sat, 25 Nov 2000, Jorge Arellano Cid wrote: > Although a bit outdated, there's a "project notes" link that > has some hints on this subject. If you want to work on it, let me > know, cause I'll have to tune the networking part in order to > make it functional. Have you considered using w3c libwww for networking? It's part of GNOME now and is available for most ditros. (It does HTTP/1.1, SSL and lots of other nice things) [Dillo-dev]Misc. From: Jorge Arellano Cid - 2000-11-25 14:18 Hi there! When reading the mailing-list, there were several topics so, here go my comments and answers: > If I hit a slow site, I get no visual clue that dillo is doing > anything. To see this first hand, go to bugs.debian.org, and look > up the package lintian. dillo seems to hang, then magically load > the page. The Stop button may suggest some activity, but that's certainly not enough... I thought that a small panel to the left of the progress bars can be a good solution. I mean a table like this: ,--------------. | Dns | Send | [Image progress bar] | Retr | Close | [Text progress bar ] `--------------' Those words in the panel can be highlighted when the correspondent activities are taking place. > Robert investigated the problem (bug#97) a little more. It appears to > be a gcc 2.95.2 bug for K6 with -O2 option. I heard about that for > non-i386 cpu (SH4, ARM). It seems -O2 is not reliable for now, except > for intel processors. > > We should put a warning in the README about this. OK, that'll be in the README file of our next release. > * Downloads > Rather than having loads of dialogs like in Netscape, a single > window with a list of downloads and status would be cool (imagine > a more simple GetRight style window). Although a bit outdated, there's a "project notes" link that has some hints on this subject. If you want to work on it, let me know, cause I'll have to tune the networking part in order to make it functional. --- Some patching suggestions: - A "Copy link location" item (to allow pasting) - There's a small bug within a_Url_squeeze. When there're too much "/.." sequences in the URL, it removes the server! - Make Dillo work in a 64bit machine! (I worked on this a lot, but I no longer have access to the machine I was using). Patches for these BUGs, and their bug-track-entries, would be appreciated. --- Finally, when integrating the Find-text patch, I noticed that it only works for the first match, so I got into it and now I have a Find-text that does it. I'll send a note to the list when I upload the code to the CVS. Jorge.- RE: [Dillo-dev]Suggestions From: Will Newton - 2000-11-24 15:50 On Thu, 23 Nov 2000, Sean 'Shaleh' Perry wrote: > I have suggested this. Jorge is not 100% convinced. Plus dillo needs some > rework to handle css. html items have no storage of names. So you can't do >
. I'll write a CSS parser if you want (a quick search on google says the only C CSS parser is in Amaya... I advise any children watching to avert their eyes from that code...). RE: [Dillo-dev]Suggestions From: Eric GAUDET - 2000-11-24 13:26 -- En reponse de "[Dillo-dev]Suggestions" de Will Newton, le 24-Nov-2000 : > > Here are a few ideas/suggestions I have: > > * Errors ... > Also, an error log window would be cool (it could be global so all > browser windows could output to it). Then I can see why the image > didn't load or get more verbose error messages. You could use > simple printfs to the console for this, but I think this way is > better. > Warnings an errors are reported to stderr, so you can read them on there term where you launched dillo. However, that's true dillo's status indications needs some work. > * Images > Is alpha handled properly on images? I have seen some black borders > round icons. > See bug-track engine : transparency is supported for GIF images, but not for PNG. > * Downloads > Rather than having loads of dialogs like in Netscape, a single window > with a list of downloads and status would be cool (imagine a more > simple GetRight style window). > There's no download at all for now : where're all thinking about it, but the feature and the actual implementation have yet to but discussed. Your suggestion seem to be what most people want, though. Netscape for macintosh does this. > * CSS > CSS can be used well like in Mozilla for preferences e.g. you store > default font/colors in a user.css file and these settings "cascade" > down. > This have been discussed before, and rejected for now for two reasons : - CSS support seems very far to be implemented ; - CSS can't handle every option we want (page-rendering only). > There's also no "About" dialog. :) > Right ! Why don't you make one, just to train your skills on Dillo ? :-) > Anyway, these are just a few ideas I have, although some of them > would obviously be in the future. I'm going to look at adding CSS > support as soon as I untangle the source code a little. :) > You're welcome. Don't hesitate to ask any (stupid or not ;-) question to the list, but make sure you've read everything on dillo's homepage, documentation and mail list archive. ----------------------------------- Eric GAUDET Le 24-Nov-2000 a 22:13:29 "Le verbe inexister ... inexiste !" ----------------------------------- [Dillo-dev][patch] Dillo-plugins final From: Eric GAUDET - 2000-11-24 12:12 Hi all, I just released the final version of my patch for Dillo-plugins for you "I-want-the-last-patch" freaks, and also for those who want to give it a try and send me some comments or bug repports. Visit my dedicated web page, now permanent at : http://www.rti-zone.org/dillo/ ----------------------------------- Eric GAUDET Le 24-Nov-2000 a 21:09:42 "Le verbe inexister ... inexiste !" ----------------------------------- RE: [Dillo-dev]Suggestions From: Sean 'Shaleh' Perry - 2000-11-24 06:28 > > * CSS > CSS can be used well like in Mozilla for preferences e.g. you store > default font/colors in a user.css file and these settings "cascade" down. > I have suggested this. Jorge is not 100% convinced. Plus dillo needs some rework to handle css. html items have no storage of names. So you can't do
. [Dillo-dev]Suggestions From: Will Newton - 2000-11-24 03:31 Here are a few ideas/suggestions I have: * Errors Errors IMO should appear in the main window and not in any dialogs like in Nestcape. Dillo does this, I know, but I wanted to make sure it stays like that. :) Also, an error log window would be cool (it could be global so all browser windows could output to it). Then I can see why the image didn't load or get more verbose error messages. You could use simple printfs to the console for this, but I think this way is better. * Images Is alpha handled properly on images? I have seen some black borders round icons. * Downloads Rather than having loads of dialogs like in Netscape, a single window with a list of downloads and status would be cool (imagine a more simple GetRight style window). * CSS CSS can be used well like in Mozilla for preferences e.g. you store default font/colors in a user.css file and these settings "cascade" down. There's also no "About" dialog. :) Anyway, these are just a few ideas I have, although some of them would obviously be in the future. I'm going to look at adding CSS support as soon as I untangle the source code a little. :) RE: [Dillo-dev]Status From: Sean 'Shaleh' Perry - 2000-11-24 03:17 On 24-Nov-2000 Will Newton wrote: > > I got the latest beta of Dillo (0.3.0) and it seems pretty fast, but no > tables means it can't render much. Is tables the next feature to be added? > CSS support would also be cool. I'm currently looking through the source > to see where I can help out. > Once the new Dw widget appers, I will begin integrating my work on html.c. I have nearly all font affecting tags working. All that is left is tables and frames, and tags that do not directly affect rendering like abbr. If I recall, someone else is working on div. If you would like to make a css parser, that would be handy. CSS support will be a while, there are more pressing issues. FW: RE: [Dillo-dev]dillo 0.3.0 crashes at startup From: Eric GAUDET - 2000-11-24 01:26 Robert investigated the problem (bug#97) a little more. It appears to be a gcc 2.95.2 bug for K6 with -O2 option. I heard about that for non-i386 cpu (SH4, ARM). It seems -O2 is not reliable for now, except for intel processors. We should put a warning in the README about this. -------------------------- FW -------------------------------- Program received signal SIGSEGV, Segmentation fault. Dw_border_size_nego_y (dw=0x8054290, allocation=0x80b0a18) at dw_border.c:64 64 dw->allocation = *allocation; locals: dw = (Dw *) 0x8054290 allocation = (DwRect *) 0x80b0a18 > -O2 seems to pose problems in 2.95.2. Can you try -O or -O0 (O-zero) > with k6 ? "-O2 -mcpu=pentium" works fine. "-O0 -mcpu=k6 -march=pentium -g3" works fine also. "-O -mcpu=k6 -march=pentium -g3" works fine as well, but "-O2 -mcpu=k6 -march=pentium -g3" doesn't work. Then again "-O6 -mcpu=pentium -march=pentium -g3" DOES work. I think setting -cpu=k6, or ever worse, -march=k6 (had problems with this one before..) causes some type of code to be miscompiled in some cases when you use -O2. If -O2 fails at compile-time (for me) it starts to loop infinitly. In these cases only -O0 works, so 2.95.x is kinda buggy. I hope the 3.0 series will be better. > -g is always a good idea for debugging ;-) Yes, and you can always strip the binary when your'e done. By the way, "-W -Wall" is a LOT better at catching dirty code than just -Wall. Try it and see for yourself :). Put a note of this in some FAQ somewhere so that you can tell the next victim of this gcc bug to "read the friendly manual" :). Thanks, Robert. --------------End of forwarded message------------------------- ----------------------------------- Eric GAUDET Le 24-Nov-2000 a 10:17:57 "Le verbe inexister ... inexiste !" ----------------------------------- [Dillo-dev]Status From: Will Newton - 2000-11-24 00:52 I got the latest beta of Dillo (0.3.0) and it seems pretty fast, but no tables means it can't render much. Is tables the next feature to be added? CSS support would also be cool. I'm currently looking through the source to see where I can help out. RE: [Dillo-dev]dillo 0.3.0 crashes at startup From: Eric GAUDET - 2000-11-23 01:21 What version of dillo ? Can you send a stack dump after the crash ? Are you on a i386 arch ? if so, can you send me your dillo binary ? PS: the list is very low traffic, if you register you won't be annoyed by more than 3 or 4 messages a week. -- En reponse de "[Dillo-dev]dillo 0.3.0 crashes at startup" de Robert Holmberg, le 22-Nov-2000 : > Ahh..CRAP! Just submitted bug #97 with too many characters and only > half the bug description was submitted.. here we go: > > Dillo crashes on startup (gcc 2.95.2, glibc 2.1.2) after showing an > empty > window. It does show contents (menu, toolbar..) in the window if run > through gdb: > > (gdb) run > Starting program: /usr/local/bin/dillo > dillo_dns_init: Here we go! > Loading bookmarks... > [New Thread 3591 (manager thread)] > [New Thread 3590 (initial thread)] > [New Thread 3592] > > Program received signal SIGSEGV, Segmentation fault. > 0x8061c00 in Dw_border_size_nego_y () > > Please CC questions to me, I'm not on the list. > > Robert. > _______________________________________________ > Dillo-dev mailing list > Dillo-dev@li... > http://lists.so....net/mailman/listinfo/dillo-dev ----------------------------------- Eric GAUDET Le 23-Nov-2000 a 10:17:40 "Le verbe inexister ... inexiste !" ----------------------------------- [Dillo-dev]Dillo 0.3.0 crash on starup. From: Robert Holmberg - 2000-11-22 21:22 Hust submitted bug #97 w. more than 256 characters in description in the bugtarcker. Here's the whole deal: Dillo 0.3.0, gcc 2.95.2, glibc 2.1.2, linux 2.4-test11: (gdb) run Starting program: /usr/local/bin/dillo dillo_dns_init: Here we go! Loading bookmarks... [New Thread 3621 (manager thread)] [New Thread 3620 (initial thread)] [New Thread 3622] Program received signal SIGSEGV, Segmentation fault. 0x8061c00 in Dw_border_size_nego_y () Dillo does draw it's window first. I'm NOT suscribed to the list. Robert. [Dillo-dev]dillo 0.3.0 crashes at startup From: Robert Holmberg - 2000-11-22 21:17 Ahh..CRAP! Just submitted bug #97 with too many characters and only half the bug description was submitted.. here we go: Dillo crashes on startup (gcc 2.95.2, glibc 2.1.2) after showing an empty window. It does show contents (menu, toolbar..) in the window if run through gdb: (gdb) run Starting program: /usr/local/bin/dillo dillo_dns_init: Here we go! Loading bookmarks... [New Thread 3591 (manager thread)] [New Thread 3590 (initial thread)] [New Thread 3592] Program received signal SIGSEGV, Segmentation fault. 0x8061c00 in Dw_border_size_nego_y () Please CC questions to me, I'm not on the list. Robert. [Dillo-dev]BUG#95 From: Jorge Arellano Cid - 2000-11-22 01:27 Hi! I hope you're enjoying the 0.3.0 release! (I haven't received complaints) There're still several things to do, and I'll try them one by one... I'm still concerned with BUG#95, and followed an interesting thread at http://www.flora.org/lynx-dev/lynx-dev/9608/0278.html Even more, further research showed me that this issue is far from being easy to solve, but the "trend" seems to use what Fote suggested in the previous thread (0283.html). The page suggested in the bugtrack engine has changed to this standard, so now its browsable with dillo without any change! And I also remember that some time ago, sourceforge included directions for displaying their logo with an image URL that included a & sequence, but not anymore! So it seems that things are getting straight in an attempt to keep CGI functionality working ('&' in that context is a reserved character to separate name=value pairs, and any use of it for other purpose should be hex escaped). Please send me any comments, specially if you can prove it MUST be handled in another way, otherwise I'll delete the entry within a few days. Jorge.- Re: [Dillo-dev]Font issues From: Sebastian Geerken - 2000-11-16 01:59 Eric GAUDET wrote: > > I've just upgraded Netscape to 4.75 (mandrake RPM) which _needs_ the > package mozilla-fonts. These fonts seem to be buggy : the > -helvetica-medium-italic- is a bitmap 18pts font looking bold, and it's > ugly when resized in 12 by Dillo for regular italic. > One way around is to change Dillo to use "oblique" instead of "italic": > > in dw_page.c (v0.3.0) > in static GdkFont *Dw_page_realize_font(DwPageFont *font) > line 1115: > - font->italic ? "i" : "r", > + font->italic ? "o" : "r", And a few lines below: if (font->font == NULL && font->italic) { - /* Some italic fonts (e.g. Courier) use "o" instead of "i". */ - sprintf(fontname, "-*-%s-%s-o-*-*-%d-*-75-75-*-*-*-*", + /* Some italic fonts (e.g. Courier) use "i" instead of "o". */ + sprintf(fontname, "-*-%s-%s-i-*-*-%d-*-75-75-*-*-*-*", font->name, font->bold ? "bold" : "medium", font->size); font->font = gdk_font_load(fontname); } > That looks like an extreme solution to change Dillo when a font is > buggy, but it works. (BTW, my adode and urw helvetica don't have italic > face, only oblique : X must be switching to oblique when italic's not > found ; does any distri have helvetica-oblique ?) It's dillo, which switches to oblique. > A better way would be to make that an option in dillorc. > Perhaps Sean has something better with his fonts handling (or a better > ideas) ? Oblique and italics are two different variants, actually you get an oblique variant by shearing a font, but an italic variant is a totally new font. Helvetica Italics is quite unusual, as well as e.g. Times Oblique. On my system, only Adobe Courier provides both variants. However, HTML does not distinguish between these variants, but CSS does. Personally I prefer oblique variants, since they are more readable at low resolutions. A option letting the user decide which variant dillo should prefer, when both are available, would imho be a good possibility. Sebastian Re: [Dillo-dev]where did view source go? From: Eric GAUDET - 2000-11-15 09:20 -- En reponse de "Re: [Dillo-dev]where did view source go?" de Sebastian Geerken, le 14-Nov-2000 : >> aaarrggghhh, where did view source go? how can I tell why dillo >> chose to >> render a web page like it did if I can not see the html? > > Yes, somehow a complete menu got lost ... commented in > a_Menu_mainbar_new. It's in the popup menu now. ----------------------------------- Eric GAUDET Le 15-Nov-2000 a 18:19:45 "Le verbe inexister ... inexiste !" ----------------------------------- Re: [Dillo-dev]where did view source go? From: Sean 'Shaleh' Perry - 2000-11-15 07:13 > > BTW, never wondered about the term "View Source"? Actually, there > *aren't* sources for html documents (except pl, php, asp ...), the > menu entry should be named "View As Plain Text". > "source" as an abstraction. It may be cgi generated, static, etc. Perhaps "View Markup"? Re: [Dillo-dev]where did view source go? From: Sebastian Geerken - 2000-11-14 20:49 Sean 'Shaleh' Perry wrote: > > aaarrggghhh, where did view source go? how can I tell why dillo chose to > render a web page like it did if I can not see the html? Yes, somehow a complete menu got lost ... commented in a_Menu_mainbar_new. BTW, never wondered about the term "View Source"? Actually, there *aren't* sources for html documents (except pl, php, asp ...), the menu entry should be named "View As Plain Text". Sebastian [Dillo-dev]Bug#95 : not a bug ! From: Eric GAUDET - 2000-11-14 13:07 I just parsed the bugtrack engine, and bug#95 puzzled me : I never heard of entities in URLs before, even in the RFC. The only official escape mecanism for URLs is %XX where X is hexadecimal ; non escaped characters are UTF-8 (which is very close to ISO8859-1 ?). HTML4 specs has a note about that : some old browsers accept this undocumented URL parsing, and some bad pages rely on that. ----------------------------------- Eric GAUDET Le 14-Nov-2000 a 21:55:49 "Le verbe inexister ... inexiste !" ----------------------------------- RE: [Dillo-dev]where did view source go? From: Eric GAUDET - 2000-11-14 12:15 -- En reponse de "[Dillo-dev]where did view source go?" de Sean 'Shaleh' Perry, le 14-Nov-2000 : > aaarrggghhh, where did view source go? how can I tell why dillo > chose to render a web page like it did if I can not see the html? Would somebody be see a problem make view source a Dillo-plugin ? ----------------------------------- Eric GAUDET Le 14-Nov-2000 a 20:21:20 "Le verbe inexister ... inexiste !" ----------------------------------- [Dillo-dev]where did view source go? From: Sean 'Shaleh' Perry - 2000-11-14 08:13 aaarrggghhh, where did view source go? how can I tell why dillo chose to render a web page like it did if I can not see the html? [Dillo-dev]dillo does not give enough user feedback From: Sean 'Shaleh' Perry - 2000-11-14 08:07 If I hit a slow site, I get no visual clue that dillo is doing anything. To see this first hand, go to bugs.debian.org, and look up the package lintian. dillo seems to hang, then magically load the page. Re: [Dillo-dev]New 0.3.0 release From: Sean 'Shaleh' Perry - 2000-11-14 07:42 > I downloaded it a few hours ago, and patched the changes into my > version (a bit "handwork" was necessary). I already tested > third-button-menus and some new tags implemented by Jörgen (, > and ), and that seems to work. > awesome, those were the last three tags. I thing once my changes are merged into dillo the only tags not supported are the non display tags (like abbr or address) and sub and sup. RE: [Dillo-dev]Font issues From: Sean 'Shaleh' Perry - 2000-11-14 07:41 > A better way would be to make that an option in dillorc. > Perhaps Sean has something better with his fonts handling (or a better > ideas) ? > font handling is actually the next thing on my list. Hmm, I have Xfree 4.0.1 and all still seems well. [Dillo-dev]Font issues From: Eric GAUDET - 2000-11-14 02:01 I've just upgraded Netscape to 4.75 (mandrake RPM) which _needs_ the package mozilla-fonts. These fonts seem to be buggy : the -helvetica-medium-italic- is a bitmap 18pts font looking bold, and it's ugly when resized in 12 by Dillo for regular italic. One way around is to change Dillo to use "oblique" instead of "italic": in dw_page.c (v0.3.0) in static GdkFont *Dw_page_realize_font(DwPageFont *font) line 1115: - font->italic ? "i" : "r", + font->italic ? "o" : "r", That looks like an extreme solution to change Dillo when a font is buggy, but it works. (BTW, my adode and urw helvetica don't have italic face, only oblique : X must be switching to oblique when italic's not found ; does any distri have helvetica-oblique ?) A better way would be to make that an option in dillorc. Perhaps Sean has something better with his fonts handling (or a better ideas) ? ----------------------------------- Eric GAUDET Le 14-Nov-2000 a 10:47:37 "Le verbe inexister ... inexiste !" ----------------------------------- Re: [Dillo-dev]New 0.3.0 release From: Sebastian Geerken - 2000-11-13 14:38 Jorge Arellano Cid wrote: > > Hi! > > The sourceforge shell acccount server is back and dillo-0.3.0 > is there, ready for download. > > Note that there're still some patches in my queue and they'll > have to wait until 0.3.1 (This release is a long delayed one and > I didn't wanted to procrastinate it even more). > > I hope you enjoy this release, but bear in mind that 0.3.0 > is an intermediate release before 0.3.1. > > Jorge.- > > Ps: I'll be expecting some feedback! I downloaded it a few hours ago, and patched the changes into my version (a bit "handwork" was necessary). I already tested third-button-menus and some new tags implemented by Jörgen (, and ), and that seems to work. I hope I'll be able to "release" a slow, buggy, but basicly working Dw, integrated in dillo 0.3.0, within the next few days. Sebastian [Dillo-dev]New 0.3.0 release From: Jorge Arellano Cid - 2000-11-13 13:01 Hi! The sourceforge shell acccount server is back and dillo-0.3.0 is there, ready for download. Note that there're still some patches in my queue and they'll have to wait until 0.3.1 (This release is a long delayed one and I didn't wanted to procrastinate it even more). I hope you enjoy this release, but bear in mind that 0.3.0 is an intermediate release before 0.3.1. Jorge.- Ps: I'll be expecting some feedback! [Dillo-dev]dillo-0.3.0 From: Jorge Arellano Cid - 2000-11-12 18:07 Hi! The dillo-0.3.0 version is ready and waiting for release. The funny thing is that the sourceforge server (shell accounts) refuses foreign connections! I've been trying for a couple of days and also wrote the support staff there asking for some news. When the server comes up, 0.3.0 will be there. Jorge.- RE: [Dillo-dev]CVS update From: Eric GAUDET - 2000-11-10 02:51 -- En reponse de "[Dillo-dev]CVS update" de Jorge Arellano Cid, le 09-Nov-2000 : > > Hi! > > We're getting closer to 0.3.0 release. I just updated the CVS > server (and bug-track engine) with the newest version. > I'll post the complete/clean Dillo-plugins patch against v0.3.0 then (this should make it easier for you) > - The horizontal rulers patch (worked by Eric and Allan). Which > one is the latest patch? Allan says its patch takes into > account feedback from Jörgen and Eric. And I also have a > patch from Eric :-) > Please tell me which one to apply or send me a new one > against the CVS code. > I know nothing about Allan's patch. Mine is dated Aug 26 and I haven't touched it since. Mine is complete though : all attributes are recognized and redendered,
before and after the
(see more details on my page). > - Bug#80 worked by Eric. The Web site doesn't contain a valid > diff file for it (please send me that patch Eric). > Oops, sorry. Site updated. As it's a tiny patch, it's also in this mail. > - For BUG#27. Which test pages did you use Eric? Please send me > some URLs. > It's in the Html.testsuite. (There was no link from the index to the page, but the page was here along with the images, sorry : updated on my page). Link : Image Maps PS: patch for bug #80: (my mailer is puting some \n in the way, prefer my page's version :-/) diff -pur dillo-0.2.4/src/dw_page.c dillo-0.2.4.bug80/src/dw_page.c --- dillo-0.2.4/src/dw_page.c Tue Aug 22 00:54:09 2000 +++ dillo-0.2.4.bug80/src/dw_page.c Sun Aug 27 22:15:18 2000 @@ -34,7 +34,7 @@ #include "prefs.h" #include "commands.h" -#define DW_GET_BW(page) (BrowserWindow *)((page)->status_data) +#define DW_GET_BW(page) (page->status?(BrowserWindow *)((page)->status_data):NULL) /* * Set GtkDwScroller::anchor_pos value. @@ -670,9 +670,9 @@ static void Dw_page_handle_event(Dw *dw, menu_popup.info.url = page->links[link_pressed].url; gtk_menu_popup(GTK_MENU(menu_popup.menu_over_link), NULL, NULL, NULL, NULL, button->button, button->time); - } else { - if (bw->nav_stack_size == 0) - return; + } else if (bw) { + if (bw->nav_stack_size == 0) + return; menu_popup.info.title = bw->nav_stack[bw->nav_stack_ptr].title; menu_popup.info.url = bw->nav_stack[bw->nav_stack_ptr].url; ----------------------------------- Eric GAUDET Le 10-Nov-2000 a 11:34:29 "Le verbe inexister ... inexiste !" ----------------------------------- [Dillo-dev]CVS update From: Jorge Arellano Cid - 2000-11-09 16:45 Hi! We're getting closer to 0.3.0 release. I just updated the CVS server (and bug-track engine) with the newest version. Please note how important is to use the bug-track engine and not to do silent patching. The new Changelog (in CVS) shows a lot of small patches some of whom were not recorded in the bug-track. Please avoid this kind of situation. On the progress side, the new code hasn't solved all the problems related with link-following before the page arrives, but certainly will not crash if you follow a different link while still downloading the referer page! (I hope to finish that in 0.3.1 release). I also started moving the patch queue, and I have a few doubts: - The horizontal rulers patch (worked by Eric and Allan). Which one is the latest patch? Allan says its patch takes into account feedback from Jörgen and Eric. And I also have a patch from Eric :-) Please tell me which one to apply or send me a new one against the CVS code. - Bug#80 worked by Eric. The Web site doesn't contain a valid diff file for it (please send me that patch Eric). - The Find text patch (originally from Sam, patched by Sebastian and repatched by Allan). Allan, is your final version a stable feature. I ask because Sam warned us of some bugs in the patch. I guess they were worked out by following patches, and you being the last one to work out the code, and most probably the one who has tested it more extensively, may have the best informed answer. - For BUG#27. Which test pages did you use Eric? Please send me some URLs. Jorge.- PS1: Having a user manual (an HTML page) has started to be necessary. A very concise guide. Volunteers? PS2: There's a small bug within a_Url_squeeze. When there're too much "/.." sequences in the URL, it removes the server! (a patch for this BUG, and its bug-track-entry would, be highly appreciated too). PS3: If someone takes the time to update the bug-track with the missing entries (found in the CVS-Changelog) that'd be very helpful too. [Dillo-dev]Dillo Weekly News From: Allan Clark - 2000-11-05 15:45 Just posted you can view it here http://www.shark.pwp.blueyonder.co.uk/news/news051100.html Allan Clark allan@al... shark@bl... [Dillo-dev]Internal release From: Jorge Arellano Cid - 2000-11-03 22:49 Hi! The internal release has a new code layer that's expected to ease the fix of the reloading, double clicking and every related problem. Those fixes are not there yet! I just wanted to give this prototype a little testing before adopting it. It was a 100Kb diff file! Jorge.- Re: [Dillo-dev]Internal release From: Sebastian Geerken - 2000-11-02 14:48 Hi! I finished a first version of dillo, with the new Dw integrated. It compiles, and runs halfway, with thousands of warnings and criticals (but no crash), furthermore some features are still missing (mouse events, colors etc). Eric GAUDET wrote: > I can't say I found it more stable yet on the multiple click on links > side : still crashing. Wasn't it supposed to fix this ? (or is it that > it's suppose to _ease_ the fixing of this ?) I wrote a test php3 page (I'll send it Eric, if you wish to), which prints the current time (to check wheater reloading works), and has a few delays in it (to simulate a slow connection), and tested it with three versions: 0.2.4, 0.3.0-pre1, and 0.3.0-pre1 with the new Dw integrated. Testing results have been always the same. Reloading after transmission is complete, works perfectly for all versions, interrupting transmission by reloading didn't work correctly in any version: 0.3.0-pre1 crashed, and 0.2.4 as well as 0.3.0-pre1 + new Dw *both* did not crash, but displayed the old page (with the old time). When debugging 0.3.0pre1, I found that the segmentation fault happens in a_Dw_page_update_begin, with argument page == NULL. I wasn't able to examine it more, since the stack was destroyed, but this reminds me somehow of bug #77 (where there was somewhere a NULL pointer, and the crash could be prevented by a simple check). When anchors work in the new Dw, it could be a good idea to check whether bug #77 still arises. BTW, I tried to insert a few checks, but didn't get it run neither. Sebastian Re: [Dillo-dev]Internal release From: Sebastian Geerken - 2000-11-02 10:36 Jorge Arellano Cid wrote: > [...] > Sebastian: you can use this version in your effort to integrate > the new Dw (or if you prefer, you can wait the testing news). > Anyway, from Dw's perspective, changes to the interface are > pretty minimal!) Ok, I'll integrate it into 0.3.0. (Should not be too hard to patch it into any version.) > -- > > Sean: > > > Jorge, would you like for this to happen after Sebastian's > > Dw changes? > > Yes, please wait until Sebastian integrates the new Dw. > An entry in dillorc to disable FONT processing looks OK to > me. (FONT tag is still a tricky issue: please be careful). I suppose, Sean's changes are mainly in the Html module, and since there aren't much changes in the interface of DwPage, that should not matter. Anyway, after integration, there will be still much to do, my current plan is: 1. integration (will not take too much time), then 2. add missing features, and 3. bugs fixing, optimization, comments etc. I could send you a unfinished version after 2, which will be stable, with defined interfaces, but still a bit buggy and slow. Sebastian RE: [Dillo-dev]Internal release From: Eric GAUDET - 2000-11-02 03:53 -- En reponse de "[Dillo-dev]Internal release" de Jorge Arellano Cid, le 01-Nov-2000 : > > Hi! > > I just uploaded dillo-0.3.0-pre1 to the sourceforge server. > This internal release is mainly intended for testing. If it > proves to be at least as stable as 0.2.4, it will be the base for > 0.3.0 release. > I can't say I found it more stable yet on the multiple click on links side : still crashing. Wasn't it supposed to fix this ? (or is it that it's suppose to _ease_ the fixing of this ?) > Those of you that feel interested, please download it and let > me know what you think about it (compared with 0.2.4). I'd > appreciate your comments, specially if new BUGs arise. > > Most probably 0.3.0 will be this code-base plus a few patches, > and a couple of weeks later 0.3.1 will integrate the rest of the > patch queue. > I'll update the Html.testsuite indications when 0.3.1 comes out then. ----------------------------------- Eric GAUDET Le 02-Nov-2000 a 12:44:29 "Le verbe inexister ... inexiste !" ----------------------------------- RE: [Dillo-dev]Internal release From: Sean 'Shaleh' Perry - 2000-11-02 02:48 > > Sean: > > > Jorge, would you like for this to happen after Sebastian's > > Dw changes? > > Yes, please wait until Sebastian integrates the new Dw. > An entry in dillorc to disable FONT processing looks OK to > me. (FONT tag is still a tricky issue: please be careful). > will wait for Sebastian then. Gives me some time for some other touch ups. [Dillo-dev]Internal release From: Jorge Arellano Cid - 2000-11-01 16:43 Hi! I just uploaded dillo-0.3.0-pre1 to the sourceforge server. This internal release is mainly intended for testing. If it proves to be at least as stable as 0.2.4, it will be the base for 0.3.0 release. Those of you that feel interested, please download it and let me know what you think about it (compared with 0.2.4). I'd appreciate your comments, specially if new BUGs arise. Most probably 0.3.0 will be this code-base plus a few patches, and a couple of weeks later 0.3.1 will integrate the rest of the patch queue. Source code location: http://dillo.so....net/dillo-0.3.0-pre1.tgz -- Sebastian: you can use this version in your effort to integrate the new Dw (or if you prefer, you can wait the testing news). Anyway, from Dw's perspective, changes to the interface are pretty minimal!) -- Sean: > Jorge, would you like for this to happen after Sebastian's > Dw changes? Yes, please wait until Sebastian integrates the new Dw. An entry in dillorc to disable FONT processing looks OK to me. (FONT tag is still a tricky issue: please be careful). Jorge.-