- Dillo-dev - lists.mailman3.com

building on AIX, gcc, dillo 0.7.3
by Riccardo Jan. 27, 2004

Jan. 27, 2004

Hwllo, I sent a private message, of which a reply was CC to this list AFAIK, about building dillo on my AIX system. The problem are threads. Pth can't be used and so I need the system's native threads. A solution was offered me but another email asked for more information, so I report first this information here before trying blindly. configure reports: checking target system type... rs6000-ibm-aix4.2.1.0 as far the system concerns. The error during configure is: checking whether threads work with -pthread... no. Now we will try some libraries. checking for library containing pthread_create... no configure: WARNING: *** No pthreads found. *** gcc -v yields: Reading specs from /usr/local/lib/gcc-lib/rs6000-ibm-aix4.2.1.0/2.95.3/specs gcc version 2.95.3 20010315 (release) while lslpp -l | egrep -i 'thread|xlc|gcc' bos.rte.libpthreads 4.2.1.0 COMMITTED pthreads Library xlC.cpp 3.1.4.0 COMMITTED C for AIX Preprocessor xlC.rte 3.1.4.0 COMMITTED C Set ++ for AIX Application greets, Ric

3 3

[Dillo-dev]Buggy HTML, the new parser and webmasters
by Andreas Schweitzer Jan. 27, 2004

Jan. 27, 2004

Hi all, First : my apologees for a rather large mail ! Now, that the bug counter is in dillo, and with Madis' last post, it's a good coincidence that I just tried to convince webmasters of 3 sites I usually visit to fix their HTML. From 2 I haven't heard a thing (and they are not even THAT huge), but from www.osnews.com I got a reply very fast - and it turned out to be an extremely interesting conversation. I asked her for permission to forward it to the list, and here it goes. I sorted it chronolically, with some comments of mine in between. I omitted my first mail - it was just the HTML warnings of the new parser along with a kind request to fix the page. WARNING, BIG REMINDER : Don't flame her - it's not about this individual site, it's about general practice and the general attitude towards standards ... After all, the e-mail conversation started in privat and she didn't have in mind that a lot of people would read it. And I think she deserves credit for taking her time and lay out her arguments and explaining everything in a very friendly manner. On Sat, Jan 24, 2004 at 06:36:27PM -0800, Eugenia Loli-Queru wrote: > Hello Andreas, > thanks for the heads up. > I looked at the error messages, and I know what is wrong. Please note that > all forms and tables ARE closing in the code, it is just that your parser is > not realizing it. > You see, I start the form inside a table's cell, but I close the form > outside the table. This is not 100% valid code, but it is like that for a > reason. This is causing your parser to spit all these errors (all the errors > below the form error are just a side effect of the first error). > So, the reason I close the form outside that table is because I am trying to > go around bugs on Netscape 4 and IE. As you can see, the cell that holds the > form has a black background color. If I close the form inside that cell, IE > and NS4 create a new paragraph inside that cell, even if my code never asked > for it. And that looks ugly. Up to here, you could argue a lot about how to design a web page, to not wanting to dictate how the page looks and leave it to the browser, to us CSS and whatnot. With enough arguing one could even convince this very author to change (not likely though, see below). BUT : the interesting part comes now : > So, the "trick" that many web developers of my time use a lot (see: > hand-written HTML), is to close the </FORM> outside of tables or places > where it screws up the design. It is a well-known trick, and so your new > parser is going to meet it a number of times in the future. I suggest you The above is frightening .... > fix your parser to understand "non-normal" code like that, and realize that > the form does close as expected. Of course, for us that's not fixing but breaking ;-) .... Here is now my reply (rather long, and you will know the arguments if you follow the lamiling list) : > Hi Eugenia, > > Thanks for the reply and explanation. > First of all : I am not a core developer of that browser - just > a steady user and occasional contributor. > > However, if this mail came from one of the developers or if I told > them to follow your advice, you would either be ignored or flamed ;-) ... > I will try to do neither. So please do keep in mind that my response > is meant as a friendly food for thought or possibly some constructive > criticism. > > Discussions about fixing broken HTML do come up on the mailing list > from time to time. And the consent is to just not do it - for a number > of reasons (I probably will forget some). > - what good are standards if nobody conforms to them, and instead an > own (not very strict !) psuedo-standard develops ? > - It's not a browser's job to guess what the author wanted to have displayed. > - It costs man-power and CPU cycles to detect and fix broken HTML. > We prefer to spend the man-power on developing something useful instead > of broekn HTML-detection code. And we prefer to give the CPU cycles > to the user (BTW, dillo is the snappiest grafic browser I know - maybe Links2) > > All right, but we already detect the problem of your page, you may say :-) .. > and indeed, the *old* parser was closing out of order. In fact, I could > configure dillo to still use the old parser and your page will look great. > BUT : then other pages break :-( - they have buggy HTML, too. And there, > closing tags with the new method is better then with the old method. > So, in chosing which parser to use, the default is the standards complying > method of interpreting HTML (also, the old parser was buggy and not quite > right). Oh - and if we care about the web sites we also tell/told the > authors if these other pages that they have buggy code ! > Now, why then are Mozilla or the other browsers displaying all correctly ? > Probably because they did spend man-power and CPU-cycles to guess what the > author wanted. > I had to smile when you wrote, "Fix your parser to understand non-normal HTML", > because I thought, "That's not fixing it, that's breaking it". > > When I design web pages, I usually make them pass (if possible) > validator.w3.org and if they look ugly in ancient - and buggy ! - browsers, > bad luck ! I realize, my pages are nothing professional like your site, > but still, is it worth to support buggy broswers ? How many visitors > are actually using Netscape 4 ? Probably more than dillo, but the comparison > should be non-buggy vs. buggy browsers. About IE - are all versions > of IE bahving like this ? Then my argument would be, why support MS's > interpretation of standards ? (and outside of HTML you probably will > agrre ;-) ... after all, don't we all know lots of alternatives and fight > for them ?) You as a professional web site probably do have to support > buggy broswers, but the question still may be asked. > > Finally, some links to archived previous discussions : > http://news.gmane.org/group/gmane.comp.web.dillo.devel/thread=1324/force_lo… > (if you don't want to read all of it, read at least Eric Gaudet's posts) > And here is the explanation of the new parser : > http://article.gmane.org/gmane.comp.web.dillo.devel/1434 > > By now, you know that the other problematic site was slashdot ;-) ... > (among others !) which displays ok with the current parser - but also is > buggy HTML - and they know it ! > > And before I forget : Yes, I saw a lot of pages that look quite ugly > with the strict interpretation. And I also only tell the webmasters > of pages for which I care :-) > > And as a really last word : the new parser may indeed have room for > improvement, and some more closing may indeed be possible as a compromise, > BUT, as I said, I'm not a core developer. > > As I said in the beginning, this was meant as food for thought, not flaming. > If you should decide to change HTML after all, that would be fantastic. If > not, hopefully I at least did explain myself well. > > Cheers > Andreas And here is her 2nd answer: > Thank you for the email Andreas, > however, I must say that the code in that form won't change. IE has more > than 70,000 viewers daily on osnews, and Dillo has about 20. So, as you can > understand, being compatible with the broken IE is more important to us, > even if that means that we have to not be validated. In fact, not being > validated, it is a given for big sites, because in order to give a > specific/certain look, you have to do tricks like the one I described to you > in order to render the same on all big browsers. Now you could, of course, go and say "screw this site" .... but the following is at least ineresting food for thought. I don't fully agree, but I can understand the point of view. > Also, please put in mind: my husband is in the browser business. He is a web > browser developer for OpenWave.com among his team. They don't have the > specific problem of out-of-order-tag-closing because they know that many > sites are like that, so they have to go around this problem, OR they won't > be able to sell their browser to anyone. > > This is a market. Even for the free Dillo. Dillo's job is not create the > most "correct" parser in the world, but to render most sites in the world. > If the first was true, then Dillo is nothing but an "Amaya", a test, a proof > of concept. But if the second is true, then Dillo is really a usable by > people browser. > > So, depends what its developers want. If they want a proof of concept, no > problem with me, osnews and most other sites won't work. If they want a > browser that works for most people, they need to sit down and go around > buggy HTML (created by either amateur developers or from developers that > they had to go around bugs). For example, IE even goes around this bug <a/> > and understands that the developer wanted to write </a> instead. Big > browsers do such error checking and go around buggy HTML, because it is the > only way they can be widely used. In my case of course, it was me that I had > to go around their bugs. > > So, it all depends what the Dillo people want. It is not what I, or you, > want, but what their real targets are: proof of concept project or a real > USABLE browser. > > If that was me, it would have being the second. And it ain't my fault that > IE has these bugs that I have to go around, it is my job to make sure my > page renders well for most of my readers. :) > > Rgds, > Eugenia > > --- > Editor in Chief > http://www.OSNews.com When I asked her for permission to forward her mails to the list she added another intersting quote : On Sun, Jan 25, 2004 at 08:53:17PM -0800, Eugenia Loli-Queru wrote: > You can go ahead and quote some of the stuff I said. > BTW, I discussed the matter with my husband (who as I said is in the browser > business) and he told me: > > "Creating an HTML browser and a Web browser are two different things. The > first one just needs to do the job right and parse HTML correctly, the other > one just has to be able to render web pages in any possible mean. These are > two different beasts and targets --unfortunately." > > I am afraid this is the reality of any product though, not just browsers... > Reality market... :( > > Eugenia > Bottom line : it may be possible to convince individual sites to fix their HTML, but what about really big sites that use these ... err ... "workarounds" ? Do they really care how a little browser interprets their HTML ? They care if HTML looks good in IE - reality sucks. It's extremely frustrating. And even more so, when I have to look at garbeled pages ;-) .... The situation may improve over time when more and more sites use CSS and don't have to rely on such dirty formating tricks - but then again, how many bugs has IE there, that authors want/will work around ? And to repeat my reminder : Don't flame her - it's not about this individual site, it's about general practice and the general attitude towards standards ... And again I think she deserves credit for taking her time and lay out her arguments and explaining everything in a very friendly manner. Cheers, Andreas -- **************************** NEW ADDRESS ****************************** Hamburger Sternwarte Universitaet Hamburg Gojenbergsweg 112 Tel. ++49 40 42891 4016 D-21029 Hamburg, Germany Fax. ++49 40 42891 4198

3 2

worried about dillo (cathedral and bazaar)
by linux seaq Jan. 27, 2004

Jan. 27, 2004

Hi, i want to express my gratitude with the dillo developers, this is a really nice product, i use it everyday. I use mozilla only for banking, hotmail and flash sites. I'm only a user i'm not a developer, so my point of view is mainly from an user's perspective. Since dillo's win32 fork and previous discussion, it was clear to me that dillo development was, in some way, closed. That odd feeling was confirmed by the "memo" sent by main developers about patches and additions to dillo. Since that mail, all the "foreign" contributors stopped to share their improvements to dillo project.Or it seems to me at this time (PLEASE PROVE ME WRONG) That's really sad 'cause what i'm seeing right now is some kind of slowness on the development of dillo, and somehow is getting behind in needed functionality (frames,https,downloads,images, a lot of stuff out there). - from an user point of view - I recently re-read the cathedral and bazar paper, and thought to my self that dillo was heading in another direction. It seems to me that there's no intention to receive help from others, or get into account other's contributions. I believed that the main force behind a project weren't its main developers but its users and contributors. That mix made projects grow and grow. And dillo rejected that "magical" force. I'm just writing these thoughts, because as i said, i'm worried about dillo. I can't even imagine how painful would be to surf the web without it. Please everyone who loves dillo, do something so this nice project can grow without restrictions. Please don't get me wrong i don't want to start a trial, i just want dillo to evolve and get better everyday, with everyone help. Regards, Andres

3 2

small politically incorrect html parser patch
by Madis Janson Jan. 27, 2004

Jan. 27, 2004

when you want to use dillo for browsing instead of html test tool, then this little patch is useful for many stupid webs, which randomly close or not close font tags whenever happens and whereever happens... diff -pru -x CVS -x configure -x 'aclocal*' -x 'Makefile*' -x config.h.in dillo-tab/src/html.c dillo/src/html.c --- dillo-tab/src/html.c 2004-01-25 04:37:21.000000000 +0200 +++ dillo/src/html.c 2004-01-27 04:27:18.000000000 +0200 @@ -4280,7 +4372,7 @@ static const TagInfo Tags[] = { {"dt", 'O', Html_tag_open_dt, Html_tag_close_par}, {"em", 'R', Html_tag_open_em, Html_tag_close_default}, /* fieldset */ - {"font", 'R', Html_tag_open_font, Html_tag_close_default}, + {"font", 'O', Html_tag_open_font, Html_tag_close_default}, {"form", 'R', Html_tag_open_form, Html_tag_close_form}, {"frame", 'F', Html_tag_open_frame, Html_tag_close_nop}, {"frameset", 'R', Html_tag_open_frameset, Html_tag_close_default}, It makes a bit less likely, that the font tags screw up every other one (like tables and forms)...

1 1

Dillo cookies
by pete Jan. 27, 2004

Jan. 27, 2004

I am getting this error with Dillo 0.8.0: [pete@happy:.dillo, 07:25 PM, Mon Jan 26]> dillo dillo_dns_init: Here we go! The cookies file has a file lock: disabling cookies in this dillo! Nav_open_url: Url=>about:splash< How do I go about 'unlocking' this file? It was working initially with 0.7.3 Thanks, -- Pete http://milneweb.com http://nomorevirus.com

1 0

An odd request, I know...
by Tom Barnes-Lawrence Jan. 26, 2004

Jan. 26, 2004

Hi, Firstly, I'd like to say I've been using Dillo for over 2 years now, and find it fantastic. Well, apart from a few little details (which I'll probably post to the list in a couple of days, when I'm less busy). So anyways, very well done for this piece of software! Now, my strange request: I'm currently writing a bunch of m4 macros to (statically) generate my personal website. Unfortunately, after ages of trying to come up with a decent layout that rendered properly in various browsers, I came up with pretty much nothing. But I kept seeing, every time I started Dillo, the tidy little splash screen... ...So basically, would Jorge (or whoever is responsible for that) mind if I effectively copied the HTML? I wouldn't bother to ask, as I'm just getting my macros to churn out the tags, rather than copying anybody's *content*, but I thought that might be covered by copyright anyway- in which case, I'm being extra cautious before sticking anything up on my site. Thanks in advance, Tom Barnes-Lawrence

4 6

my.yahoo.com
by pete Jan. 26, 2004

Jan. 26, 2004

Hi all, I am now running the CVS version of Dillo. I am seeming to have a lot of lockups with it. Now I know this is common with CVS but is this something that should be mentioned here? Just wondering what is the proper thing to do. Another, I am still trying to access my yahoo with it but still not possible. Am I missing something that I need to do or is this just typical for Yahoo? -- Pete http://milneweb.com http://nomorevirus.com

1 0

Thanks!
by Arjen Meek Jan. 26, 2004

Jan. 26, 2004

Just a quick "thank you" to all Dillo-devs here. It has replaced Firebird as my default browser less than 5 minutes after installing it. Some experiences: - Dillo combines perfectly with the Ion window manager. A browser shouldn't have to perform window manager tasks, and tabbed browsing becomes a rather silly concept when a sensible window manager combines with a sensible browser. - 0.7.3 compiles & runs cleanly on my 60 Mhz HP PA-RISC system (712/60 running Debian Woody-hppa), and actually renders pages faster on that system than Firebird does on my 600 Mhz Athlon. - A suggestion: why not have CTRL+L switch focus to the URL bar when it is visible, instead of popping up a dialog box? Keep up the good work! Arjen

1 0

Mailing list
by Jorge Arellano Cid Jan. 26, 2004

Jan. 26, 2004

Hi Thorben! Is good to see you around. You've surely noticed I recently switched dillo-dev to only accept posts from its members. This is because it was receiving too much SPAM; the kind that's crafted to bypass SPAM filters. I don't know the current filtering layout of our list, but the fact is that now I receive (as the list administrator) a confirmation email for each rejected message (?) coming from a non-member. :( If the list SPAM-filters everything before checking whether it comes from a member, it's a sane default I suppose, but considering that I have near 3 confirmation requests a day (that repeat themselves for each day I leave these duties unattended), I wonder if it's possible to simply discard them. Regards Jorge.-

1 0

Bug meter
by Jorge Arellano Cid Jan. 26, 2004

Jan. 26, 2004

Hi there! I just commited the new (big) patch that adds a "bug meter" to Dillo's UI. From the Changelog: + Added a bug-meter button that shows the count of detected HTML errors (left click shows the errors, right click offers a validation menu). + Removed the "show_html_warnings" option from dillorc. + Cleaned up the code for the progressbar widgets. Most probably this CVS, plus a few improvements to our configure stuff (compiling in old distros, png detection and proper library linking) will be the basis for dillo-0.8.0-rc1. Please test it and enjoy! Cheers Jorge.- PS: I'm for letting "combat mode" rendering and https for 0.8.1 as its been too much time without a release.

3 2