WebMagick script generates html pages, which are not valid html. They lack DOCTYPE declaration, include non-standard elements, some of the tags are in wrong elements etc. The proposed patch (which can be just placed in files directory of the port) makes the generated html-files be valid html 4.01. Changes in brief. - add DOCTYPE declaration at the beginning of each generated html-file - change default charset to iso-8859-1 (from default not use charset at all) - do not use non-standard <NOBR> element - change case of case sensitive tags (border, frameborder) - move tags border, frameborder from <FRAMESET> to <FRAME> - delete nonstandard FRAMESPACING tag - add ALT tags where missed - add TYPE tag to <SCRYPT> elements - avoid empty <HEAD> element by use of <META> and <TITLE> in it - change <HR> to <BR> in one place; I did not check why the validator dislikes <HR> here Please note that FRAMESPACING tag is introduced by files/patch-aa Please also note complains from portlint -AC /usr/ports/www/WebMagick How-To-Repeat: Feed an URL of a WebMagick-generated page to http://validator.w3.org/ e.g. http://validator.w3.org/check?uri=http://webmagick.sourceforge.net/samples/one/ (this is a sample from WebMagick homepage).
Responsible Changed From-To: freebsd-ports-bugs->ache Over to maintainer.
State Changed From-To: open->feedback > WebMagick script generates html pages, which are not valid html. The main goal here is to produce most compatible html, i.e. rendered identically by most of browsers, and not pure standard compliance. I not see how your fixes can help identical rendering, but I see that it broke it in several places. Some changes are pure nosense, like changing tags case, some does nothing to rendering. If you'll find browser that render generated pages in wrong way and will find a fix for it which not broke other browsers in the same time (at least Mozilla, IE, old Netscape, Opera should be tested), you are welcome.
Since my reply to the maintainer did not make its way to gnats, I repeat it with some clarifications. Bottom line: Close the PR. Briefly: We have different goals, therefore different opinions and demands to tools. Without standard complience as a goal the patch is not needed. Broader. I never consider identical rendering on different platforms as a goal for html (generated or handwritten). For this purpose I use jpeg, png and similar formats. Even pdf is a compromise from identical rendering to other conviniences of use. For html I see logical structure of rendering as a goal. After validator.w3.org I planned to try Bobby validator, which promises reasonable apearence on all major browsers for complient pages. That needs syntactic correctness before checking the accessibility recomendations. I have some experience in attempts to make things more standard, it is nearly impossible, and not only in FreeBSD. I give up. (Yes, I know, some brave souls manage to do progress in this field - Thanks, guys! There are people appreciating your efforts!) My PRs standards/46504, www/63551, and www/63552 can be closed too. I cannot return to them within a year in any case. Best regards, -- V.Chukharev Reply to maintainer follows. On Mon, 24 May 2004 03:21:59 -0700 (PDT), Andrey A. Chernov <ache@FreeBSD.org> wrote: > Synopsis: [Patch] WebMagick generates nonstandard html > The main goal here is to produce most compatible html, i.e. rendered > identically by most of browsers, and not pure standard compliance. I not > see how your fixes can help identical rendering, but I see that it broke I thought a secondary goal exists, being standards complient. Mea culpa, it does not. > it in several places. Some changes are pure nosense, like changing tags > case, some does nothing to rendering. If you'll find browser that render > generated pages in wrong way and will find a fix for it which not broke > other browsers in the same time (at least Mozilla, IE, old Netscape, > Opera should be tested), you are welcome. I tested only with Mozilla, IE, Opera and w3m. Didn't see any change in rendering for one browser (rendering by IE and w3m are not identical). Little speed up does not count - suspect myself being inclined. Sure, testing was insufficient. You do not show what it broke, I cannot comment on that. Close the PR. > http://www.freebsd.org/cgi/query-pr.cgi?pr=67070 Best regards, -- V.Chukharev
State Changed From-To: feedback->closed Some of your fixes integrated with just committed port update