NetSurf 3.11 Released
NetSurf 3.11 is now available to download from https://www.netsurf-browser.org/ The biggest change to the core is improved CSS support, including support for the “display: flex” property value, which improves the layout of some web pages. There is also a new option in the Choices to disable CSS. A huge number of other changes have also been made including new features, significant performance optimisations, stability improvements, and other fixes. Change Log == NetSurf 3.11 ### Core / All platforms * HTML: Fixed failure to reflow SVGs if fetched and ready before layout. * HTML: Support for `display: flex` layout. * HTML: Improved minimum/maximum sizes in box layout. * HTML: Improved handling of percentages. * HTML: Minor cleaning up of layout code. * HTML: Use new LibCSS API for unit conversion. * HTML: Improved ordered list handling. * CSS: Updated selection callbacks to latest LibDOM API. * Options: Added core option to disable CSS. * Options: Added core option to prefer dark mode. * Options: Improved user choices file processing. * Bitmap: Opaque testing now implemented in core. * Bitmap: Added core support for premultiplied-alpha. * Bitmap: Format logged on startup. * Bitmap: Added support for pixel colour component order configuration. * Bitmap: Added colour component order conversion functions. * Bitmap: Generally more optimal on all platforms for all image formats. * Image: JPEGXL image handler. * Image: Compatibility with latest rSVG version. * Image: Improved handling for broken GIF images. * Image: Updated to new LibNSGIF API. * Image: Updated all image format handlers to use new core bitmap capabilities. * Fetch: Updated to new libcurl API. * Fetch: Disabled TLS1.0 and TLS1.1. * Fetch: Improved handling of bad SSL connections. * Fetch: Change to libcurl to optimise HTTPS connections (upstreamed). * Local history: More robust rendering. * Resources: Updated certificate bundle. * JavaScript: Minor updates to DOM bindings. * JavaScript: Updated to Duktape 2.7.0 release. * JavaScript: Console: Don't log through closed window. * Utility: Cleaned up UTF8 handling. * Utility: Improved recursive directory removal. * Utility: Add support for xx_YY format language codes. * CI: Various improvements to build automation and testing. * General: Various warning fixes. * General: Aligned UserAgent with compatibility spec. * Documentation: Updated URLs to https. * Documentation: Added front-end development guide. * Text areas: Clear selection on word left/right. * Buildsystem: Fixed handling of removed header files. * Disc cache: Minor fixes. * Debug: Added generated charts to image cache stats page. * Debug: Added descendant bounding boxes to HTML box tree dumps. * Built in: Cleaned up generated `about:` pages. * LibParserUtils library 0.2.5 (parser building utility functions): - Optimised consuming from buffer. - Optimised endian detection. - Added new API to append vector to buffer. * Hubbub library 0.3.8 (HTML parser): - Massively optimised element type detection using perfect hash. - Optimised and updated performance tester. - Fixed bitrot in tests. - Improved example client code. - Buildsystem improvements. * LibCSS library 0.9.2 (CSS parser and selection engine): - Added support for SVG `fill-opacity` property. - Added support for SVG `stroke-opacity` property. - Added support for CSS property wide `revert` value. - Added support for CSS property wide `unset` value. - Added support for CSS property wide `initial` value. - Added support for CSS `position` property `sticky` value. - Added support for CSS `display` property "grid" values. - Added support for `prefers-color-scheme` media query. - Added new public API for CSS unit conversion. - Added support for predefined counter styles. - Optimised media query handling. - Made selection code generator deterministic. - Various selection code generator improvements. - Squashed leak of system font names. - Improved internal handling of property units. - Improved internal string map. - Minor buildsystem improvements. - Improved example code. - Added new tests. * LibDOM library 0.4.2 (Document Object Model): - Fixed XML parser error handling. - Fixed XML parser empty document handling. - Added DOMTokenList implementation. - Added DOM tree walking function. - Improved example code. - Fixed HTML Element int32 attribute getter to handle signed values. - Various stability improvements. - Buildsystem improvements. * LibNSGIF library 1.0.0 (GIF support): - Complete rewrite. - New API that doesn't expose internal state. - Much better handling of bad or broken GIFs. - Support for decoding to client's choice of pixel colour component order. - Many fixes. - Faster decoding. - Updated documentation. * LibSVGTiny library 0.1.8 (SVG support): - Fixed X11 example utility build. - Implemented path arc correctly. - Updated documentation. ### RISC
Please test the latest build
Hello, Please could you test the latest builds (5377 or later) from: https://ci.netsurf-browser.org/builds/ Particularly please test the Amiga OS4 and RISC OS builds as we have updated the versions of 3rd party libraries that we build against. Amiga OS4 builds have had about 3 years worth of updates and RISC OS builds about 1.5 years. In addition we have patched the version of libcurl we build against which should make HTTPS connections much faster for the RISC OS and Amiga OS4 builds. Best regards, -- Michael Drake https://www.codethink.co.uk/ ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: NetSurf wont start with certain (required) fonts
On 11/10/2022 02:53, Mike Hobbs wrote: > I just upgraded MessengerPro (R-Comp) and to be able to display > unicode characters certain fonts are required (e.g. Libertine, > DeJaVu). However, when these fonts are installed Netsurf fails to > start up. > > Is there a workaround for this? What version of NetSurf are you using? Please try a recent development build (e.g. 5367): https://ci.netsurf-browser.org/builds/riscos/?C=M;O=D -- Michael Drake https://www.codethink.co.uk/ ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: Using Netsurf cookies with wget
On 21/01/2022 13:17, Harriet Bazley wrote: > Then the browser history getting updated with the new page and a > FETCH_REDIRECT from the login page to the user home page. No record of > what data was sent to the server, that I can see. In your Choices file, try setting: suppress_curl_debug:0 -- Michael Drake https://www.codethink.co.uk/ ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: Build 5293 on Atari
On Sun, 23 May 2021, 13:10 Peter Slegg, wrote: > > > > > Date: Sat, 22 May 2021 12:35:02 + (GMT) > > From: Peter Slegg > > Subject: Build 5293 on Atari > > To: > > Message-ID: <0013ad5a.04251490b...@smtp.freeola.net> > > > > I just updated to 5293 and while it doens't crash it never displays > > the local page that I had opened. > > > > > > > > Peter > > > Does it work with the latest build? > ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: Crashing after #5252
Hi John, On 24/02/2021 09:33, Petit Four wrote: > On 24 Feb, petit.f...@free.fr wrote: > >> For example, #5255 crashes with the attached error log. Please try with #5261 or later. Best regards, -- Michael Drake ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: Dragging from URL icon
On 02/12/2020 22:24, Harriet Bazley wrote: > On 2 Dec 2020 as I do recall, > Martin Avison wrote: > >> In article <725eadd858.harr...@bazleyfamily.co.uk>, >>Harriet Bazley wrote: >>> I'm on RISC OS 5.27. >>> Neither dragging the little yellow star nor dragging the URL text is >>> currently working. > > Also, if the star is the thing you're supposed to drag ('favicon'), > isn't there a problem with accidentally creating a favourite in the > process? Having a draggable object distinct from the icon text sounds > like a good solution in principle, but using that particular one seems > awkward. The star is a hotlist/bookmarks/favorites toggle. The favicon is either a website-specific icon, or if none is provided, the NetSurf content icon should be used as a fallback: http://www.netsurf-browser.org/documentation/images/content.png -- Michael Drake ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: Dragging from URL icon
This is all getting very confusing. For what it's worth, this is the target behaviour that NetSurf's aiming for, note it varies depending on whether the version of RISC OS supports text selection / C in writable icons: RISC OS with text selection support === RISC OS Versions: * RISC OS 4/6 from 4.37 onward (I think, it's been a while.) * RISC OS 5 from 5.28 onward Dragging the URL text should perform text selection, and you should be able to cut/copy/paste the text with Ctrl+X/C/V. Dragging the favicon should drag-save the whole URL. RISC OS without text selection support == RISC OS Versions: * RISC OS 2/3/4 up to 4.37 * RISC OS 5 up to 5.28. Dragging the favicon or the actual URL text should drag-save the whole URL. If that's not happening, it's not as intended, however, we'll need a RISC OS developer to fix it for us. Cheers, -- Michael Drake ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: Dragging from URL icon
Hi, On 25/11/2020 16:51, Harriet Bazley wrote: > I upgraded to v5223 and have just noticed that dragging a URL from the > URL icon no longer produces any result - you have to go via Page->Save > location->Text, or try to select, copy and paste. > Is this an intentional change? Not completely. Ideally if you're on a version of the OS that supports text selection in writable icons, a URL drag should select text. It needs a RISC OS developer to make any sense of it, however. Contributions welcome! Cheers, -- Michael Drake ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
NetSurf 3.10 released
NetSurf 3.10 has been released. It is available from: https://www.netsurf-browser.org/ NetSurf 3.10 features many fixes, better performance and improved functionality. In particular, handling of authentication, certificates and fetch errors have been updated and unified. The GTK front end has had a major overhaul too. -- NetSurf Developers ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: HTML5 element support
On 14/04/2020 17:18, Ewen Pring wrote: > At least some of my problem can be seen at these 2 simplified test pages: > > https://timebus.co.uk/riscos/webtest/using-html4.htm > https://timebus.co.uk/riscos/webtest/using-html5.htm Those are great, thanks! It looks like a problem with our Hubbub library. (The HTML5 spec has moved on a long way since it was written.) > If Netsurf is supposed to support HTML5 elements perhaps I should log this > as a bug on the tracker. Please do! Cheers, -- Michael Drake https://www.codethink.co.uk/
Re: HTML5 element support
On 14/04/2020 13:39, Ewen Pring wrote: > Specifically, my web site has over 60 pages that use the HTML5 figure and > figcaption elements to contain content (images and their captions > alongside). Netsurf doesn't recognise these and doesn't act upon styles > applied to them in CSS. The page renders as if they were not there. They work here. What version of NetSurf are you running? -- Michael Drake https://www.codethink.co.uk/
Re: Please test webp image format
On 28/12/2019 23:48, Chris Young wrote: > Not working on AmigaOS 4/PPC, just fails to recognise them at all (once > I'd remembered to disable my old WebP datatype). I think the Amiga (and Atari) toolchain builds need to have libwebp enabled in the SDK. -- Michael Drake https://www.codethink.co.uk/
Please test webp image format
Hello, Recently we added the webp library to our SDK. NetSurf Builds from our CI should now have webp support. So far we've only tested on Linux. Please could users of other platforms visit https://developers.google.com/speed/webp/gallery1 and let us know if it's working? The CI builds are available from here: https://ci.netsurf-browser.org/builds/ Cheers, Michael
Local history update
Hi all, The local history window has had a few updates. * It will now open scrolled to show the current entry. * It can now be navigated with the keyboard. (Cursor Left, Right, Up, Down, and Return.) * Thumbnails are now slightly larger, and their dimensions take account of the screen DPI. Cheers, -- Michael Drake https://www.codethink.co.uk/
Re: The 3.9 Release / documentation for Media Queries etc
On 22/07/2019 11:00, Jim Nagel wrote: > As I understand it from the bottom of my learning curve, Media Queries has > to do with designing a website so that it automatically adapts according > to whether the user is viewing it on a large or small screen. Yes, exactly. It's a way to make conditionalise the application of groups of CSS rules. So you can give elements different CSS properties depending on, for example, the width of the window. Here's a trivial example. It will show "large", "medium", or "small" depending on the width of the page. http://test.netsurf-browser.org/html/mq.test.html Note, in NetSurf we don't support dynamic changes to the computed style yet, so as you change the width of the window, you'll need to click reload to see the effect. Cheers, -- Michael Drake https://www.codethink.co.uk/
Re: NetSurf 0002562 - not quite fixed
On 30/08/18 20:21, David Higton wrote: > In message > David Pitt wrote: > >> David Higton, on 29 Aug, wrote: >> >>> but >>> if you delete an entry of the hotlist when all of the list is displayed >>> (i.e. there is empty space at the bottom of the window), the bottom line >>> is repeated. >> >> Also trying to delete a single entry is a bit odd. > > I discovered, by fluke, another way to get an instant crash. Open > the hotlist, open an ordinary window in front of it. Drag the mouse > from the hotlist into the ordinary window. Bang. We need a RISC OS maintainer to step up to work on these sorts of things. They simply aren't reproducible on the GTK front end that the core developers develop on. Cheers, -- Michael Drake https://www.codethink.co.uk/
Re: Please try the latest build
On 15/08/18 21:33, Peter Young wrote: > Yes, I got one of those too. Why ask for a reply to a closed group to a > message to the general public? I simply sent the e-mail to two mailing lists, so that it would be seen by people who are only on one of them. Both are public, but you might not be subscribed to both. Cheers, -- Michael Drake https://www.codethink.co.uk/
Re: new release and nested
On 15/08/18 12:55, Jim Nagel wrote: > Here's an example: http://stjohns-glastonbury.org.uk > At first I groaned and presumed it's the usual Javascript impasse. But > looking at its HTML code, you can see that its navigation system uses > (unordered list) for the three menu items that Netsurf shows, but > WITHIN each of these list items is nested a second set that > Netsurf does not display. > > Any chance that the new Netsurf release could add this trick? I've no > idea how difficult it would be. (But at least it ain't Javascript!) Our current layout engine doesn't support dynamic changes to the page. The layout engine needs a rewrite to handle this, which is a big job. It won't happen for the release. Cheers, -- Michael Drake https://www.codethink.co.uk/
Re: Please try the latest build
On 15/08/18 12:08, Jim Nagel wrote: > Michael Drake wrote on 15 Aug: >> Please try entering the incorrect credentials, the correct >> credentials, and canceling the login, by closing the login window. > > Problem arises in testing this: After a successful login, Netsurf > will return to the site without asking again for credentials. > > How do we make it forget so that we can try again to see what happens > with bogus credentials? With https://httpbin.org/basic-auth/foo/bar you can adjust the "correct" credentials by altering the URL. For example: https://httpbin.org/basic-auth/elephant/potato Cheers, -- Michael Drake https://www.codethink.co.uk/
Re: Please try the latest build
Hi Brian, On 15/08/18 11:57, Brian wrote: > In article , >Michael Drake wrote: >> Which platform are you testing on? > > Win 7, VRPC, RISC OS 4.02, NetSurf #4402 Just spotted this! It needs to be #4403 or later. Please could you try with #4403 or the latest one (currently #4405). Cheers, -- Michael Drake https://www.codethink.co.uk/
Re: Please try the latest build
Hi Brian, On 15/08/18 11:57, Brian wrote: > In article , >Michael Drake wrote: >> Can you send me the log file for the crash. > > I've saved the log file but forgotten where to send it Please e-mail it to me direct. Cheers, -- Michael Drake
Re: Please try the latest build
On 15/08/18 09:05, Brian wrote: > In article <3c705f3f-e461-c7a7-3b70-54d63192a...@netsurf-browser.org>, >Michael Drake wrote: >>https://httpbin.org/basic-auth/foo/bar > >>(Where "foo" is the username, and "bar" is the password, >>and you can change the "correct" username and password >>by changing the URL.) > > Seems not to work. Serious error - crashes NetSurf. Which platform are you testing on? What exactly did you do? Can you send me the log file for the crash. Cheers, -- Michael Drake
Please try the latest build
Hello, Could people please test build #4403 or later? (Or build from latest master, if you compile from source.) https://ci.netsurf-browser.org/builds/ We are close to making a release, and would like to check that nothing is broken. I am particularly interested to hear from anyone who can run the following platforms: - Amiga - Atari - Haiku - RISC OS - Windows We have just changed the way that HTTP authentication works, so I would be grateful if people could check that anything that requires an HTTP basic auth login still works with the latest build. If you don't know any sites which use HTTP authentication, you can use: https://httpbin.org/basic-auth/foo/bar (Where "foo" is the username, and "bar" is the password, and you can change the "correct" username and password by changing the URL.) However, testing with real-world sites is better. Please try entering the incorrect credentials, the correct credentials, and canceling the login, by closing the login window. Note: the Windows build doesn't yet support HTTP auth, so for that platform, I'm just looking for a general indication of whether the browser functions. Cheers, -- Michael Drake http://www.netsurf-browser.org/
Re: why is site php.net so slow in NetSurf?
On 16/07/18 23:18, Jim Nagel wrote: > Trying to learn stuff from http://php.net is so painful on NetSurf > because any page from the site takes ages to load and a further age to > scroll. (Using #4342) > > I'm curious about the mechanisms involved. Any wisdom, please? It's because of this SVG being used as a tiled background on the HTML element: http://php.net/images/bg-texture-00.svg Since its an SVG it bypasses all the optimisations we have for tiled bitmap redraw. It also isn't a filled rectangle, its just strokes, so it doesn't get "knocked out" by our normal rendering optimisations. Finally it uses opacity, which wouldn't help either, although I think we ignore that anyway for now. It could potentially be optimised by pre-rendering it once to a bitmap, but it would be a fair amount of effort for what is a pretty uncommon case. For now, I'd try just toggling background images off in the menu for that window, and forcing a redraw (F12, Return). http://www.netsurf-browser.org/documentation/guide.html#BrowserWindowMenuDisplayImages Hope that helps, -- Michael Drake https://www.codethink.co.uk/
Re: Image won't display
On 12/06/18 14:07, Richard Torrens (lists) wrote: > A very weird problem... > > The page > http://stronged.torrens.org/man/modes/syntaxmodefile.html > which I am writing. It has a line > > But Netsurf won't display the image. Chrome does it properly. It's getting caught by NetSurf's ad blocking CSS. Looks like it matches the following selector. img[src*="/click"] You can disable ad blocking in the preferences. -- Michael Drake https://www.codethink.co.uk/
Re: Persistant error boxes.
On 13/05/18 18:08, lists wrote: > I get persistant error boxes from NetSurf "failing to identify the > authenticity of an SSL certificate". Whether I select reject or accept, I > get a continued succession of similar boxes and can end up with dozens all > over the screen. Do you have a URL that causes it? Which version of NetSurf? Have you tried the latest build? http://ci.netsurf-browser.org/builds/ Cheers, -- Michael Drake http://www.codethink.co.uk/
Re: URL crashes Netsurf
On 13/04/18 14:02, Richard Torrens (lists) wrote: > www.siemens-home.bsh-group.com/uk > > causes NetSurf to totally crash on the ARMX6 with 5.23 (18-Feb-18). Works for me with the GTK front end on Linux. What version of NetSurf are you using, and have you tried the latest development build? Cheers, -- Michael Drake http://www.codethink.co.uk/
Re: linking to root of website, not root of drive
On 19/02/18 15:46, Jim Nagel wrote: > - NetSurf displays the pages served up via WebJames perfectly, > except that it is NOT obeying the stylesheet I guess WebJames is not serving the CSS file with the "text/css" MIME type. I think there's a mimemap file that lives somewhere in !Boot, which may need updated with an entry for CSS. WebJames may ship with an example mimemap for this. Cheers, -- Michael Drake http://www.codethink.co.uk/
Re: linking to root of website, not root of drive
On 19/02/18 10:35, Jim Nagel wrote: [snip] Is there a way around this? Use a webserver to serve the root directory of the web site you're developing, and point your browser at http://localhost/ I believe WebJames can do this on RISC OS. Cheers, -- Michael Drake http://www.codethink.co.uk/
Re: CSS support for display:table
On 15/02/18 16:00, Jim Nagel wrote: That seems to mean Mozilla and Webkit. Is there perhaps an equivalent for NetSurf? No, we have no plans to implement draft specifications under a vendor prefix. We don't have enough developer bandwidth to implement all the stuff that is currently standard, never mind trying to keep up with volatile draft standards. Cheers, -- Michael Drake http://www.codethink.co.uk/
Re: CSS support for display:table
On 15/02/18 12:17, Jim Nagel wrote: I'm wondering if Netsurf CSS supports these commands (if that's the correct term) display:table or width:min-content Supported: display:table Unsupported: width:min-content The min-content value for width is just at editors draft stage. https://drafts.csswg.org/css-sizing-3/ So it's subject to change, and not standardized yet. As the mozilla documentation says, it shouldn't be used in production code. The third idea uses table:caption; caption-side:bottom; -- again, does Netsurf support it? In my attempts the caption lands at the right of the picture, not beneath it. (See "tablecaption-X.html".) NetSurf doesn't support table captions. As for how I'd do it, I think there would be more mileage in constraining the outer figure element with image dimensions, rather than the image. Something like (untested, and requiring polish): figure { display: block; width: 33%; max-width: 400px; min-width: 100px; } figure.left { float: left; } figure.right { float: right; } figure > img { display: block; width: 100%; } figure > figcaption { display: block; width: 100%; } Then you'd have e.g. Lots of text... -- Michael Drake http://www.codethink.co.uk/
Re: CSS for a RiscOS font and button
On 09/02/18 14:56, Jim Nagel wrote: In CSS, is it possible to specify a particular RiscOS font? No, NetSurf only supports the generic font families. In CSS, how would I define a button to look like a familiar RiscOS action button? Input type image is probably your best bet, but it won't match desktop button style if users have customised the RISC OS button rendering in any way. -- Michael Drake http://www.codethink.co.uk/
Re: Failure to initialise Unicode font library
On 07/12/17 11:12, John Williams wrote: My RPi has suddenly developed an inability to load NetSurf: The Unicode font library could not be initialized. Please report this to the developers. [snip] Any suggestions as to what to look at next? I think the usual cause is the installation of a subtly broken font. Have you installed any fonts recently? -- Michael Drake http://www.codethink.co.uk/
Re: Experimental treeview search feature
On 03/10/17 16:09, Bernard Boase wrote: By the way, 'Global history' doesn't appear to include the list given by the main window's drop-down 'URL suggestion icon'. That list is stored in file Choices.WWW.NetSurf.URL. I'm unclear which URLs are placed there, but I do see that it is unaffected by deleting the items from Global history. That's something separate from the core global history that is managed by the RISC OS front end. It stores the URLs that are typed manually into the URL bar, I think. Cheers, -- Michael Drake http://www.codethink.co.uk/
Re: LibCSS: Flexbox property support review
On 30/09/17 21:05, Tim Hill wrote: In article <0a012258-3662-78c0-93ca-5dac0982b...@codethink.co.uk>, Michael Drake <t...@netsurf-browser.org> wrote: This is great news even though I suspect you didn't mean to post it to 'users' :). Yes, I meant to send it to the dev list, sorry. This work adds handling of flexbox properties to LibCSS, but there has currently been no work to implement the Flexbox layout algorithms in NetSurf. We only change version numbers when we do a release. Cheers, -- Michael Drake http://www.codethink.co.uk/
LibCSS: Flexbox property support review
(min_height == CSS_MIN_HEIGHT_AUTO) { uint8_t display = get_display(style); if (display != CSS_DISPLAY_FLEX && display != CSS_DISPLAY_INLINE_FLEX) { min_height = CSS_MIN_HEIGHT_SET; } } return min_height; } Finally, I don't think it's OK to override the value to *_SET and not set appropriate values for length and unit at the same time. When the computed value is set to *_AUTO in the cascade it doesn't clear any values for length and unit. So we should be them up for 0px where we override min_height. Same for both css_computed_min_height() and for css_computed_min_width(). In css_computed_display(): +} else if (display == CSS_DISPLAY_INLINE_FLEX) { + return CSS_DISPLAY_FLEX; Spaces used for indent on the first inserted line. In src/select/computed.h: http://source.netsurf-browser.org/libcss.git/diff/src/select/computed.h?h=lcneves/flexbox-tidy=2fd135cd43d6821e944110bbdf8f77bea5cc40fb + * 35 yyybbbaa overflow-y | background-repeat | align-content + * 36 bbaj flex-basis | align-content | justify_content + * 37 fffcccjj flex-direction | clear | justify_content Since align-content and justify-content are split over two areas, I would say somthing to hint at that in the comments, e.g.: * 35 yyybbbaa overflow-y | background-repeat | align-content1 * 36 bbaj flex-basis | align-content2| justify_content1 * 37 fffcccjj flex-direction | clear | justify_content2 In the long term perhaps we could consider using a uint32_t array for the bit data than uint8_t, so its easier to pack things in without splitting stuff. +css_fixed flex_basis; + +css_fixed flex_grow; + +css_fixed flex_shrink; + +int32_t order; Spaces used for indent. In src/select/dispatch.c: http://source.netsurf-browser.org/libcss.git/diff/src/select/dispatch.c?h=lcneves/flexbox-tidy=2fd135cd43d6821e944110bbdf8f77bea5cc40fb + }, +{ + PROPERTY_FUNCS(align_content), Spaces for indent on the middle inserted line above. In src/select/propget.h http://source.netsurf-browser.org/libcss.git/diff/src/select/propget.h?h=lcneves/flexbox-tidy=2fd135cd43d6821e944110bbdf8f77bea5cc40fb -#define BOX_SIZING_INDEX 34 +#define BOX_SIZING_INDEX 23 Good spot! Ditto for the similar change in propset.h. In get_align_content(): +/* Most significant bit out of three */ + bits_b <<= 2; + +uint8_t bits = bits_a | bits_b; Spaces for indent on the first and last lines above. Similar in get_justify_content(). Selection: Logic for the flexbox properties === http://source.netsurf-browser.org/libcss.git/commit/?h=lcneves/flexbox-tidy=587926f812375472962628845226d18f5df95cf1 In src/select/properties/flex_basis.c: http://source.netsurf-browser.org/libcss.git/diff/src/select/properties/flex_basis.c?h=lcneves/flexbox-tidy=587926f812375472962628845226d18f5df95cf1 In css__cascade_flex_basis(): + uint16_t value = CSS_FLEX_BASIS_INHERIT; + css_fixed length = 0; +uint32_t unit = UNIT_PX; Spaces for indent on last line above. In src/select/properties/order.c: http://source.netsurf-browser.org/libcss.git/diff/src/select/properties/order.c?h=lcneves/flexbox-tidy=587926f812375472962628845226d18f5df95cf1 + /* Undo the <<10, because this is an integer */ + order = *((css_fixed *) style->bytecode) >> 10; Use CSS_RADIX_POINT instead of 10. (From include/libcss/fpmath.h) Or use the FIXTOINT macro from include/libcss/fpmath.h. -- Michael Drake http://www.codethink.co.uk/
Re: Experimental treeview search feature - a problem
On 25/09/17 14:19, Richard Torrens (lists) wrote: I have just spotted a (hopefully small) problem - it's impossible to delete items! Is #4213 any better? Cheers, -- Michael Drake http://www.codethink.co.uk/
Re: Experimental treeview search feature
On 22/09/17 02:20, Harriet Bazley wrote: Other thoughts: from a practical point of view, I think it would be preferable if any active search were deleted when the window is closed. Otherwise it gets a bit disconcerting when you go to reopen your history/hotlist and there are only two or three entries visible, due to a search you carried out earlier and have long since finished with - particularly as the icon in question doesn't have the input focus when the window reopens, I agree, it would be better to cancel the search when the window is closed. Unfortunately the core doesn't know when the front end closes the window at the moment, so it's not straightforward to change that just yet. and clearing the old search string is a pain! Once the search field has focus, either ESC or CTRL+U should clear it. Cheers, -- Michael Drake http://www.codethink.co.uk/
Re: Experimental treeview search feature
On 22/09/17 01:11, Harriet Bazley wrote: The size of the work area gets a bit confused if you toggle the 'expand entries' icon with a search active (seems to reset to what it would be if the whole window were being displayed, rather than just the visible entries). It corrects itself if you then carry out a fresh search. Good spot, that should be fixed in the latest build. Cheers, -- Michael Drake http://www.codethink.co.uk/
Experimental treeview search feature
Hello, An experimental search feature has been added to the following: 1. Hotlist viewer (bookmarks) 2. Global history viewer 3. Cookie manager window To use it, click in the area next to the search icon at the top of the window, and start typing. It has currently only been tried under the GTK front end. It would be helpful if people could put it thoroughly through its paces. Any feedback gratefully received. Cheers, -- Michael Drake http://www.netsurf-browser.org/
Re: Redraw oddity
On 16/05/17 13:23, Richard Torrens (lists) wrote: https://www.quora.com/Where-does-iPhoto-store-photos-locally At first this displays a blank page. But wipe over it with another window... Please try with CI build #4098 or later. Cheers, -- Michael Drake http://www.codethink.co.uk/
Re: Codethink sponsored work on NetSurf
On 05/05/17 13:46, Kevin Wells wrote: In message <20170505074516.gf12...@kyllikki.org> Vincent Sanders <vi...@netsurf-browser.org> wrote: On Wed, May 03, 2017 at 12:51:54PM +0100, Kevin Wells wrote: It has fixed a bug I submitted. :) I tried to inform the bug tracker, but it appears to be down. Thanks for reporting this, It was due to a configuration error after last weeks upgrades, fixed now I hope. Please let me know if there are any additional issues with the bug tracker. This bug has been fixed, I could not find a button on there to update the status: <http://bugs.netsurf-browser.org/mantis/view.php?id=2393> I've closed it. Thanks for the feedback! Cheers, -- Michael Drake http://www.codethink.co.uk/
Codethink sponsored work on NetSurf
Over the past few weeks I've been able to spend time contributing to open-source projects at work. Codethink [1] values open-source software, and so I was approved to work on NetSurf and related components while I was between projects. I'd like to thank Codethink for this opportunity, which I very much value and appreciate. What follows is a summary of the development I've been able to accomplish during this time. [1] http://www.codethink.co.uk/ LibCSS (CSS parsing and selection library) == * Added parsing and selection support for the CSS3 box-sizing property. NetSurf === Core: HTML layout engine * Added support for the CSS3 box-sizing property. This helps us assign element widths and heights as the web page author intended. * Extended :after pseudo element handling to cover the table display type. This improves layout of pages that use a variant of the clearfix CSS trick for clearing floats. * Fixed the handling of the CSS min-width and max-width properties to correctly affect the auto-width left and right margins. * Added support for width attribute on HR element. * Fixed use of uninitialised value in CSS presentational hint handling, that had potential to reduce the likelihood of computed style sharing. * Improved layout tree dump, to make debugging layout problems easier. Core: Content handling -- * Improved logging, making it easier to track HTML contents through a session, for debug purposes. * Removed bogus assertion in the HTML object fetch handler, which would sometimes crash the browser. * Fixed bug in content state machine, which made aborted fetches cause their contents to get stuck in the cache. This would happen on JavaScript location assignments. Core: URL handling -- * Improved file: URL handling. * Removed redundant code. * Improved hexadecimal handling. * Improved test suite. Core: Hotlist (bookmarks/favourites) * Added support for saving the hotlist automatically after entries are added or removed. This helps prevent loss of hotlist changes if NetSurf is unable to save the file on exit. * Added support for read-only hotlist mode, which can be selected by frontend code. In this mode, the hotlist file may be read, and the hotlist updated, but changes are never saved to disc. GTK frontend * Fixed bug where different pango layout contexts were used for measuring and painting text, which in some situations caused either lines with overlapping text, or large spaces between bits of text on a line. * Slight optimisation to the text line break handling code. * Rewrote bitmap rendering plotter. The new implementation is many times faster for tiled rendering, such as for repeating background images, and simplifies the handling of scaled plots. RISC OS frontend * Improved handling of external hotlist managers, by setting the core hotlist to read-only mode, when configured to use an external manager. NetSurf's JavaScript engine (Duktape) = NetSurf uses the excellent Duktape JavaScript engine. URL: http://www.duktape.org/ * Updated NetSurf and nsgenbind (our JavaScript binding generator) from Duktape 1.6 to Duktape 2.0. We had been stuck on the 1.X series due to API changes in Duktape 2, and lack of time to make the transition. Duktape 2.0 provides greater performance, amongst many other improvements. * Added support for HTML style comments (ES6 spec) to Duktape and upstreamed it to the Duktape author, Sami Vaarala. This feature had been requested by our users. * Updated NetSurf from Duktape 2.0 to the recently released Duktape 2.1, which gives further performance gains, adds the HTML stlye comment handling, and other improvements. * On the advice of the Duktape author, enabled Duktape's fastint build option, for faster integer arithmetic. LibNSGIF (GIF decoding library) === * Complete rewrite of LZW decompression. The new implementation is both 30% faster and cleaner code. It also now handles LZW codes that span three sub-blocks. (Although these are unlikely to actually appear in the real world.) LibSVGTiny (SVG decoding library) = * Build system improvements to fix warnings about use of gperf-generated code. Cheers, Michael
Re: history and hotlist not saved; CSS site
On 28/04/17 11:18, Michael Drake wrote: We could change it not to save the hotlist on exit when an external hotlist manager is used, since that would be a simple change. Done.
Re: history and hotlist not saved; CSS site
On 28/04/17 10:51, Michael Drake wrote: In any case, I can make it more robust by consulting the external hotlist option before removing from the hotlist. Actually it already does that, so NetSurf shouldn't be saving the hotlist when the hotlist is edited, when configured to use an external hotlist manager, however, it will still save the hotlist on exit. We could change it not to save the hotlist on exit when an external hotlist manager is used, since that would be a simple change. Adding support for querying the external hotlist about whether it has a URL and removing URLs from external requires more work, and would need a RISC OS developer. If there are any developers interested in working on the RISC OS front end, they would be most welcome. None of NetSurf's core developers use RISC OS, so we have trouble maintaining and supporting it. Cheers,
Re: history and hotlist not saved; CSS site
On 27/04/17 18:03, Michael Drake wrote: On 26/04/17 18:30, Nick Roberts wrote: Can you just confirm that NetSurf doesn't save the file if there is an external hotlist manager? I don't know, off the top of my head. I'm unfamiliar with the external hotlist option. If I have time tomorrow I could have a look at the code. As far as I can tell there are three differences in behavior when the RISC OS front end's external hotlist option is enabled: 1. Opening the hotlist will Filer_Run the external application, rather than opening the NetSurf hotlist manager window. 2. The hotlist_has_url test always returns false for external hotlists. (Meaning the star in the URL bar will always be an empty outline, rather than filled yellow, even when the current page is in the external hotlist.) 3. When adding pages to the hotlist, NetSurf's core hotlist code isn't informed. This means that the NetSurf won't save the file when adding URLs. So NetSurf won't save the hotlist file when adding pages, and set to use an external hotlist manager. It looks like the RISC OS front end code for removing pages from the hotlist first checks whether the core hotlist has the page, and if it does, it instructs the NetSurf core hotlist to remove it, rather than the external application. The remove from hotlist handling has no code to tell an external hotlist to remove pages. This means that if this code is called, and the the core hotlist had the URL, it will remove it, which will cause NetSurf to save the hotlist file. I can't remember exactly how the RISC OS UI presents the interface for removing pages from the hotlist. It might have been select clicking on a filled URL bar star, or adjust clicking on a URL bar star, or something else. If the former, then because of point 2 above, it would not let the remove page code get called in the first place, so the file wouldn't get saved. In any case, I can make it more robust by consulting the external hotlist option before removing from the hotlist. Cheers,
Re: history and hotlist not saved; CSS site
On 26/04/17 18:30, Nick Roberts wrote: Can you just confirm that NetSurf doesn't save the file if there is an external hotlist manager? I don't know, off the top of my head. I'm unfamiliar with the external hotlist option. If I have time tomorrow I could have a look at the code. Cheers,
Re: history and hotlist not saved; CSS site
On 23/11/16 09:58, Harriet Bazley wrote: On 23 Nov 2016 as I do recall, Jim Nagel wrote: Does NetSurf not save its hotlist and history till the user quits the program? No, it doesn't. As of CI build #4070, the hotlist file is saved shortly after URLs are added / removed. Cheers,
Re: Horizontal rules in NetSurf 3.5 & 3.6
On 08/04/17 21:41, pjpe...@freeuk.com wrote: This displays as expected in NetSurf 3.0 and Firefox but in NetSurf 3.5 and 3.6 it produces a full width rule. Should be fixed in CI build #4060. http://ci.netsurf-browser.org/builds/ Cheers, -- Michael Drake http://www.codethink.co.uk/
Re: Printing from Netsurf
On 03/04/17 22:52, Harriet Bazley wrote: Can anyone remind me again of the reason why Netsurf can't print to Postscript printers? (I've simply been working round it all the time by using Drawprint, but I'm trying to explain to my mother why it doesn't work...) Please see the Printing section here: http://www.netsurf-browser.org/documentation/roinfo.html#Printing Cheers, -- Michael Drake http://www.codethink.co.uk/
Re: Non-displaying page
On 05/01/17 13:27, Richard Torrens (lists) wrote: Is this a bug or a style fault on the page? The "body { visibility: hidden !important; }" style means "hide the whole page from view". There is some JavaScript to unset it, revealing the page. It's daft, because the page won't display at all in any browser with JavaScript disabled. In NetSurf, it doesn't work with JavaScript enabled because our JavaScript support is far from complete. -- Michael Drake http://www.codethink.co.uk/
Re: Comments and the JS engine
On 30/11/16 13:37, Brendan Stephenson wrote: Either the javascript engine is choking on
Re: No-display page
On 24/09/16 11:34, Richard Torrens (lists) wrote: http://www.petervis.com/Radios/sinclair-micromatic-pocket-radio/sinclair-micromatic-pocket-radio.html Does not display. Should be fixed in the next build (#3717). Cheers, -- Michael Drake http://www.netsurf-browser.org/
Re: Fail to load
On 23/08/16 09:22, Gerald Dodson wrote: Further to my posting just now, Organizer has been affected or is also involved. When I try to load it the error message "file does not start with orgdata" appears. It has been overwritten by NS. Sounds like it might be disc corruption. Have you tried DiscKnight? https://armclub.org.uk/products/discknight/ -- Michael Drake http://www.netsurf-browser.org/
Re: Slow ?
On 12/05/16 20:18, Peter Slegg wrote: Build 3242 required 5m 15s Build 3537 required 8m 28s 3242 was built without JavaScript for Atari. -- Michael Drake http://www.netsurf-browser.org/
Re: Slow ?
On 11/05/16 20:48, Peter Slegg wrote: I am going to work back through the versions to see if it is a real or imagined performance drop. If the performance drop came around the time JS builds were enabled for Atari, try running with JS configured off. -- Michael Drake http://www.codethink.co.uk/
Atari front end
Are there any Atari users around? Could you let us know if the current m5475 and m68k builds are working? http://ci.netsurf-browser.org/builds/atari/?C=M;O=D -- Michael Drake http://www.netsurf-browser.org/
Re: Javascript features list
On 19/02/16 20:11, Peter Slegg wrote: Is there a maintained list of available JS features ? The closest is this: http://git.netsurf-browser.org/netsurf.git/tree/Docs/UnimplementedJavascript.txt -- Michael Drake http://www.netsurf-browser.org/
Re: Review: Codethink sponsored work on NetSurf, LibCSS and LibDOM
On 12/02/16 17:46, Jim Nagel wrote: NetSurf (#3382) timed out when I tried to look at the Codethink site, but when I immediately tried again it displayed immediately. We made a change to the fetch behavior some time ago that resulted in us not waiting very long for a server to respond. We've moved back to the old behavior in build CI #3406. Cheers, -- Michael Drake http://www.codethink.co.uk/
Re: Backgrounds fill stopped working?
On 13/02/16 16:08, Bob Latham wrote: Is it me or has stopped working? Yep. Should be fixed in the next build, thanks. -- Michael Drake http://www.netsurf-browser.org/
Review: Codethink sponsored work on NetSurf, LibCSS and LibDOM
file changed, 199 insertions(+), 9 deletions(-) 2443cc4 Attempt to squash warning that happens on CI. 1 file changed, 3 insertions(+), 3 deletions(-) b1dbc04 Add paragraph proto test. 1 file changed, 23 insertions(+) 9fb755b Handle known HTML elements without specialisations. 1 file changed, 7 insertions(+), 1 deletion(-) 3cc80b6 Add test for document.write() adding a STYLE element. 1 file changed, 15 insertions(+) Next was a complete rewrite of CSS presentational hints handling. Presentational hints are mostly legacy HTML element attributes which affect the presentation of the elements. They affect the CSS computed style when no other styles have affected it. Using the new LibCSS API coupled with the new LibDOM element tag type feature has made support for presentational hints much faster. 2784514 Rewrite hints handling to be more efficient. 2 files changed, 788 insertions(+), 1123 deletions(-) eb67607 Merge branch 'tlsa/upfront-hints' I noticed and investigated an apparent memory leak. Some changes were made to improve the cache logging, which ruled the cache out as the cause of the problem. The leak was tracked to a particular issue in the GTK front end's text rendering. 63fb2cb Add URL to llcache object destruction logging. 1 file changed, 2 insertions(+), 1 deletion(-) fc2d766 Style: Wrap some long lines. 1 file changed, 31 insertions(+), 18 deletions(-) 255361a Log the cache limit along side size, after clean. 1 file changed, 1 insertion(+), 1 deletion(-) 9cb522c Don't haemorrhage Pango layouts out of nsfont_paint. 1 file changed, 1 insertion(+) A few tidy-ups to GTK font rendering and core CSS hint handling: df2d62a Use readonly get_line, since it is supposed to be faster. 1 file changed, 1 insertion(+), 1 deletion(-) f5be1eb Add switch case fallthrough comments. 1 file changed, 5 insertions(+) 5108424 Add switch case fallthrough comment. 1 file changed, 1 insertion(+) 3af77ea Can free pango font desc as soon as it's set. 1 file changed, 12 insertions(+), 10 deletions(-) Finally I fixed several bugs on the bug tracker. The first was an assertion (crash) that could occur during HTML layout. The issue stemmed from NetSurf's allowing :before and :after pseudo elements to be generated for replaced elements. This behaviour was unspecified in CSS 2.1, but in CSS 3 it is not permitted. Second was to remove some code which was complicating the status bar updating for HTML contents. Third was to fix the object accounting that is done to keep track of fetches used for additional resources such as images when rendering an HTML content. This accounting went wrong when a fetch for an HTML content was stopped, preventing the content ever reaching the DONE state, and causing misleading status bar messages. 3bee7b7 Layout: Don't generate :before and :after boxes for replaced elements. 2 files changed, 17 insertions(+), 6 deletions(-) d4a01d5 HTML: Remove some status bar updating code. 4 files changed, 43 deletions(-) 04e61b7 Fix object accounting for aborted HTML contents. 1 file changed, 4 insertions(+), 3 deletions(-) I also did quite a bit of testing, particularly of the JavaScript related code, to ensure that there is nothing major blocking the forthcoming NetSurf 3.4 release. Overall, there were 20 commits to NetSurf through the course of this stint. Overview and thanks --- I was able to get many hours of work done on the NetSurf project over this period. It has been a very nice experience, working on this project at work. Thanks to the work environment, I found myself more focused than I usually am when I'm working on it in my spare time. Many thanks to [Codethink](http://www.codethink.co.uk/) for providing this opportunity. I'm very grateful and I've really enjoyed it. The stats and logs presented were gathered using the following commands in the NetSurf, LibDOM and LibCSS repositories. $ git shortlog -n -s -e | grep "michael.dr...@codethink.co.uk" $ git log --author="Michael Drake <michael.dr...@codethink.co.uk>" --oneline --shortstat --reverse -- Michael Drake http://www.codethink.co.uk/
Re: Big push on testing needed
On 27/01/16 18:05, John Rickman Iyonix wrote: Daniel Silverstone wrote Have you reported this, along with an *attached* test case, to the BTS? If so, can you let me know the issue number? Michael Drake has reported it upstream to Duktape That doesn't change the need for a report in the NetSurf bug tracker. The upstream may decide that the fix doesn't belong in duktape, or they could implement it, but that would fix nothing for NetSurf users unless we update duktape in the NetSurf source. We have an issue tracker to keep track of all the things we need to do. It's impossible to remember everything. I've created an issue here: http://bugs.netsurf-browser.org/mantis/view.php?id=2416 I'd be grateful if someone could attach a test case. Cheers, -- Michael Drake http://www.netsurf-browser.org/
Re: Big push on testing needed
On 27/01/16 10:33, John Rickman Iyonix wrote: Assuming that that JavaScript is entirely defined by the ECMAScript standard, it seems clear that html style comments are valid. Thanks John. This has been reported upstream to the maintainer of Duktape (the JavaScript engine we use). You can follow the issue here: https://github.com/svaarala/duktape/pull/564 Cheers, -- Michael Drake http://www.codethink.co.uk/
Re: Latest builds missing.
On 26/01/16 07:49, David Pitt wrote: Currently the Changes page shows #3312 as the latest successful build but the Downloads page only goes up to #3307 for all builds. The downloads are still there, but "json" has been removed from the filenames. http://ci.netsurf-browser.org/builds/riscos/?C=M;O=D -- Michael Drake http://www.netsurf-browser.org/
Re: Big push on testing needed
On 25/01/16 21:43, Dave Higton wrote: One other thing: the CI builds are coming out with logging enabled. This will be changed for the release version. All RISC OS releases and CI builds have had logging enabled. I would not suggest turning it off. If something does go wrong there would be nothing to help us fix the problem In the very rare cases when we've logged so much that it makes a significant impact on performance it's because we're aware something is buggy and want to gather the info needed to help fix it. The reason I suggested turning logging off to you last night was to determine what the problem was without needing to rebuild anything. Since then, I've disabled the excessive JavaScript debug output, so there should be no reason for users to do anything unusual with the !Run file. Cheers, -- Michael Drake http://www.netsurf-browser.org/
Re: JavaScript enabled builds for Atari
On 23/01/16 23:06, Peter Slegg wrote: Using Netsurf JS today it has crashed a few times. It isn't as stable as the last non-JS versions. How about when you use it with "enable_javascript:0"? -- Michael Drake http://www.netsurf-browser.org/
JavaScript enabled builds for Atari
We have enabled JavaScript builds (json) for Atari on the CI server. You can get them here: http://ci.netsurf-browser.org/builds/atari/?C=M;O=D I'm not sure if the Atari front end exposes the JavaScript option in its GUI. If not, you can edit the Choices file setting either: enable_javascript:0 or enable_javascript:1 I'd be interested to get confirmation on whether its working from Atari users. You can test by visiting: http://git.netsurf-browser.org/netsurf.git/plain/test/js/js-fractal.html -- Michael Drake http://www.netsurf-browser.org/
Re: El Reg formats differently these days - why?
On 21/08/15 19:36, Dave Higton wrote: NS used to render The Register in three columns, like other browsers. Recently it renders with each line of three headings broken over two lines, with the additional oddity that some single items appear left justified and some appear centred. Go to http://www.theregister.co.uk and you'll see. They changed the way they did the layout, exposing a layout issue in NetSurf. NetSurf CI builds from 2934 contain a fix. -- Michael Drake http://www.netsurf-browser.org/
Re: Updated disc cache summary
On 08/04/15 12:41, Chris Newman wrote: So given all this, on my RiscPC Strong ARMv4 Adjust 4.39 with Unipod, to what should I set the Cache parameters in NetSurf Choices? Too slow to be useful. Set disc cache size to 0. -- Michael Drake http://www.netsurf-browser.org/
Re: Change.org massively scaled images
On 03/12/14 19:10, Cristopher Dewhurst wrote: Does anyone know why NetSurf scales the images on www.change.org to what seems like several thousand percent? They use media queries, a CSS3 feature that we don't yet support. http://www.w3.org/TR/css3-mediaqueries/ Support needs added to both libcss and NetSurf. Cheers, -- Michael Drake http://www.netsurf-browser.org/
Re: links no longer work
On 22/10/14 22:26, Tony Moore wrote: Visit about:testament and if you have user in the Provenance column for the scale setting, then you'll have some other value set. The default is 100%. about:testament displays the following Oops, I meant to write about:config Yes, the problem has gone away! Good. :) -- Michael Drake http://www.netsurf-browser.org/
Re: links no longer work
On 22/10/14 00:00, Tony Moore wrote: Using RiscPC, RO 6.20, NS #2190, the links no longer work ... Does anyone else see this bizarre problem? I'm guessing you have your own scale setting configured? Visit about:testament and if you have user in the Provenance column for the scale setting, then you'll have some other value set. The default is 100%. Anyway, assuming that non-standard scale configurations are the problem, it should be fixed in the latest build. Cheers, -- Michael Drake http://www.netsurf-browser.org/
Maps
Since NetSurf doesn't have enough JavaScript support to use Google Maps, I've added a simple front end to the Google Static Maps API. You can have a go with this if you visit: about:maps in the latest NetSurf development build. -- Michael Drake http://www.netsurf-browser.org/
Re: bug about SSL security certificate -- Mantis grrrr
On 27/09/14 10:30, Martin Bazley wrote: The correct response to invalid input in a form is to display the form again, with the input the user just submitted, and the offending fields highlighted and annotated, not to redirect to a blank (and uninformative) error page and rely on a convenience feature in the browser to prevent the user's data from being summarily deleted. You could make the feature request to the Mantis developers. Anyway, this is way off topic for here. -- Michael Drake http://www.netsurf-browser.org/
NetSurf 3.2 released
The NetSurf developers are happy to announce NetSurf 3.2. This release contains many bug fixes and improvements. It is available to download from http://www.netsurf-browser.org/ Binaries are available now for RISC OS and AmigaOS 4 systems. NetSurf 3.2 is primarily a bug-fix release. In addition to bug fixes and tidy-ups, a disc cache feature has been added, and a little work has been done to improve CSS3 support. Several of the front ends have received quite a bit of attention, with new features and improvements; notably the GTK, AmigaOS 4 and Framebuffer front ends. We recommend all users upgrade. A more detailed and complete list of changes is given below. Core / All platforms * Improved building on OpenBSD. * Improved interface between front ends and core. * Improved internal URL handling. * Cleaned up many internal interfaces. * Added support for disc caching. * Fixed treeview selection issue. * Fixed line breaking of plain text with multi-byte sequences. * Improved out-of-memory handling. * Added support for overflow-x and overflow-y CSS3 properties. * Improved web search feature. * Improved internal task scheduling. * Improved support for Unicode URLs. * Fixed issue relating to CSS imports. * Improved debug dumping of box trees. * Fixed SSL certificate viewer lifetimes bug. * Fixed HTML ordered list numbering issue. * Fixed DOM tree debug dump. * Fixed handling of malformed URLs. * Fixed text selection. * LibCSS library (CSS parser and selection engine): + Added support for CSS3 overflow-x and overflow-y properties. + Fix to !important handling with play-during property. + Improved test coverage. * LibDOM library (Document Object Model): + Implemented many missing specialisations of HTMLElement. RISC OS-specific * Fix for toolbar width calculation. * Code clean-ups. * Added option to use disc cache. GTK-specific * Improved configuration handling. * Improved commandline usage. * Enabled disc caching. * Improved context menu. * Fixed gdkpixbuf usage issue. * Improved view source, and debug features. * Improved main menu. * Improved developer options. * Added options to open view source, etc, in new window, tab, or editor. * Many code clean-ups. * Internationalisation fixes. AmigaOS-specific * Removed Cairo. * Improved installer script. * Enabled disc caching. * Improved line splitting behaviour for text wrap. * Improved web search. * Improved hotlist menu. * Improved task scheduler. * Wait for network activity, rather than polling for it. * Improved throbber operation. * Fixed tab bar update failure issue. * Improved favicon handling. BeOS/Haiku-specific --- * Various fixes. Atari-specific -- * Made several little bug fixes. Framebuffer-specific * Added Unicode support to internal font. * Added codepoint rendering to internal font, for missing glyphs. * Added scaled glyph rendering to internal font. * Added support for window resizing, were underlying surface supports it. * LibNSFB library (NetSurf Framebuffer): + Made SDL surface resizeable. + Fixed rendering for 1bpp glyphs with greater than 1 byte width. Also included are many smaller bug fixes, improvements and documentation enhancements. -- Michael Drakehttp://www.netsurf-browser.org/
Re: Memory still not being released
On 29/08/14 16:53, m...@johnwoodhouse.plus.com wrote: the problem seems to persist You say persist and still; is there some context for this that I am missing? Do we have a log files for such runs? If there isn't one for this already, please open a bug report[1] with the relevant details and logs. [1] http://bugs.netsurf-browser.org/ Cheers, -- Michael Drakehttp://www.netsurf-browser.org/
Re: CSS Base
On 05/07/14 14:19, Peter Slegg wrote: Does anyone know what is wrong with the Atari daily builds? Any page just opens a dialogue box saying CSS Base (attached). No pages are rendered. Looks like it's failing to find its resources. Neither the default style sheet or the Messages file have loaded. I don't know how the Atari front end deals with loading them. -- Michael Drakehttp://www.netsurf-browser.org/
Re: Not enough application memory to start Basic
On 14/06/14 13:48, David Pitt wrote: All the NetSurf needs to run its cache is a value for Cache$Dir pointing to the chosen location. Indeed. The original proposal for Cache by Rob Kendrick[1] simply set two variables in its !Run file. These were: Caches$Write - Expands to where you should write new cache data Caches$Path - Expands to where you should read cached data We recently decided to make use of !Cache as we assumed it would be suitable and correct. (None of the core developers are actually using RISC OS at the moment, so we hadn't checked it actually worked.) Looking at the content of the supplied !Cache it seems to contain a lot of extra stuff and unexpected complexity. As shipped, it currently looks like this: !Cache/ !Cache/!Boot,feb !Cache/!Help,feb !Cache/!Run,feb !Cache/!RunImage,ffb !Cache/!Sprites22,ff9 !Cache/!Sprites,ff9 !Cache/Caches !Cache/Caches/Blank !Cache/Caches/Single !Cache/Caches/Single/MimeMan !Cache/Caches/Single/MimeMan/MimeBase,ffd !Cache/Caches/Single/MimeMan/PluginMime,ffd !Cache/Resources !Cache/Resources/MultiError,ffb !Cache/Resources/ResFind,ffb !Cache/Resources/UK !Cache/Resources/UK/Help !Cache/Resources/UK/Messages !Cache/Resources/UK/!Meta !Cache/Resources/UK/Templates,fec While we'd expect something like: !Cache/ !Cache/!Boot,feb !Cache/!Help,feb !Cache/!Run,feb !Cache/!Sprites,ff9 !Cache/!Sprites22,ff9 !Cache/Caches We can either create our own !Caches that behaves as we expect, or we can try to get the upstream !Cache fixed and continue to use that. Although !Cache does seem over-complex; I'm not sure why it needs a !RunImage, for example. Also, whichever we choose, we should move the RUfl cache to the persistent cache, rather than storing it in Scrap. [1] See this thread: http://vlists.pepperfish.net/pipermail/netsurf-users-netsurf-browser.org/2006-June/005356.html -- Michael Drakehttp://www.netsurf-browser.org/
Re: Not enough application memory to start Basic
On 14/06/14 14:30, Andrew Pinder wrote: We recently decided to make use of !Cache as we assumed it would be suitable and correct. (None of the core developers are actually using RISC OS at the moment, so we hadn't checked it actually worked.) If you give some clues as to how to test it, some of us may be able to give feedback :-) Thanks, but I think it's already been tested enough to show it doesn't work, as it stands. You should be able to put it in !Boot without it throwing up errors or having other side-effects. Our options are to get Cache from http://www.snowstone.org.uk/riscos/ fixed, and continue using it, or create our own simpler solution. Steve Fryatt: any thoughts on this? Cheers, -- Michael Drakehttp://www.netsurf-browser.org/
Fixed: BBC Sport overlapping content
At the recent RISC OS show in Wakefield, someone pointed out to me that we were making a mess of the BBC Sport site, and I said I'd post here once I'd got around to fixing it. The problem was that BBC was using the overflow-x property from CSS3 and neither LibCSS or NetSurf supported that. (We only supported the overflow property.) So LibCSS has been updated to support overflow-x and overflow-y, and so has NetSurf. The original overflow property is now a shorthand that sets both overflow-x and overflow-y. Cheers,
Re: MousAxess
In article 53d4c22179li...@torrens.org.uk, Richard Torrens (lists) li...@torrens.org.uk wrote: I hav found Netsurf very inclined to crash. It feels like it's caused bey switch bounce, usually from the menu button. But removing MouseAxess seems to stop it. Please could someone who is seeing this crash file a bug report with a Log file attached. Otherwise it is unlikely ever to be investigated. Cheers, -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: Iconv report
In article 20140125160616.ga20...@parsifal.org.uk, John-Mark Bell j...@netsurf-browser.org wrote: This is generated by UnixLib. See http://vlists.pepperfish.net/pipermail/netsurf-users-netsurf-browser.org/2012-September/010906.html for details. And GCCSDK hasn't had a release since UnixLib was fixed[1], so we're still using the old gccsdk 4.1.2 r2 release from October 2012. [1] http://www.riscos.info/websvn/revision.php?repname=gccsdkpath=%2Ftrunk%2Fgcc4%2Frecipe%2Ffiles%2Fgcc%2Flibunixlib%2Flocale%2Ficonv.crev=6017peg=6017 -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: Output from forms different in Netsurf than other browsers.
In article 53cdf35ab4brian.jord...@btinternet.com, Brian Jordan brian.jord...@btinternet.com wrote: A bit more checking show that 3.1 (Dev CI #1586) submits emails formatted the sames as other browsers whereas (Dev CI #1598) submits the different formatting. I don't have versions between these to enable me to narrow it down further. Please report it on the bug tracker and include the HTML for the page with the form. -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: Paypal bank transfer
In article 31a32dcd53.harr...@blueyonder.co.uk, Harriet Bazley li...@orange.wingsandbeaks.org.uk wrote: I can't seem to withdraw money to my bank account from Paypal using Netsurf any more (it just gives a page saying there may have been an error) -- is it a temporary glitch or have they introduced a JavaScripty dependency since I last withdrew money in October? You don't say which version of NetSurf you're testing. Anyway, try #1635, which has a SELECT element bug fix. -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: Atari daily build
In article 001f2b43.01dc85002...@smtp.freeola.net, Peter Slegg p.sl...@scubadivers.co.uk wrote: The Atari daily build is still broken. It renders the url underscores but not the fonts and then exits abruptly. At which build number did it stop working? -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: Bugs, their care and stomping
In article 20131217171425.ga3...@kyllikki.org, Vincent Sanders vi...@netsurf-browser.org wrote: [1] http://http://bugs.netsurf-browser.org That's http://bugs.netsurf-browser.org/ of course. -- Michael Drake (tlsa) http://www.netsurf-browser.org/
CSS selection speedup
Recent development builds now use an updated version of our libcss library. It makes a significant optimisation to the amount of time NetSurf spends applying CSS rules to document elements to discover their style. The improvement is in NetSurf CI builds from #1508 onward. For more details about the change to libcss and profiling results, see this posting: http://vlists.pepperfish.net/pipermail/netsurf-dev-netsurf-browser.org/2013-December/003230.html -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Summary of recent changes
Here is a brief list of the main changes made to NetSurf's core code since the release of NetSurf 3.0: + Choices behaviour improvement - Now only saves options which differ from the defaults. + about:config improved - Visiting about:config shows which settings were modified by the user. + Textarea undo/redo - Supports keyboard shortcuts to undo/redo changes made in textareas. - The shortcut keys are platform dependent. + Rewritten treeviews [*] - Faster (especially application startup time). - Fixes rendering glitches. - Improved new look. - Nicer behaviour. - Keyboard navigation. - Treeview clients rewritten for new treeview: - Hotlist - Global history - Cookie manager - SSL certificate viewer - Several bugs fixed, e.g. cookie delete. + Hotlist has default folder [*] - New additions go into the default folder, rather than into root. - Default folder has different icon. + Directory listing of file: URLs - Improved sort to fix File1, File10, File2. - Icons shown for folder/file differentiation. + Several leaks squashed - Memory leaks in various places in NetSurf fixed. - Reference counting leaks in LibDOM fixed. + URL bar hotlist indicator [*] - Indicator shows whether current URL is in the Hotlist. - Click to add/remove from hotlist. + LibCSS speedup - Faster handling of CSS selection. [*] Core change also requires front end updates which may or may not have happened for any given platform. I've left out all purely front end developments. Please test the latest development builds and let us know how you get on. Cheers, -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: Daily build 1374 - crash on Yahoo
In article 0033b38d.01dc65003...@smtp.freeola.net, Peter Slegg p.sl...@scubadivers.co.uk wrote: I noticed that it does seem to depend on which Yahoo sign-in page is used. I've not been able to reproduce this. Please give the URL of the sign-in page that leads to a problem, and the exact steps to reproduce it. -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: 3.1 (Dev CI #1330)
In article 5394b9c0b6bbai...@argonet.co.uk, Brian Bailey bbai...@argonet.co.uk wrote: Looks just great here. I can now see what is happening. Good good. Thanks for the feedback eveyone. -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: Cut to clipboard
In article 5595b59453.pitt...@iyonix.home, David Pitt pit...@pittdj.co.uk wrote: Using 1369 as far as I can see F8 and F9 work as expected. Excellent. -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: 3.1 (Dev CI #1330)
In article 538573f7ddt...@netsurf-browser.org, Michael Drake t...@netsurf-browser.org wrote: In article 53857230d8joh...@ukgateway.net, John Williams joh...@ukgateway.net wrote: My treeview display pluses and minuses have changed to 25/b8 and 25/be. They should be triangles, but you don't have any fonts containing the correct glyphs. The latest build uses the polygon plotter to draw the triangles instead. On some platforms, this means the triangles are no longer anti-aliased. -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: 3.1 (Dev CI #1330)
In article 5394515da2t...@netsurf-browser.org, Michael Drake t...@netsurf-browser.org wrote: The latest build uses the polygon plotter to draw the triangles instead. On some platforms, this means the triangles are no longer anti-aliased. That looked dreadful, so I've changed it to generate bitmaps of anti-aliased triangles of the appropriate size and colours at runtime. Is anyone testing this stuff? Also the undo/redo textarea handling. -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: Cut to clipboard
In article 5379858e4bt...@netsurf-browser.org, Michael Drake t...@netsurf-browser.org wrote: In article 20ed3b7953.harr...@blueyonder.co.uk, Harriet Bazley li...@orange.wingsandbeaks.org.uk wrote: would it be possible to have areas deleted by overtyping automatically cut to the clipboard, a la Impression? Or at least have a do you really want to do this? warning, a la Impression II? Editing blunders should be handled by undo/redo. NetSurf's textarea doesn't currently support this. Supported via keyboard shortcut in the latest build. The shortcut depends on the front end. ++==+==+==+ | GTK Framebuffer | RISC OS | ++==+==+==+ | Undo | Redo | Undo | Redo | |+--+--+--| | Ctrl+Z | Ctrl+Y or| F8 | F9 | || Ctrl+Shift+Z | | | ++==+==+==+ The other front ends don't yet pass KEY_UNDO / KEY_REDO to the core. Note that on RISC OS, F8 is undo when the caret is in a textarea, and view source otherwise. -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: Cut to clipboard
In article out-524b1c1a.md-1.4.17.chris.yo...@unsatisfactorysoftware.co.uk, Chris Young chris.yo...@unsatisfactorysoftware.co.uk wrote: I've removed the shortcut from Clear selection as I don't think it needs one anyway (the Style Guide didn't have an opinion on that or Redo). The Escape key normally acts the same as clear selection. Are there any can_undo()/can_redo() type functions? No. At least, not yet. -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: Hotlist editing
In article 6f29e49253.harr...@blueyonder.co.uk, Harriet Bazley li...@orange.wingsandbeaks.org.uk wrote: It is (as of at least v1356, and currently in v1359) no longer possible to edit/manually update the URLs in a hotlist, something which I used do on a regular basis. Alt+Click? It should behave like the filer. -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: Atari daily build 1344
In article 0003b9c1.0234e500a...@smtp.freeola.net, Peter Slegg p.sl...@scubadivers.co.uk wrote: bash-2.05a# ./ns020.app ns020.app: atari/gemtk/guiwin.c:810: gemtk_wm_get_grect: Assertion `win != ((void *)0)' failed. Abort Can you get a backtrace? Also, run with -v and put the log output in a pastebin[*]. [*] e.g. http://pastebin.com/ -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: Cut to clipboard
In article 20ed3b7953.harr...@blueyonder.co.uk, Harriet Bazley li...@orange.wingsandbeaks.org.uk wrote: would it be possible to have areas deleted by overtyping automatically cut to the clipboard, a la Impression? Or at least have a do you really want to do this? warning, a la Impression II? No, the clipboard should only be modified by the operations that specifically affect it; cut to clipboard, and copy to clipboard. Editing blunders should be handled by undo/redo. NetSurf's textarea doesn't currently support this. -- Michael Drake (tlsa) http://www.netsurf-browser.org/
Re: Sourceforge
In article 8ecaf06e53@abbeypress.net, Jim Nagel nets...@abbeypress.co.uk wrote: Michael Drake wrote on 7 Jul: Has anyone managed to submit a new bug report with NetSurf since SourceForge changed their site? Any upshot from this, Michael? Noone replied, so I assume it doesn't work in NetSurf now. I'll raise the issue when the developers meet on Aug 10th, if nothing's been sorted out by then. -- Michael Drake (tlsa) http://www.netsurf-browser.org/