Re: Atari build
On 29/12/2023 12:02, psl...@scubadivers.co.uk wrote: Hi all, I hadn't updated Netsurf on my Atari 68060@60MHz) since before January. I have always found the Atari build to be extremely slow, 5-10 minutes or more to render a page being normal and I think it's the ssl that hurts it most. So it's not greatly usable but I do use it download FreeMint OS updates. 60MHz is not very much to work with, tbh, so I am not entirely surprised that things are slow, particularly given the ever-increasing complexity of the modern web. However, the last few versions I had tried would timeout before rendering anything. That was why I was still using build 5366 from 2022. Yesterday I installed 6637 and it still gives Timeouts. The only page that would work was http. I increased the curl timeout to 300 with no change. All of the support libraries have been upgraded in the last year, including migrating to OpenSSL 3. These are significant changes so may have introduced some brokenness at some point. Sadly, the Atari frontend has no maintainer (any changes are mostly to make it build after changes to other parts of the code) so, although CI continues to spit out binary builds, these are never tested (and none of the core developers has any specific Atari knowledge or any appropriate hardware/emulator so cannot test them, either). As ever, if someone is able to work out a fix, we'll happily take a patch containing it! John-Mark. ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: allocator_claim_pages:449
On 01/12/2023 09:57, Martin Avison wrote: In article <8e67b10b5b.harr...@bazleyfamily.co.uk>, Harriet Bazley wrote: Since downloading v6511 of Netsurf I keep seeing "Unknown dynamic area" errors pop up in my Reporter window. These are 'silent' errors that don't generate a WIMP error box, so presumably the user is not supposed to be aware of them, but it gives the impression that the app is running out of memory Reporter 2.72 (15 Aug 2020) List Fri 1st Dec 2023 01:11 01:06:46.54 ** Clear ** from Menu allocator_claim_pages:449 - Unknown dynamic area allocator_claim_pages:449 - Unknown dynamic area allocator_claim_pages:449 - Unknown dynamic area allocator_claim_pages:449 - Unknown dynamic area allocator_claim_pages:449 - Unknown dynamic area allocator_claim_pages:449 - Unknown dynamic area allocator_claim_pages:449 - Unknown dynamic area allocator_claim_pages:449 - Unknown dynamic area allocator_claim_pages:449 - Unknown dynamic area Reporter 2.72 (15 Aug 2020) Listed 10 lines They do not look like messages caused by a RISC OS error being raised, but more like messages produced by a specific Reporter call within the program. These appear to emitted by the ARMEABISupport module. There is, to my knowledge, no "official" release build of this, so we made sure to provide the same binary as was being shipped by other applications to ensure that everything was consistent. That may well mean that it's actually a debug build of the module, hence the logging. John-Mark. ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: Typing Cyrillic
On 21/05/2022 10:09, Harriet Bazley wrote: On 20 May 2022 as I do recall, John-Mark Bell wrote: Additionally, note that the character encoding used to submit forms on websites is determined from the web site itself (and has nothing to do with whatever settings apply to the OS on which the browser is running). In the case of Google, the search page they serve to NetSurf does not specify a charset to use for form submission, so the encoding of the web page will be used. Page -> Info will tell you that this is ISO-8859-1 (i.e. Latin 1), which is not able to represent Russian, thus you will find that attempting to search Google for Russian text will end up with NetSurf submitting a load of question marks, instead. Other search engines (e.g. DuckDuckGo, Yahoo) work fine as NetSurf is able to submit UTF-8 encoded text to those (and thus Russian is representable). I wondered if using www.google.ru would work (on the assumption that this page presumably expects Russian input), but it doesn't. (Page Info gives "windows-1251", but the search result page returned then shows up as ISO-8859-1 again) And clicking on "Google offered in: russkiy" (in the absence of any ability to submit Cyrillic in e-mail!) at the bottom of the page gives an error; doing the same thing on google.de, for example, does successfully switch the page over into searching in German. As of CI build #5346, Google will serve up pages using the UTF-8 encoding, so all this mess goes away. J. ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: #5345 keeps scanning the fonts
On 29/05/2022 17:37, John-Mark Bell wrote: On 29/05/2022 13:16, Paul Sprangers wrote: Dear developers, The latest build (#5345) keeps scanning the active font folders at each start up and quits with an error before ending it. I had to revert to #5342 to get rid of this behaviour. Please send me (off-list) a log file showing the error, and any and all fonts that it cannot process. To close the loop here: in this case, the issue was not with any individual font. Instead, it was caused by the presence of more than 255 unique fonts on Paul's system -- a single code path failed to properly account for this with the result being corrupted memory and undefined behaviour. This is fixed in build #5346. J. ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: #5345 keeps scanning the fonts
On 29/05/2022 19:10, Paul Sprangers wrote: Please send me (off-list) a log file showing the error, and any and all fonts that it cannot process. The font that it cannot process is UniFont. When I remove that one from the font directory, NetSurf behaves nicely and won't scan the fonts next time that it is started. I will send you the log file privately. Do you wish to receive the font in question as well? Yes. J. ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: #5345 keeps scanning the fonts
On 29/05/2022 13:16, Paul Sprangers wrote: Dear developers, The latest build (#5345) keeps scanning the active font folders at each start up and quits with an error before ending it. I had to revert to #5342 to get rid of this behaviour. Please send me (off-list) a log file showing the error, and any and all fonts that it cannot process. Pi4 / RISC OS 5.29 (19-Feb-21) and some large unicode fonts. Switching the unicode fonts off avoids the error, but NetSurf will still keep scanning the fonts at each new boot. If the scan completes, then it will not scan again unless one or more of the following is the case: 1. The installed fonts change 2. The cache file in !Scrap does not exist Note that the cache file has moved in recent versions of NetSurf so, if you have something that cleans out scrap on boot, it will need to be updated to preserve the RUfl cache. J. ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: Typing Cyrillic
On 20/05/2022 17:58, Harriet Bazley wrote: How do I type Cyrillic into Netsurf? I have the necessary fonts for the browser to *display* it, but I can't, for example, type a Russian word into a Google search box or quote a line of Russian poetry in my blog. (I'm pretty sure I have managed it in the past by the incredibly laborious technique of searching using transliterated Latin text for a Web page containing the relevant word in the right case and capitalisation and copying and pasting the text from one window to another, but the process is rather like trying to compose a letter from words cut out of the newspaper!) On RISC OS 5, you can use the following (cumbersome) approach: 1. Open a Task window and enter *Country This will output the current Country setting. (We're assuming here that you don't have some non-standard setup where you have explicitly configured an Alphabet and/or Keyboard the differ from the Country's default -- if you have done so, then enter *Alphabet and *Keyboard to obtain their current settings). 2. Enter *Country Russia in the Task Window This will set the Country to Russia, the system alphabet (*Alphabet) to Cyrillic and the keyboard (*Keyboard) to Russia. (If you prefer, or have non-standard settings as described above, you can set *Alphabet and *Keyboard explicitly instead of using *Country. Do make sure that you set both settings, though -- the system alphabet must be either Cyrillic or UTF8 for this to work, and you obviously need to select the correct keyboard driver, too) 3. Press the left Alt and Shift keys together The Russian keyboard driver has two layers, which may be switched between by pressing the left Alt and Shift keys simultaneously. The base layer is equivalent to a US keyboard, the alternate layer is the Russian 104 key layout. Here we have switched to the alternate layer. 4. Type the text you want into NetSurf The keycaps on your physical keyboard won't help you now -- you'll need to know the layout of a Russian keyboard. If you don't, then you can find a Drawfile containing the relevant layout at [1]. 5. Press the left Alt and Shift keys together This switches the keyboard driver back to the base layer (US layout) 6. Enter *Country in the Task Window This sets the Country/Alphabet/Keyboard settings back to what they were before. here is that output in step 1, above. Again, use *Alphabet and *Keyboard instead, if these are relevant to you. If you think the above is unpleasant, then you would be correct. Unfortunately, this is an excellent example of an area where RISC OS is resolutely stuck in the 1980s. It will require significant work on the OS itself to improve matters. Additionally, note that the character encoding used to submit forms on websites is determined from the web site itself (and has nothing to do with whatever settings apply to the OS on which the browser is running). In the case of Google, the search page they serve to NetSurf does not specify a charset to use for form submission, so the encoding of the web page will be used. Page -> Info will tell you that this is ISO-8859-1 (i.e. Latin 1), which is not able to represent Russian, thus you will find that attempting to search Google for Russian text will end up with NetSurf submitting a load of question marks, instead. Other search engines (e.g. DuckDuckGo, Yahoo) work fine as NetSurf is able to submit UTF-8 encoded text to those (and thus Russian is representable). J. 1. https://gitlab.riscosopen.org/RiscOS/Sources/Internat/IntKey/-/blob/master/Layouts/Russia,aff ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: Netsurf suddenly fails to start
On 16/08/2021 19:09, Dave wrote: VRPC-DL RISC OS 6.20 Help please. Recently it was okay, but suddenly today Netsurf fails to run, presenting the following. " ** WimpError ** from NetSurf Error : &0001 Message: The Unicode font library could not be initialized. Please report this to the developers." What, why and how to fix it? It is likely you have recently added some new fonts to your system which are causing the library to get confused. Shift-double-click on !NetSurf, and run the OpenScrap script within. Then send the log file to me (off list) and we'll see what it says. J. ___ netsurf-users mailing list -- netsurf-users@netsurf-browser.org To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
Re: TLS Security NS 3.8
On 05/07/2019 05:01, ferrit...@yahoo.com wrote: Little more than a week ago I posted about the Security Certs for NS 3.8. I was not aware at that time that NS 3.9 was already available (I was using a link provided for D/L of 3.8). Since there has been other bugs/problems, I thought to provide the actual results. The location of this Qualys Client Test is https://www.ssllabs.com/ssltest/viewMyClient.html Presuming the Certs are within NS 3.8, it would appear that the "weak" certs be removed for added security. I did not receive an answer to the question if the certs are tapped from the Distribution or the Browser. So, here are the results... This test does nothing with certificates. However, the answer as to which certificates get used depends upon the platform you are using. If Linux, it will, by default, use the standard system-wide CA certificate store (usually found in /etc/ssl/certs). Protocols TLS 1.3 No TLS 1.2 Yes* TLS 1.1 Yes* TLS 1.0 Yes* SSL 3 Yes* SSL 2 No These are not an accurate reflection of reality -- the test relies on support for more Javascript (and associated things) than NetSurf has. NetSurf supports TLS1.0/1.1/1.2. SSL2/3 are disabled. TLS1.2 will always be used by preference. Cipher Suites (in order of preference) TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384 (0xc030) Forward Secrecy 256 TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384 (0xc02c) Forward Secrecy 256 TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA384 (0xc028) WEAK 256 TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA384 (0xc024) WEAK 256 TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256 (0xc02f) Forward Secrecy 128 TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256 (0xc02b) Forward Secrecy 128 TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256 (0xc027) WEAK 128 TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA256 (0xc023) WEAK 128 TLS_DHE_RSA_WITH_AES_256_GCM_SHA384 (0x9f) Forward Secrecy 256 TLS_DHE_RSA_WITH_AES_256_CBC_SHA256 (0x6b) WEAK 256 TLS_DHE_RSA_WITH_AES_128_GCM_SHA256 (0x9e) Forward Secrecy 128 TLS_DHE_RSA_WITH_AES_128_CBC_SHA256 (0x67) WEAK 128 TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA (0xc014) WEAK 256 TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA (0xc00a) WEAK 256 TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA (0xc013) WEAK 128 TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA (0xc009) WEAK 128 TLS_DHE_RSA_WITH_AES_256_CBC_SHA (0x39) WEAK 256 TLS_DHE_RSA_WITH_AES_128_CBC_SHA (0x33) WEAK 128 TLS_RSA_WITH_AES_128_CBC_SHA (0x2f) WEAK 128 TLS_EMPTY_RENEGOTIATION_INFO_SCSV (0xff) - (1) When a browser supports SSL 2, its SSL 2-only suites are shown only on the very first connection to this site. To see the suites, close all browser windows, then open this exact page directly. Don't refresh. Qualys currently marks all CBC ciphersuites as "weak", as a result of a preponderance of padding oracle issues in implementations. If built against a modern OpenSSL, there are no currently known issues here. CBC suites will remain enabled in NetSurf until such time as they are not required for compatibility with web servers that don't support GCM. J.
Re: shortcut "font" in CSS
On 22/06/2019 19:09, Jim Nagel wrote: Does the shortcut version of the "font" property in CSS work properly in Netsurf? Yes, it does. You can find extensive test data at [1]. This CSS gives me the display I want: font-family: serif; font-style: italic; font-weight: bold; font-size: 24pt; This single CSS expression is supposed to be equivalent, but Netsurf ignores it: font: serif italic bold 24pt; Or am I failing to understand the spec? I'm afraid so. The spec[2,3] says that (assuming none of the single-token values are specified -- e.g. caption or inherit): * font-style, font-variant, font-weight appear first (in any order), if required * font-size comes next (and is always required) * an optional line-height may be specified next * font-family appears last So, given the long-hand you have specified above, you want something like: font: bold italic 24pt serif; J. 1. https://git.netsurf-browser.org/libcss.git/tree/test/data/parse2/font.dat 2. https://www.w3.org/TR/CSS21/fonts.html#font-shorthand 3. https://www.w3.org/TR/css-fonts-3/#font-prop
Re: Crash on loading NetSurf
On 19/02/2019 22:29, Andrew Pinder wrote: Message from NetSurf: "The Unicode font library could not be initialised. Please report this to the developers." I don't believe this is a bug - it's the aftermath of a crash. This ARMini (RO 5.22) had been out of use for a period due to a house move. I had been using release 4335, which seemed stable. I downloaded release 4528 as the then most recent. This ran but wasn't displaying some pages correctly. I think the next day I had a crash while downloading email with NetFetch. This caused disc problems. After using DiscKnight and repairing Messenger I then tried loading NetSurf. I got the above error message. I deleted r4528 and reverted to r4335. The error message still occurs on attempting to load NetSurf. I double checked the modules in System to make sure they were the ones that came with r4335. My best guess is that a Unicode file has been corrupted. How do I check? Where do I find an up to date version of Unicode to install? The Unicode resource (!Unicode) has very little to do with the Unicode font library used internally by NetSurf. Please provide a NetSurf log file (which can be found by running the OpenScrap obey file stored in the NetSurf application directory -- shift double click on !NetSurf to open it) from a crashing start -- this contains copious information which will allow the developers to understand what is going on. In all likelihood you have a font installed that causes it to get confused. I thought we'd found and fixed all of the issues here over the last decade, but clearly not. J.
RISC OS/RPi 3 support
All, Build 3460 or later should run fine on the Raspberry Pi 3 under RISC OS. As none of the developers have such hardware, please could somebody who does please test this and ensure it works. Thanks, John-Mark.
Re: MousAxess
On Fri, Feb 07, 2014 at 03:19:20AM -0800, Dave Higton wrote: On Thu, 6 Feb 2014 17:25:34 + John-Mark Bell wrote: More generally; all issues should be reported on the issue tracker, with sufficient detail to describe the problem fully. Any issue which is not reported in this way will almost certainly not be investigated, and will certainly not be tracked properly, so will be forgotten. So, if you want something fixed, then report it on the issue tracker! In general, I agree with you. However, the MousAxxess and sockets issues were not obviously Netsurf's fault. Discussion on list has helped add information. This discussion would have not taken place if the OP had done nothing more than reported to the issue tracker. Discussion having taken place, all relevant information should go into a report in the issue tracker. I sense you're missing my point: if the issue is not reported in the issue tracker, it will not cross the radar of the developers. Discuss what you want, where you want, but the only sure way of ensuring that a developer is aware of the issue is to report it on the issue tracker. In fact, if you can go a step further and include a link to the relevant discussion in the mailing list archives, that would be even better! John-Mark.
Re: MousAxess
On Fri, Feb 07, 2014 at 12:05:12PM +, Richard Torrens (lists) wrote: In article 53d5cd2686t...@netsurf-browser.org, Michael Drake t...@netsurf-browser.org wrote: That's http://bugs.netsurf-browser.org/ of course. Well I went there, then to the next page. How do I report a bug? You log in, click the Report Issue link, fill in the form, and click Submit Report. Life's too short to open accounts and log in or to work out complicated routines, sorry. The report form requires merely 3 fields to have content. However, the more information you provide in a report, the more likely it will be looked at. A report saying it crashed, and attaching a log file is very unlikely to have much priority; particularly when other reports contain much more circumstantial information about *what* the user was doing when things went wrong. This kind of information is invaluable when investigating issues, even if it seems irrelevant to the reporter, so please include as much detail as possible when reporting issues -- more information is always better in this case. As for creating an account, we require this so that we can continue to contact the reporter for feedback when things are not clear from the information provided. Anonymous reports prevent this, and are a waste of our time, and the reporter's. As I said in another thread; the *only* place to report issues to the developers is the issue tracker. If you are unable or unwilling to report issues there, then they are guaranteed not to get investigated or fixed, unless someone else happens to report the same issue. John-Mark.
Re: MousAxess
On Fri, Feb 07, 2014 at 03:09:51PM +, Daniel Silverstone wrote: Unfortunately the old tracker was hard for the developers to use and stopped being accessible from NetSurf entirely thanks to SourceForge fecking up. All of this was covered in Vince's email announcing the new bug tracker: http://vlists.pepperfish.net/pipermail/netsurf-users-netsurf-browser.org/2013-December/012100.html Please read this if you haven't already done so! John-Mark.
Re: MousAxess
On Fri, Feb 07, 2014 at 02:19:05PM +, Richard Torrens (lists) wrote: In article 20140207125751.gb5...@parsifal.org.uk, John-Mark Bell j...@netsurf-browser.org wrote: You log in, click the Report Issue link, fill in the form, and click Submit Report. That page, at the bottom has a link: bug reporting interface. That takes me to http://bugs.netsurf-browser.org/mantis/my_view_page.php which has no Report Issue link... I suspect the key is the my_view_page bit. It probably only works for logged-in members As I explained in the previous email, you must log in to the bug tracker to report issues. John-Mark.
Re: MousAxess
On Thu, Feb 06, 2014 at 04:26:13PM +, Michael Drake wrote: 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. More generally; all issues should be reported on the issue tracker, with sufficient detail to describe the problem fully. Any issue which is not reported in this way will almost certainly not be investigated, and will certainly not be tracked properly, so will be forgotten. So, if you want something fixed, then report it on the issue tracker! John-Mark.
Re: Iconv report
On Sat, Jan 25, 2014 at 03:23:42PM +, m...@johnwoodhouse.plus.com wrote: Can someone explain why (at least in !Netsurf 3.0) the !Run file contains : RMEnsure Iconv 0.12 NetSurfRMLoad System:Modules.Iconv RMEnsure Iconv 0.12 Error NetSurf requires Iconv 0.12 or later. This can be downloaded from http://www.netsurf-browser.org/iconv/ - and this is the version I have running according to !Verma, NetSurf requires Iconv 0.12, which the Run file is ensuring. and yet, running !Reporter, I see that when running !Netsurf for the first time, there is a delay while this is repeated no fewer than 189 times: ! [Appl/NetSurf] RMEnsure Iconv 0.04 RMload System:Modules.Iconv [Appl/NetSurf] RMEnsure Iconv 0.04 Error 16_10F iconv support requires the Iconv module 0.04 or newer on subsequent page loads it is repeated a mere 66 times. Apart from the question of why this is repeated so many times, so slowing up page loads, where is this coming from when Iconv 0.04 is called for in the !Run file, and in any case Iconv 0.12 is running? This is generated by UnixLib. See http://vlists.pepperfish.net/pipermail/netsurf-users-netsurf-browser.org/2012-September/010906.html for details. John-Mark.
Re: Fwd: Re: Netsurf/atari build 1501
On Wed, Nov 27, 2013 at 09:28:52PM +0100, Ole wrote: Btw. since libdom is used (2.9 used libxml) css matching got slower. Evidence, please. With libdom, this should be significantly faster, as string comparisons no longer need consider the actual string data. J.
Re: Orpheus Home page
On Wed, May 01, 2013 at 04:45:29PM +, Tony Moore wrote: On 1 May 2013, Daniel Silverstone dsilv...@netsurf-browser.org wrote: As for the 'choices bug' we have no idea what's going on there. Again I don't think we've touched anything in that yet. There has been a discussion of the problem on csa.apps, several people having noticed the same effect. Issues with NetSurf should be reported here or on the bug tracker. Anything reported anywhere else is highly unlikely to attract the attention of NetSurf's developers. Certainly, none of the core developers read csa.* at all. J.
Re: Upper character truncation - CSS handling
On Mon, 2013-04-15 at 21:59 +0100, John Williams wrote: It needs sorting! Patches are always welcome. J.
Re: NetSurf on Virtual Acorn
On Sat, 2013-03-30 at 10:18 +, Michael Drake wrote: In article op.wuqk4kifif396l@alansall, Alan Sally jare...@xtra.co.nz wrote: 2. On my Risc PC running 3.5 the NetSurf icon disappears with all versions I've tried since #734. When loaded there is a blank space on the icon bar. I can click on it and NetSurf appears and runs normally. The icon appears in the directory viewer immediately after download but disappears from the viewer leaving only the name once the programme is loaded. Sounds like something went wrong here: http://source.netsurf-browser.org/netsurf.git/commit/!NetSurf?id=a6b117fcb43b8afea584ff2e6ec3e808f27f8d9f I think the answer is probably to simply remove the ASprites. Or just ensure that a sufficiently modern SpriteExtend[1] is present, which is what I've done. However, it's worth reiterating that the earliest version of RISC OS we actually support is 4.02, so anyone running anything earlier is running the risk of things breaking. J. 1. Version 0.99, as found in RISC OS 3.60, which is rapidly approaching its 18th birthday. Seriously, people, upgrade your OS or don't expect modern software to run on it.
Re: Iconv 0.12 released
On Wed, 2013-01-30 at 09:50 +, ChrisF wrote: In message 1359504204.31757.24.camel@duiker John-Mark Bell j...@netsurf-browser.org wrote: Version 0.12 of the Iconv module is now available for download. This is a recommended update for all users - new builds of NetSurf will require it. Hi, I notice the Stubs folder. After Boot and System merging, can I remove the Stubs folder from my computer or should it be stored somewhere, and if so where? The ReadMe says the following: Note for developers: The stubs directory contains source for a set of C stubs. See the ReadMe file in that directory for further information. Are you a developer? If not, then you can ignore the stubs directory. J.
Re: Log-in problem
On Wed, 2013-01-30 at 21:22 +, Richard Porter wrote: I log into my eBlah forum permanently using a cookie. Previously when I visited the forum I would be logged in automatically, but for the last couple of weeks or so I have had to log in every session. Not even my username is remembered. Could this have been caused by a change in NetSurf? It's impossible to say, given the information you've provided. Please set suppress_curl_debug to 0 in the Choices file, produce a suitable login failure, and submit a bug report, attaching the generated log. Thanks, J.
Iconv 0.12 released
Version 0.12 of the Iconv module is now available for download. This is a recommended update for all users - new builds of NetSurf will require it. New in this version: * Add proper transliteration behaviour when requested using //TRANSLIT. Fixed in this version: * Correct handling of trailing valid shift sequences. Previously would erroneously report EINVAL, instead of silently accepting them. Changed in this version: * Master alias mapping file now lives in ROOL repository. About Iconv === Iconv is a module providing character set conversion facilities for third-party applications. It is currently used by a number of applications, most notably NetSurf. UnixLib's iconv support also requires this module. Updating from earlier versions == If updating from earlier versions, please perform _both_ a !Boot and !System merge with the contents of the Iconv distribution archive as documented in the Installation section of the ReadMe file. Where to get it === Iconv may be downloaded from http://www.netsurf-browser.org/projects/iconv/
Re: Clipboard not working #824
On Mon, 2013-01-21 at 09:35 +, John Rickman Iyonix wrote: There appears to be a problem with the clipboard in 3.0 #824. Within a NetSurf window ctrl-c and ctrl-v work as normal but anything put on the clipboard from outside the NS window does not get inserted into writable icons using ctrl-v. Yes; the way the clipboard is handled internally has changed, and it's waiting for the RISC OS frontend to be updated to cope. J.
Re: RISC OS Netsurf on Beagleboard
On Mon, 2013-01-07 at 18:38 +, Christoper Dewhurst wrote: Hi John In message 1357543705.19631.12.camel@duiker John-Mark Bell j...@netsurf-browser.org wrote: On Sat, 2013-01-05 at 17:21 +, Christoper Dewhurst wrote: Hi all Does anyone have any ideas what might be going wrong? Not with the information you have provided so far. Please provide a NetSurf log file from the crash. The error boxes Internal error on data transfer at followed immediately by Filecore in use render the machine unusable, i.e. it nearly freezes (nearly meaning when you move the mouse the pointer will respond after about 5 seconds but otherwise does not respond to input.) Sadly, you've elided the most important part of the error report; namely the address at which it happens. What, precisely, does contain in this message? J.
Re: RISC OS Netsurf on Beagleboard
On Mon, 2013-01-07 at 21:08 +, CHRISTOPHER DEWHURST wrote: Hi John Sadly, you've elided the most important part of the error report; namely the address at which it happens. What, precisely, does contain in this message? It changes each time however some examples are: 202A4D70 202C8430 202B48B0 (so I guess all in the region 2020) Right, so the abort is in a module. Please do the following: 1) Run NetSurf 2) Open a TaskWindow, and run *modules 3) Save the contents of the TaskWindow to a file 4) Do whatever it is you do to provoke the crash 5) Note down the exact value of Then, send me the value you noted in step 5 and the file you saved in step 3. Thanks, J.
Re: RISC OS Netsurf on Beagleboard
On Sat, 2013-01-05 at 17:21 +, Christoper Dewhurst wrote: Hi all There is a discussion over on the ROOL site about Netsurf crashing on my Beagleboard: https://www.riscosopen.org/forum/forums/11/topics/1610?page=1 Tried switching everything off/on/round in the options, checking other apps aren't running etc. but with no success, Does anyone have any ideas what might be going wrong? Not with the information you have provided so far. Please provide a NetSurf log file from the crash. J
Re: RISC OS Netsurf on Beagleboard
On Sun, 2013-01-06 at 21:48 +0100, Simon Voortman wrote: How many colours are you using? I am one of the people who's using NetSurf on a Beagleboard, image quality is set at Error diffused and Dithered, BUT... I have to use 256 or 16 M colours. If I dare to use 32 k colours, as soon as Netsurf has to redraw some graphics, I'm getting an errorbox which tells me NetSurf has detected a serious error and must exit. This is precisely the symptom expected if you have not configured image rendering to Use OS. The Tinct module performs unaligned data accesses when rendering images in 32k modes. J.
Re: RISC OS Netsurf on Beagleboard
On Sun, 2013-01-06 at 23:16 +0100, Simon Voortman wrote: Not that I ever had to do something special to get NetSurf working on the Beagleboard (besides setting Alignment exceptions off) that I remember... There is no reason to disable alignment exceptions: NetSurf works perfectly well with them enabled. J.
Re: NetSurf Developer Workshop
On Wed, 2012-11-07 at 22:16 +, Chris Gransden wrote: In article 20121107215800.GB2863@somnambulist.local, Daniel Silverstone dsilv...@netsurf-browser.org wrote: On Wed, Nov 07, 2012 at 07:07:00PM +, george greenfield wrote: Re javascript and NetSurf, could the 'jsoff' component in the test-build titles be meaningful, I wonder ;-) Indeed. Until and unless we can get libmozjs ported to RISC OS, it'll be jsoff there all the time :-( D. I got 1.8 and 1.7 to build for RISC OS. The javascript shell runs ok but !NetSurf crashes as soon as it hits some javascript. That's as far as I got. Sadly, this doesn't provide enough information to hasten the addition of libmozjs to our SDK or to help us diagnose the issues you're experiencing. J.
Re: Request for feedback
On Sat, 2012-10-06 at 01:14 +, Chris Young wrote: Platform: AmigaOS 4.1 File: NetSurf-2012-10-05_23-21-14.lha URL: http://www.danvk.org/wp/dragtable The above site freezes NetSurf as it fetches/displays. I can't reproduce on the same Windows auto-build and have no way of testing on any of the other builds currently. We can't reproduce this either, which suggests that it's an Amiga-specific issue. That page does contain copious amounts of non-latin text, if that helps. J.
Re: Request for feedback
On Sat, 2012-10-06 at 13:17 +, Chris Young wrote: Platform: AmigaOS 4.1 File: NetSurf-2012-10-05_23-21-14.lha URL: http://us.asmodee.com Visiting the above site and clicking on Games causes a crash in libdom, as per the below stack trace: Fixed. J.
Re: Crashes on sites with frames
On Sat, 2012-10-06 at 16:31 +0100, Chris Gransden wrote: Platform RISC OS OS5.19 on Pandaboard version netsurf-2012-10-06_14-18-18/zip browsing http://www.archivemag.co.uk error log Fixed. J.
Re: User feedback requested (esp. RISC OS)
On Tue, 2012-10-02 at 09:53 +0100, David H Wild wrote: In article 52d7fa45b1t...@netsurf-browser.org, Michael Drake t...@netsurf-browser.org wrote: In article 52d7f97c49dhw...@talktalk.net, David H Wild dhw...@talktalk.net wrote: The readme file needs replacing, as it applies to the older system and tells you how to install the boot and system files which are not supplied with the new version. The !Boot and !System directories will be added at some point. I still think that the readme file should reflect what is in the ip file. It wouldn't be much work to change it when needed. Firstly, development builds come with a very large health warning as they represent the current state of development which, in all likelihood, is broken. Please don't confuse them with stable, numbered, release versions. Secondly, we have not yet announced the return of development builds for general use. Michael requested early feedback about the stability of the software only so we could see how widespread a problem the machine freezing behaviour was. There will be an announcement when development builds are available for general use; right now, they should be considered harmful. J.
Re: User feedback requested (esp. RISC OS)
On Sat, 2012-09-29 at 21:54 +0100, Martin Bazley wrote: I've never had problems of this scale with any other site. Whatever andyfanton.com does, NetSurf really doesn't like it! In the interests of preventing this regression occurring a fifth time, perhaps 'Jenkins' could be amended to make sure NetSurf can cope with a fetch and display of that specific page? Stopping regressions what automated testing is for, right? A Jenkins job just runs a command. There is currently no infrastructure for this kind of testing. Patches to improve the situation are always welcome. J.
Re: Current version?
On Tue, 2012-05-29 at 22:15 +0200, John Williams wrote: In article 5285034c43joh...@ukgateway.net, John Williams joh...@ukgateway.net wrote: We will enable the autobuilder again once trunk NetSurf reaches a satisfactory state. Could this be announced here, please, when 'normal service resumes'? Is there any news on this? No. John-Mark.
Re: NetSurf crashing on cpc site
On Fri, 2012-05-25 at 09:35 +0100, lists wrote: The home page: http://cpc.farnell.com/ works fine so does a search on something like inverter but trying to select or search for an individual item such as PW02762 results in an immediate crash Please provide the following information[1]: 1. The version of NetSurf you are using 2. What hardware and operating system version you are using 3. Whether this issue is reproducible 4. The full URL of the page that causes problems John-Mark. 1. We document our bug reporting guidelines here: http://www.netsurf-browser.org/documentation/info#Bugs Please endeavour to follow them as it saves everybody time.
Re: Fw: snag on installing Netsurf
On Tue, 2012-05-08 at 16:08 +0100, DJ BRADLEY wrote: I am in process of installing Netsurf on my Riscpc. I downloaded the system on to a CD on my PC computer, then used this to transfer to the Risc machine. This method seems to work excpt that I have not succeeded with the Boot Merge. I get a well known error message saying that an application to load the file has not been found, etc. I have not been able to get round this - can you advise please? The above sounds very wrong. What, exactly, did you do when attempting the Boot Merge operation? Resource installation is documented here: http://www.netsurf-browser.org/documentation/resinstall Alternatively can you expand on drag the supplied !Boot over your existing boot structure? (quoting from readme) As it says: drag the !Boot from the NetSurf archive over the !Boot in the root of your hard drive. John-Mark.
re: netsurf-sdl on Windows?
On Tue, 2012-05-08 at 09:44 -0400, LM wrote: Why not use the native Windows frontend? The framebuffer frontend is totally unsuitable for use on systems that have a full-featured window manager. I was hoping to use the SDL version on a couple of platforms including Windows. I suppose I could build each one natively though. If I can get it to build, long-term plan was not to use it as a web browser, but more of a cross-platform replacement for dialog. Was thinking I could possibly replace the part that sends form data to a server with the ability to store the information locally to a file. I wouldn't even need the web access capabilities, just the capabilities to interpret HTML and CSS and render pages. There are cross-platform replacements for dialog, but I'd really like to have something that can use standard web syntax (HTML, CSS, etc.) for creating interfaces. I really don't think NetSurf is suited to this scenario at all. There are many assumptions within NetSurf that there is a server to talk to. You'd probably be better off looking for a standalone rendering engine, which NetSurf is not (and is unlikely to be in the short to medium term). When I run make TARGET=windows I see no error messages. I redirected standard output and standard error to a file. The reason I asked for the exact build output is that it saves us both time and speculation. Please do the following: 1) make TARGET=windows clean 2) make TARGET=windows Q= buildlog-win 21 3) Send the verbatim buildlog-win to me Checking the build-windows-windows directory, I see object files (108 of them). There are no libraries or executables built. The output binary is placed in the top level of the source tree, not in the build-* directory. For the Windows target, it is called NetSurf.exe When I tried running make TARGET=framebuffer, I got several errors back. Some of them include: image/mng.c:29:20: fatal error: libmng.h: No such file or directory [ I thought I had MNG disabled, but appears framebuffer mode still wants the library installed. ] It uses the same configuration file as all the other frontends, so it's unlikely to explicitly want it. How have you disabled MNG? framebuffer/thumbnail.c:66:16: error: 'NSFB_SURFACE_RAM' undeclared (first use in this function) framebuffer/thumbnail.c:71:41: error: 'NSFB_FMT_XBGR' undeclared (first use in this function) framebuffer/thumbnail.c:73:2: error: incompatible type for argument 1 of 'nsfb_init' framebuffer/thumbnail.c:90:2: error: too many arguments to function 'nsfb_plot_copy' These failures are symptomatic of building against the wrong version of libnsfb. If you are building from a release tarball, then the correct libraries are bundled with it. If you are building from an SVN snapshot, please ensure that you have snapshots of the libraries from the same revision. framebuffer/convert_image.c:22:17: fatal error: png.h: No such file or directory [ My png.h is in /usr/local/include. I'm guessing if this was switched from png.h to png.h, it'll find it. ] No, it won't. You need to ensure that /usr/local/include is on the include path (or, ideally, ensure that pkg-config reports the correct paths when asked) cc.exe: error: build-windows-framebuffer/caret_image.c: No such file or directory cc.exe: fatal error: no input files compilation terminated. make: *** [build-windows-framebuffer/build-windows-framebuffer_caret_image.o] Error 1 Again, please do the following: 1) make TARGET=framebuffer clean 2) make TARGET=framebuffer Q= buildlog-fb 21 3) Send the verbatim buildlog-fb to me John-Mark.
Re: netsurf-sdl on Windows?
On Mon, 2012-05-07 at 15:32 -0400, LM wrote: Tried out netsurf-sdl on Debian and was attempting to figure out if something similar could be built on Windows. Why not use the native Windows frontend? The framebuffer frontend is totally unsuitable for use on systems that have a full-featured window manager. I attempted to build netsurf-2.8-full-src.tar.gz, but hit some issues. What issues? Please provide the exact build output. I invoked the make file with: make target=windows This builds the native Windows frontend, not the framebuffer. If you really want to build the framebuffer frontend, then you want make TARGET=framebuffer. It seems to build several object files in the build-windows-windows directory, but it never creates an executable. Again, please provide the exact build output. Without it, there's nothing we can do to help. John-Mark.
Re: Netsurf Running Out of Memory
On Mon, 2012-01-02 at 18:08 +, Dave Symes wrote: Does Netsurf log any private details, like accounts numbers and passwords? Not deliberately. If in doubt, search the log file for likely things. Alternatively, if the log file contains anything like Harriet quoted, there's no need to send it to me, as the issue is likely fixed in r13365. John-Mark.
Re: Netsurf Running Out of Memory
On Sat, 2011-12-31 at 12:36 +, Dave Symes wrote: On 31 Dec, d...@triffid.co.uk wrote: [Snippy] How interesting Brian, I was about to write up a note and post it when your posting arrived... We had the very same problem with LloydsTSB a few day ago, so we updated to version r13351 (27 Dec 2011) but the memory problem persists. That's on SARPCs one with Select 4.39 and this one with 6.20. Anyone else had the return of this old problem? Dave Should have mentioned, I still have the log fron a LTSB visit earlier today. Please send it to me. John-Mark.
Re: Fetching.BadType
On Thu, 2011-12-22 at 23:18 +, Harriet Bazley wrote: On 22 Dec 2011 as I do recall, Brian Bailey wrote: From time to time, when accessing a website, I see Fetching.BadType. What is the significance of this information, please? I believe it indicates an image which, when fetched, turns out not to be the type the server declared it to be (e.g. a GIF served with an image type of JPEG). No. It simply means that an object being fetched is unacceptable for some reason. Usually, this occurs when a site serves an object in a format that NetSurf does not support, or if it tries to serve a CSS stylesheet, say, in place of an image. It may also occur if NetSurf is completely unable to compute a type for the object, though this is a much rarer scenario. Certain Windows browsers ignore the image type information and only look at the filename Perhaps surprisingly, the filename is about the one thing they don't look at. Instead, in certain circumstances, they inspect the data served to them for signatures of known types. NetSurf does this, too. Hence why, a GIF served as a JPEG is still treated as a GIF by NetSurf. John-Mark.
Re: NetSurf on ARM9 Fedora based distribution?
On Wed, 2011-12-21 at 00:12 +0100, Karsten Jeppesen wrote: On 12/19/2011 13:14, John-Mark Bell wrote: On Sat, 2011-12-17 at 16:35 +0100, Karsten Jeppesen wrote: My first inclination was to export the trunk and then form 2 rpms. NetSurfLib and NetSurf itself. It's not clear why you'd want to install the libraries, but not the browser, unless you're building shared libraries, in which case, a package per library would make more sense. O - thats easy to explain. Netsurf itself will look for libraries and headers (its own or dependencies oc) amongst others, in the root file system (/usr/include, /usr/lib etc) but in a cross compiler that doesnt work that way. It will look in a number of places, depending on the parameters passed to make. As I said before, however, building the GTK frontend is entirely dependent upon pkg-config. So you need to get dependencies compiled first and then installed into the cross compiler before continuing. Yes, and this is what we do for all of the platforms for which we cross-compile NetSurf. At the moment I think I will end up with 3 rpms - a commonlib, a lib and netsurf itself. Injecting commonlib into the cross in order to compile lib and so on. Up to you, I guess. Unless you're intending to ship the libraries, I don't understand why you don't just build them into a prefix, and tell NetSurf's buildsystem to look in there. However I am hitting a few kinks which I believe is enough to start asking these questions. Amongst others I am hit with the problem that gcc in a full cross will refuse dumping the specs which in turn confuses NetSurfs makefile so it wont find the libnsgif simply because it cant figure out what compiler is in play. This doesn't make much sense to me. NetSurf's buildsystem doesn't use -dumpspecs to identify the compiler, or determine where libraries are installed. It does rely on pkg-config, however, and expects that the environment has been configured to allow pkg-config to find the libraries you have installed. Well - it actually does. In order to determine if it is gcc or some other compiler. Makefile.tools. Grep for dumpspecs and you will see. As I said below, this is the buildsystem used for the libraries. NetSurf itself uses a different one. Regardless, please provide full build logs of your progress so far. At present, it's not clear to me that you have built the libraries that NetSurf depends upon, let alone the browser itself. The buildsystem used for the libraries does use -dumpspecs, but falls back to parsing the output of $(CC) --version if that fails (which it shouldn't, even with a full cross) Shouldn't != doesn't. Gcc 4.4.2 does not. Which kinda sucks, but thats the compiler I am locked too for now. Please provide the verbatim output of both gcc -dumpspecs and gcc --version for the toolchain you're using. If you have successfully built and installed the libraries into a prefix, then you can build NetSurf as follows: export PKG_CONFIG_PATH=/path/to/install/prefix/lib/pkgconfig: $(PKG_CONFIG_PATH) make TARGET=gtk \ CC=/path/to/cross/bin/gcc \ PREFIX=/path/to/install/prefix He he - have you any idea how problematic libtools and pkgconfig are to dance with in a cross environment? They really aren't that complex. For pkg-config, export PKG_CONFIG_LIBDIR with the path to your cross environment's lib/pkgconfig directory. Well - CC always points to the compiler. Which of course contains both the native and cross. I don't follow. CC contains the path to the compiler binary, not the bin directory in which the toolchain binaries are stored. I will see if I can get pkgconfig to work, but its really a drag. Consider in a crossbuild that like libtools this program is built and running on the same host but collecting information on the target but is not supposed to target the target rather the host. For this purpose libtool takes the ARCH as option, but as far as I can see no such luck with pkg_config. All pkg-config does is parse text files, so as long as you have a pkg-config binary that runs on the machine on which you're compiling things, it's fine -- just point it at the correct pkgconfig directory, as I describe above. - oo - and remember to state that automake ver 1.1 if you want it to compile. Otherwise you will get some strange errors, NetSurf's buildsystem does not use autotools at all, so I don't understand what it is that requires automake. Somehow its involved or I wouldn't write it. The full build log would help here, too. John-Mark.
Re: Problem launching NetSurf on RPCEmu/Win7
On Wed, 2011-12-21 at 15:18 +, george wrote: I‘m running a copy of RPCEmu originally from one of ROOL‘s London Show memory sticks on a Windows 7 machine (Dell XPS, 3.4GHz Core i7, 8MB RAM). It runs very well, averaging around 300 MIPS, but I find NetSurf 2.8 for RISC OS (installed from a fresh download) will not launch. I‘ve updated !Boot and !System, but clicking on the NetSurf icon produces an endlessly cycling hourglass. RPCEmu is not yet bridged to the web, but I assume this isn‘t necessary in order to use NetSurf as (for example) an HTML viewer? Unfortunately, you've not given us enough to go on to have any chance of helping you. Please provide a log file from an attempt to start NetSurf on your system. Thanks, John-Mark.
Re: NetSurf on ARM9 Fedora based distribution?
On Sat, 2011-12-17 at 16:35 +0100, Karsten Jeppesen wrote: Hi there, Im new to this project, so bear with me if Im barking up the wrong tree. I have made a Fedora based distribution to an ARM9263 processor including using the rpm package system. My tool chain includes a full gcc4.4.2 cross. Is there any documentation on how to proceed in the porting? I don't think so, no. My first inclination was to export the trunk and then form 2 rpms. NetSurfLib and NetSurf itself. It's not clear why you'd want to install the libraries, but not the browser, unless you're building shared libraries, in which case, a package per library would make more sense. However I am hitting a few kinks which I believe is enough to start asking these questions. Amongst others I am hit with the problem that gcc in a full cross will refuse dumping the specs which in turn confuses NetSurfs makefile so it wont find the libnsgif simply because it cant figure out what compiler is in play. This doesn't make much sense to me. NetSurf's buildsystem doesn't use -dumpspecs to identify the compiler, or determine where libraries are installed. It does rely on pkg-config, however, and expects that the environment has been configured to allow pkg-config to find the libraries you have installed. The buildsystem used for the libraries does use -dumpspecs, but falls back to parsing the output of $(CC) --version if that fails (which it shouldn't, even with a full cross) If you have successfully built and installed the libraries into a prefix, then you can build NetSurf as follows: export PKG_CONFIG_PATH=/path/to/install/prefix/lib/pkgconfig: $(PKG_CONFIG_PATH) make TARGET=gtk \ CC=/path/to/cross/bin/gcc \ PREFIX=/path/to/install/prefix So: - where to look for a porting procedure? See above. - what libs and parts from the trunk are needed for the netsurf build? Read Docs/BUILDING-GTK in the NetSurf source tree. The PACKAGING-GTK document will probably be of some interest, too. - oo - and remember to state that automake ver 1.1 if you want it to compile. Otherwise you will get some strange errors, NetSurf's buildsystem does not use autotools at all, so I don't understand what it is that requires automake. John-Mark.
Re: accented characters in Google translate
On Mon, 2011-11-28 at 19:25 +0100, rick...@argonet.co.uk wrote: There is a problem with accented characters in some webpage text areas. (and as this note contains accented characters I fear it may not be readable by all mail clients) These two links demonstrate the problem. Link (a) accepts accents, link (b) does not. (a) http://babelfish.yahoo.com/translate_txt (b) http://translate.google.com/ You can see the behavior by:- setting the language pairs to Spanish/English, and pasting the following text SILLÓN: respuesta positiva de Yoko Ono a John. then clicking Translate. For me, at any rate, the Google page replaces the accented character by a question mark. If I replace the accented ø by O Google manages to read it. The Yahoo page works without complaint. I think this is a NetSurf feature rather than Google's as on Linux Firefox Google accepts the accents. And FWIW NetSurf has always behaved this way. No. This is a bug in Google Translate. It serves the page to NetSurf in the ISO-8859-1 character set. NetSurf then submits your data in this character set, encoding the O-acute correctly. This then fails to be interpreted correctly by the server. The reason it works in Firefox, is that Google serve the page in the UTF-8 character set to that browser. John-Mark.
Re: Error converting PNG
On Sun, 2011-11-20 at 22:10 +, Dave Higton wrote: Every picture I view from the Railway Herald site's Imaging Centre causes Error converting PNG to flash briefly on the bottom line, but only when the image is viewed for the first time - refreshing the image, or going back to one recently visited, doesn't show the message. The main image is always displayed successfully, though. Example URL: http://www.railwayherald.com/imaging.centre/showimage.php?image=214023gallery=X3 Any explanation? At least one of the graphics on that page is not a commonly understood image format. The website serves it as a PNG, hence the message you see. John-Mark.
Re: Problems with NS and Orpheus Squirrel Mail
On Sun, 2011-11-06 at 21:33 +, Steve Fryatt wrote: On 6 Nov, Chris Young wrote in message out-4eb6f630.md-1.4.17.chris.yo...@unsatisfactorysoftware.co.uk: On Sun, 6 Nov 2011 19:14:14 - (UTC), Gerald Dodson wrote: Has no one else had the problems I have experienced since 2.8 using Orpheus Sqirrel Mail? I have SquirrelMail set up on my domain which I use occasionally. I just tried it from NetSurf and it tells me I'm not logged in, so I can't actually get as far as replicating your problems. Yes, I can confirm this too, having just tried. Trying to access Squirrel Mail on a local IP address (https://192..) just tells me that I'm not logged in. Please set suppress_curl_debug to 0 in your choices file, do whatever it is that provokes this issue, then send me the log file. John-Mark.
Re: is this a bug or a site source error?
On Sat, 2011-11-05 at 21:13 +, Dave Higton wrote: Please try: http://opensource.com/life/11/11/drm-graveyard-brief-history-digital-rights-management-music (I hope I've copied that out correctly; I can't seem to drag it out of NS.) I get a warning from NetSurf: Error while processing content unencoding: invalid code lengths set Is this a bug in NetSurf, or is it an error in the site's source code? If it's a bug, I will happily report it. The site responds with a gzip-encoded stream containing trailing garbage. This causes the HTTP library we use to complain with the message you see. r13129 contains an egregious hack around the problem which should make it work as well as can be expected. John-Mark.
Re: is this a bug or a site source error?
On Sat, 2011-11-05 at 22:45 +, Tony Moore wrote: On 5 Nov 2011, John-Mark Bell j...@netsurf-browser.org wrote: On Sat, 2011-11-05 at 21:13 +, Dave Higton wrote: [snip] http://opensource.com/life/11/11/drm-graveyard-brief-history-digital-rights-management-music [snip] The site responds with a gzip-encoded stream containing trailing garbage. This causes the HTTP library we use to complain with the message you see. In that case, why does r13077 have no problem with the site? Most likely because you're fetching it through a proxy server which fixes the brokenness. Certainly, there's no reason for r13077 to work any better than later versions in this respect. John-Mark.
Re: puzzling email
On Mon, 2011-10-17 at 21:56 +0100, Brian Bailey wrote: I ran the html file which was with the message, which appears as an attchment, in !Pluto, in NetSurf. The BMP file was also an attachment to the message. NetSurf tried to load the file. End of. NetSurf has full support for BMP images. For us to be able to debug this issue, we need the following: 1) What filetype does Pluto give the BMP file? 2) Please email me the log file from NetSurf's attempt to load the HTML document in question 3) Please email me both the HTML and BMP files in question Without _all_ of the above, there is precisely nothing we can do to help. Thanks, John-Mark.
Re: Loading local copy of Builds page
On Tue, 2011-10-04 at 14:00 +0200, John Williams wrote: Logfile sent to J-MB Please do not assume that it's appropriate to send a log file directly to me for every issue you encounter. If a log file is required, then you will be asked for one by the person best placed to deal with the problem. Invariably, this is not me! Alternatively, report the problem on the bug tracker and attach the log file there. Thanks, John-Mark.
Re: Warning from NetSurf
On Wed, 2011-09-07 at 23:50 +0200, John Williams wrote: In article 1315431420.8060.5.camel@duiker, John-Mark Bell j...@netsurf-browser.org wrote: By running NetSurf, doing the thing you want, and then quitting it. The log file will be in Wimp$ScrapDir.WWW.NetSurf, as always. Log file sent. Fixed in r12778. John-Mark.
Re: visited link made visible
On Thu, 2011-09-08 at 01:36 +0100, Jim Nagel wrote: Now I'm curious: what mechanism is involved, and why is it difficult for Netsurf? It's not difficult -- we've had an implementation in the sources for at least two years. The trouble is, it's disabled because it increases page load times significantly. Enabling it is dependent on two things: 1) Changing the way NetSurf handles URLs internally 2) libdom Don't expect either of those changes to happen soon. Is it perhaps that CSS lacks the feature and overrides the earlier system which catered for visited links? No. CSS has support for visited links. John-Mark.
Re: Warning from NetSurf
On Wed, 2011-09-07 at 22:04 +0200, John Williams wrote: Unfortunately this warning box resists the efforts of the NoError utility module to close it, and the content doesn't appear in SysLog's WIMP log like it might with other multitasking programs. It would: NetSurf doesn't use Wimp_ReportError for those windows. Really, I should improve my script to avoid the empty window being generated, but I am curious as to what, if anything, this warning message actually means and as to how helpful it is to have it appear in this indecipherable form? Without a log file, there's nothing we can tell you. John-Mark.
Re: Netsurf won't get into the Woman's hour site
On Tue, 2011-08-23 at 07:28 +0100, Michael Bell wrote: I am using NetSurf 2.7 - Is this the latest? It is the most recent stable release, yes. It won't get into any of the Woman's hour sites, it downloads forever. It seems to behave normally on other sites. Fixed in r12644. John-Mark.
Re: HTTP auth and NetSurf
On Wed, 2011-06-29 at 18:43 +0100, Ste (PlusNet) wrote: Hi all, When I go to view a page that's on an HTTPS site which also uses the Apache auth stuff (meaning I need to enter a username and password) many sites require me to log in multiple times with the same credentials just to view a single page (e.g. one place I get no less than eight challenges for Site Authentication) presumably for fetching related objects (CSS, images, etc). Please set suppress_curl_debug to 0 in your NetSurf choices file, run NetSurf and access the site in question, quit NetSurf, then send me the resulting NetSurf log file (in Wimp$ScrapDir.WWW.NetSurf). Thanks, John-Mark.
Re: default form submission by keypress Return
On Wed, 2011-05-11 at 14:55 +0100, Mike Hobbs wrote: When page content includes something like this: form method=POST action=some.cgi ... input input type=submit value=Go ... input type=submit value=Clear /form and the user presses Return, every other browser will send the post request with the various arguments from the input elements (i.e. they default to doing the submit as if the user clicked the Go button), but NS seems to default to actioning the Clear button. The example you give above will submit nothing for either submit button, as neither have name attributes. If they do have names, however, NetSurf will pick the first one it sees, which is Go, if the above example is modified so the submit buttons are given names. Do you have a real example that displays the problem you describe, given I'm unable to reproduce it? John.
Re: Building Netsurf on arm-linux framebuffer
On Mon, Apr 18, 2011 at 10:56:11AM +0300, Sertac TULLUK wrote: Dear Netsurf developers, Firstly, I want to thank you for making such modular web browser. I have an embedded device with arm926EJ-s chipset, which is using simple linux framebuffer. And I want to port Netsurf to my device. I passed many sequences, but I finally stuck on compiling libnsfb with arm-linux-gcc. When I compile it with my GCC, it compiles perfectly. but when I compile it with my cross-tool: arm-linux gcc, I get following error: root@ubuntu:/home/stulluk/dev/MRH-1202/MRH-1201/netsurf/libnsfb# make COMPILE: src/libnsfb.c cc1: error: unrecognized command line option -Wno-overlength-strings make: *** [build-Linux-Linux-release-lib-static/src_libnsfb.o] Error 1 root@ubuntu:/home/stulluk/dev/MRH-1202/MRH-1201/netsurf/libnsfb# Which version of GCC are you using? And then, I disable with #-Wno-overlength-strings, but this time I get another error as below: root@ubuntu:/home/stulluk/dev/MRH-1202/MRH-1201/netsurf/libnsfb# make COMPILE: src/libnsfb.c COMPILE: src/cursor.c COMPILE: src/plot/api.c COMPILE: src/plot/util.c COMPILE: src/plot/generic.c COMPILE: src/plot/32bpp.c src/plot/32bpp.c: In function `get_xy_loc': src/plot/32bpp.c:26: warning: cast increases required alignment of target type make: *** [build-Linux-Linux-release-lib-static/src_plot_32bpp.o] Error 1 Replacing the (uint32_t *) with (void *) on line 26 of src/plot/32bpp.c should prevent that warning being generated. If that is the case, let us know and we'll fix that in SVN. HTH, John.
Re: logging
On Wed, 2011-03-30 at 09:17 +0100, Brian Bailey wrote: Hi John [snip] Ok. Have you tried Vince's suggestion? Thanks for reminding me, Mark. I tried that suggestion some time ago but was unable to draw any conclusion. However, I will give it a longer test now. I have being trying that suggestion since your email and there does seem to be an improvement. However, I am now using vv r12147 and note the SNV comments, but also note that -v remains in the !Run file, It would; r12147 contained changes to the Windows frontend only. which shows Revision: 11204. Is that an anomaly or is that line inert, That is the last revision in which the !Run file changed. John.
Re: logging
On Tue, 2011-03-22 at 09:00 +, Brian Bailey wrote: In article 20110321163550.gg4...@rjek.com, Rob Kendrick r...@netsurf-browser.org wrote: Remove the -v flag. That is what enables verbose logging. Note that this simply stops NetSurf formatting and emitting the log data. It will still involve checking do I need to log this? Thanks, Rob. I also removed the -v flag, without putting an | anywhere. Commenting out bits of the !Run file using | is pretty much guaranteed to end in tears. Please don't do this. Additionally, disabling logging totally defeats the point of us asking for log files when you report a bug. NetSurf seemed to be entirely stable in both cases. As far as I could tell, there was no discernible, subjective change in time in loading and rendering files, in either case. It would be; logging simply isn't a significant proportion of runtime. NetSurf spends the vast majority of its time doing the things you'd expect (namely, fetching and rendering web pages). I guess that I must accept that my machine /is/ rather slow. But, as I am retired I can make endless cups of tea whilst things happen. 8-) I'm not sure you've told us what your machine is or what led you to believe that turning off logging would help. John.
Re: Bug Out of date
On Tue, 2011-03-08 at 08:30 +, David J. Ruck wrote: What is the reasoning behind marking a bug out of date? Particularly when it is easily reproducible with the last test build by one click on the given URL. Simple: it was 4 years old and I couldn't reproduce it (and still can't fwiw) John.
Re: Bug report r11881
On Wed, 2011-03-02 at 16:09 +, Brian Bailey wrote: I dunno, I couldn't get the bug report page to work for me, I'll probably get shouted at for doing things this way! 8-) Actually, bits of that page I couldn't understand. Clicking a photo image line in URL http://www.steamtraingallerries.latest_photos.html I get error message 'serious error' with a complete crash, using RISC OS 4.02 on an A7000+ Error Log scrap file attached. Unfortunately, the error log you attached is not from the crash, so we have no useful information to go on. John.
Re: crash on trying to get to textfile or to enter password
On Wed, 2011-03-02 at 19:47 +, Jim Nagel wrote: i filed a bug report when clicking this link crashed Netsurf earlier this afternoon: http://archivemag.co.uk/JN/2212/test.txt (the link works fine with Fresco or XP Firefox.) This has already been fixed. Please update to a newer version. similar thing happened just now in trying to download new version of Moredesk from http://www.7thsoftware.co.uk/moredesk.releases -- it asks for a username and password, which i enter, whereupon Netsurf crashes. Fresco copes OK. You've not given us anything to go on here -- it crashes isn't a particularly useful bug report. Do you have a log fix from the crash? John.
Re: font face=monospace in HTML
On Sun, 2011-02-13 at 17:03 +, Richard Porter wrote: On 13 Feb 2011 Dave Higton wrote: My experiments yesterday and today show that Netsurf ignores font attributes within HTML, e.g. font face=monospace (and all the other generic face types) are ignored, whereas the equivalent within a CSS is obeyed. Is this a deliberate design decision? No. It's a bug. I think this is the usual problem that priority has been given to CSS over plain html and anything that is deprecated even if it's in widespread use. The size and color attributes are supported though. This is not the case. There is a defined place in the CSS cascade for these legacy presentational hints. However, until we replaced NetSurf's original CSS implementation with libcss, there was no mechanism for allowing the presentational hints to be used when determining the styling of an element. What support did exist for them at the time mostly comprised layers of ugly hackery to approximate the right thing. Invariably, this went wrong. Today, however, we do have the appropriate hooks in place. Thus supporting all of these legacy presentational hints is relatively trivial and totally opaque to most of the browser. If someone feels adventurous, they could compare section 10.2 of the HTML5 specification (specifically, anything where presentational hint is mentioned) with the node_presentational_hint function in css/select.c in NetSurf's sources to see what's missing. John.
Re: font face=monospace in HTML
On Mon, 2011-02-14 at 21:58 +, Dave Higton wrote: In message 1297713411.8764.20.camel@duiker John-Mark Bell j...@netsurf-browser.org wrote: There is a defined place in the CSS cascade for these legacy presentational hints. However, until we replaced NetSurf's original CSS implementation with libcss, there was no mechanism for allowing the presentational hints to be used when determining the styling of an element. What support did exist for them at the time mostly comprised layers of ugly hackery to approximate the right thing. Invariably, this went wrong. Today, however, we do have the appropriate hooks in place. Thus supporting all of these legacy presentational hints is relatively trivial and totally opaque to most of the browser. If someone feels adventurous, they could compare section 10.2 of the HTML5 specification (specifically, anything where presentational hint is mentioned) with the node_presentational_hint function in css/select.c in NetSurf's sources to see what's missing. Thanks for your response, John. It's curious that two so similar constructions behave so differently: pfont face=monospace doesn't work; This is a legacy presentational hint -- see the explanation I gave above as to why NetSurf ignores it at present. Essentially, NetSurf is currently missing an implementation of the mapping from the face attribute of the font element to the corresponding CSS font-family representation. p style=font-family:monospace does work, The style attribute contains CSS declarations for the element. These are not treated in the same way as presentational hints, as they have differing requirements. They work in NetSurf as support for them simply falls out of having any kind of CSS implementation at all. John.
Re: NetSurf crashes on Blogger images
On Tue, 2011-01-11 at 19:18 +, Martin Bazley wrote: It seems that whenever NetSurf attempts to load a direct link to an image hosted on Blogger, it crashes after loading, but before displaying, the page. Thanks for the report. Fixed in r11293. John.
Iconv 0.11 released
Version 0.11 of the Iconv module is now available for download. This is a recommended update for all users - new builds of NetSurf will require it. Fixed in this version: * Detect missing mapping file when using 8bit codecs. This prevents spurious memory exhaustion errors. * Toolchain used to build 0.10 turns out to have produced broken code. Changed in this version: * Minor additions to the charset alias mapping file. About Iconv === Iconv is a module providing character set conversion facilities for third-party applications. It is currently used by a number of applications, most notably NetSurf. UnixLib's iconv support also requires this module. Updating from earlier versions == If updating from earlier versions, please perform _both_ a !Boot and !System merge with the contents of the Iconv distribution archive as documented in the Installation section of the ReadMe file. Where to get it === Iconv may be downloaded from http://www.netsurf-browser.org/projects/iconv/
Re: libnsfb issue
On Thu, 2010-11-11 at 17:07 +, Ario K wrote: Friendly Greetings, I've been trying to compile NetSurf for Linux frambuffer but the libnsfb is giving me nightmares. libnsfb seems to be a broken package with lots of files (like SDL.h or xcb files) missing. I downloaded many of the missing files but it still wants more. Currently the error message is like this: LibNSFB depends on you having installed the SDL and XCB development packages on your system. It sounds very much like you have not done so. You certainly shouldn't be downloading individual header files. My distro is Debian Squeeze. Please help me. In which case, apt-get install libsdl-dev libxcb1-dev libxcb-image0-dev \ libxcb-icccm1-dev libxcb-keysyms1-dev is probably all you need.
Re: Tinct and ROOL 2010-11-01 Beagle ROM with 16bpp modes
On Thu, 2010-11-04 at 16:34 +0100, Trevor Johnson wrote: Could someone please comment on the 'Use OS' approach? That's the only available fix. We do not have access to the Tinct sources and have given up trying to obtain them under suitable licencing terms. Therefore, if you are using NetSurf on ARMv7 platforms, configure it to use the OS for image rendering. Any bugs that are reported about this issue will end up being closed without prejudice as we're simply not in a position to fix them. John-Mark.
Re: Re:ROOL site: error setting certificate verify locations:
On Thu, 2010-09-09 at 21:32 +0100, Dave Higton wrote: So, whatever the difference is, it doesn't seem to be accounted for by any content of Netsurf's choices. I'm mystified. So are we. As I said in the original thread, we need a log file. Please provide one and we'll take a look. It's clearly a Beagleboard-specific issue. John.
Re: Downloading zip files (from mw-software)
On Thu, 2010-07-22 at 10:21 +0100, Rob Kendrick wrote: On Thu, 22 Jul 2010 10:07:03 +0100 Russell Hafter - Lists rh.li...@phone.coop wrote: Can you give us an example URL that demonstrates this? Given that in both cases I was downloading paid for software and had to enter username and password, I am not at liberty to do so. You would have to take that up with the owner. Are you running a (recent) version of NetSurf with all the logging turned on? If so, is it possible for you to email that log to one of us privately? There's no way for us to identify where the fault may lie without it. Has anyone actually done this yet? Without the appropriate log file, from an appropriately recent build (namely r10650 or later), this cannot be debugged. If you want problems fixed, please provide the information we've asked for. John.
Re: Downloading zip files (from mw-software)
On Thu, 2010-07-22 at 19:16 +0100, Russell Hafter - Lists wrote: In article 1279822036.22137.11.ca...@duiker, John-Mark Bell j...@netsurf-browser.org wrote: Are you running a (recent) version of NetSurf with all the logging turned on? If so, is it possible for you to email that log to one of us privately? There's no way for us to identify where the fault may lie without it. Has anyone actually done this yet? Without the appropriate log file, from an appropriately recent build (namely r10650 or later), this cannot be debugged. If you want problems fixed, please provide the information we've asked for. You will need to tell me where to find the log, if I am to do this. *Filer_OpenDir Wimp$ScrapDir.WWW.NetSurf John.
Re: Hard drive continuous access
On Wed, 2010-07-21 at 20:29 +0100, Bob Latham wrote: Hi, Using OS 4.39 and this happens on more than one machine. I've noticed that with recent versions of Netsurf, if a netsurf window is open then the hard drive is continuously being accessed non stop. r10640 works fine - no continuous hard disc access. r10651 hard disc continuous if a netsurf window is open. r10656 hard disc continuous if a netsurf window is open. Is this a bug or a feature or have I got something wrong? I have reverted to r10640. It's a feature until such time as we have sufficient information to debug a number of issues we cannot reproduce ourselves. Please do not revert to an earlier version, as that defeats the point of enabling the extra logging. John.
Re: Certificate verification query
Please do not top post on this mailing list. Thanks. On Wed, 2010-05-05 at 06:17 +0200, Trevor Johnson wrote: Thanks too, Steve. I meant to add that I'll be able to try rebuilding from source once I've bought the Acorn/Castle/ROOL C/C++ suite. NetSurf uses GCC, not the Acorn C tools. It's probably a BeagleBoard issue. It's far more likely to be a configuration problem. Please try moving your NetSurf configuration out of the way, then running it. If that doesn't work, then we need a log file. John.
Re: NetSurf 2.5 Released
On Sat, 2010-04-24 at 22:10 +0100, Michael Bell wrote: In message 510d269ff2t...@netsurf-browser.org Michael Drake t...@netsurf-browser.org wrote: The NetSurf developers are happy to announce the immediate availability of NetSurf 2.5. This release contains many bug fixes and improvements. It is available to download from http://www.netsurf-browser.org/ The NetSurf Developers Newly downloaded Such a bad omen! The middle section cut out. We cannot reproduce this, and the log file you provided contains insufficient information to allow us to debug this successfully. Do you still have the full log file? If so, please zip it and email it to me (off list). John.
Re: Website development updates?
On Wed, 2010-03-10 at 18:19 +, Grahame Parish wrote: Hi, Not a complaint, but I check the NetSurf website daily to see what the latest development changes have been. Is there any reason why the last change shown on the website is r9799? Yes. r9802 introduced an incompatible API change to libcss, and the autobuilder is not intelligent enough to know which version of libcss it should be building NetSurf against. Additionally, we're rather more interested in completing the new cache implementation than we are in fixing this issue. Doubly so as I've vetoed any such fix being applied to the trunk code as it will conflict badly with the new cache work. There have been very few modifications to the trunk code since the autobuilder stopped, so you're not missing out on much. I know it is the last released RO version No. The last released version of NetSurf (on all platforms) is version 2.1. Development builds are just that. We make absolutely no guarantees about their stability or, indeed, their general availability. , at least as far as RiscPkg is concerned. I have recently refreshed the version I run on Ubuntu, which appears to be at least release 10030, but there's no indication on the site of what changes have been implemented. I assume that you're compiling this yourself? As an aside on the GTK version, is there a way to see from the running executable what the release number is in the same way that the RO Info menu option does? Help-About will show you. If you compile it yourself, however, it will not show you the exact SVN revision the sources came from. This information is inserted by the autobuilder. I can't currently get a GTK version to compile, with css-related errors during compilation. I can't pass on the details at present, but I'll make a note next time I try. Please ensure that you are using revision 9801 of libcss. NetSurf is currently unable to use anything more modern than that. John.
Re: NetSurf partly works on the BeagleBoard
On Mon, 2010-03-01 at 21:31 +, Dave Higton wrote: I don't know if anyone else has tried it, but I have... NetSurf r9799 on a BeagleBoard renders my (text only) web site! Good. We've been compiling with GCC4 for quite some time, which ensures that the code is automatically ARMv7 compatible. It occurs to me that noone's actually verified that the one bit of assembler (the Artworks binding) in the NetSurf sources is ARMv7 ready. Perhaps someone with an interest might take a look. It falls over on the BBC News site a fraction of a second after first drawing the window. I've caught a log and I'll attempt a bug report. Most likely, it's exploding in Tinct, which is not ARMv7 ready. Unfortunately, we are not in a position to fix this as we don't have access to the sources for that module. John.
Re: NetSurf crashes as soon as it is run
On Sun, 2010-01-31 at 11:09 +, Richard Mellish wrote: NetSurf was working OK until a few days ago, when it took to crashing when started. I am not aware of making any change to my system, which is an Iyonix running RISC OS 5.11. (I am intending to install 5.16 soon but am deferring that until I have copied some applications and data files to a new computer.) Delete Wimp$ScrapDir.RUfl_cache. I strongly advise upgrading to 5.16 as, aside from the date fix, it also contains a bug fixed FontManager that allows RUfl to scan fonts approximately 7 times faster. John.
RE: New to using NetSurf...Problem with compiling.
Please do not top post when emailing this mailing list. Thanks. On Fri, 2010-01-29 at 04:48 +, Rajesh Marathe wrote: On Thu, 28 Jan 2010 16:01:27 +, wrote: css/css.c:1552: error: ‘LBRACE’ undeclared (first use in this function) --- [...] Thinking back, this problem was with the OLD CSS engine, not the libcss one. Correct, and most likely because of a missing lemon or re2c. Are you compiling the latest version from SVN or one of the releases? I was suggested to use the revision 9801 of css instead of the trunk release. The trunk release (Latest on svn) also shows this problem. The version of libcss is irrelevant here, as the compilation failure you're describing is with NetSurf itself. The advice about libcss was correct, assuming that you were trying to compile the latest NetSurf sources. From the compilation failure, it would appear that you are not attempting to compile the most recent sources for NetSurf. For us to help you further, please tell us which version of NetSurf you are attempting to compile. John.
Re: Google's PDF-as-text
On Sat, 2009-11-14 at 14:42 +, Jim Nagel wrote: when the results of a Google search include a PDF, Google offers to display it as plain text. but it comes out as nonsense in Netsurf (r9629): all the lines of text seem to be piled on top of one another. Fixed in r9675. John.
Re: Netsurf compilation error
On Thu, 2009-10-15 at 19:21 +0530, Moutusi De wrote: Hi All, I am trying to build netsurf-2.1(I believe this stable release) source code in linux with gtk. I am getting following error: netsurf/css/css.c:1539: undefined reference to `css_parser_Alloc' netsurf/css/css.c:1552: undefined reference to `css_parser_' netsurf/css/css.c:1556: undefined reference to `css_tokenise' netsurf/css/css.c:1574: undefined reference to `css_parser_Free' Please ensure that you have both lemon and re2c installed, as described in Docs/BUILDING-GTK. netsurf/desktop/save_pdf/pdf_plotters.c:484: undefined reference to `HPDF_Image_AddSMask' netsurf/desktop/save_pdf/pdf_plotters.c:423: undefined reference to `HPDF_LoadJpegImageFromMem' As described in Docs/BUILDING-GTK, NetSurf requires a more recent version of libharu than the most recent upstream release. A snapshot of an appropriate development version can be found in the NetSurf SVN respository -- see Docs/BUILDING-GTK for details. Along with that I am getting undefined reference for gtk calls, like netsurf/gtk/dialogs/gtk_about.c:74: undefined reference to `gtk_about_dialog_set_url_hook' etc. Odd. That function has been in GTK since 2.6, so should be available in 2.18. John.
Re: LloydsTSB
On Thu, 2009-10-15 at 17:41 +0100, Dave Symes wrote: FWIW. I'm using the latest NetSurf and it works perfectly okay with LTSB, and I do have a number of .co.uk cookies. You should have no .co.uk cookies at all. I strongly recommend that you delete them. By rights, NetSurf should never have accepted them in the first place. Unfortunately, this is yet another case where following any of the published specifications doesn't actually get you anywhere near reality. Additionally, J-M.Bell's cookies suggestion does seem to work if you encounter the problem, as I used it yesterday to get Fays NetSurf working again with LTSB. Of course it works -- I'd not have suggested it otherwise :) John.
Re: LloydsTSB
On Thu, 2009-10-15 at 21:19 +0100, Dave Higton wrote: In message 1255628822.25071.29.ca...@duiker John-Mark Bell j...@netsurf-browser.org wrote: On Thu, 2009-10-15 at 17:41 +0100, Dave Symes wrote: FWIW. I'm using the latest NetSurf and it works perfectly okay with LTSB, and I do have a number of .co.uk cookies. You should have no .co.uk cookies at all. I strongly recommend that you delete them. By rights, NetSurf should never have accepted them in the first place. Unfortunately, this is yet another case where following any of the published specifications doesn't actually get you anywhere near reality. I have no doubt that you're right, but can you please explain why the above is so? All the specifications say that user agents should refuse to accept domain cookies that do not have a '.' in their domain. The intention of this is to avoid accepting cookies for top-level domains such as .com. Unfortunately, this falls over when the tld registrar issues top-level domains of the form .foo.bar (e.g. .co.uk). The upshot of this is that any domain cookies for .co.uk will be sent to any site under the .co.uk hierarchy. Obviously, this is completely incorrect. The solution is for the user agent to consult appropriate DNS records to ensure that the domain referenced by a domain cookie actually exists. NetSurf currently does not do this, which is the bug. I hope that clarifies things somewhat. John.
Re: LloydsTSB
On Fri, 2009-10-16 at 00:24 +0100, John-Mark Bell wrote: All the specifications say that user agents should refuse to accept domain cookies that do not have a '.' in their domain. Ugh. That's badly worded. What I meant was, ... accept domain cookies that do not have more than one '.' in their domain.. John.
Re: amending autocompletion list
On Tue, 2009-10-13 at 14:58 +0100, Jim Nagel wrote: some months ago i stupidly typed archive.co.uk into the URL field rather than archivemag.co.uk. now, foreverafter, the wrong entry is the first one on the auto-completion list that pops up when i start typing. how can i delete the wrong entry from the list? (Netsurf helpfile doesnt touch on this question.) Open the Global History viewer (^F7), find archive.co.uk, and delete it. John.
Re: amending autocompletion list
On Tue, 2009-10-13 at 15:13 +0100, David J. Ruck wrote: Jim Nagel wrote: some months ago i stupidly typed archive.co.uk into the URL field rather than archivemag.co.uk. now, foreverafter, the wrong entry is the first one on the auto-completion list that pops up when i start typing. how can i delete the wrong entry from the list? (Netsurf helpfile doesnt touch on this question.) * Close NetSurf * Go to !Boot.Choices...WWW.NetSurf * Edit the URL file and remove the offending line Following the above instructions will corrupt the visited URL database. The URL file _must not_ be edited manually unless you really know what you're doing. Please use the global history viewer to inspect and remove items, instead -- it's what it's there for. John.
Re: Problem with fbhvc site
On Tue, 2009-10-13 at 15:24 +0100, Richard Porter wrote: I've noticed a problem with the search box on the site http://www.fbhvc.co.uk/ . It appears at the top right of most pages. I can't enter any text into the box and when I click the Go button it just returns to the same page. I can't see an obvious reason though I guess it's css/js related. The search is OK on Oregano 1.10. The search box is there. NetSurf gets its position slightly wrong. You need to click *above* the search area -- this should be obvious as the pointer will change to a caret. Entering text and clicking Go then works fine. I'll await comments here before raising a bug report. Please report it, along with a full copy of the page. John.
Re: LloydsTSB
On Tue, 2009-10-13 at 19:54 +0100, Geoffrey Baxendale wrote: In message 1255458566.32582.200.ca...@duiker John-Mark Bell j...@netsurf-browser.org wrote: On Tue, 2009-10-13 at 19:17 +0100, Geoffrey Baxendale wrote: Hi, LloydsTSB has recently stopped working for me after years of trouble free operation. I don't think it's Netsurf as V2.1 doesn't work either. I suspect they have loaded their site with Javascript or it's a Browser detection problem. Having said that Organo2 works OK. Try: https://online.lloydstsb.co.uk/account.ibc or try loging on from the home page. It really is NetSurf. Please see https://sourceforge.net/tracker/?func=detailaid=2872051group_id=51719atid=464312 Thanks John, that at least rules out finger trouble at this end, but why does it not work with V2.1 if it is due to a recent change in Netsurf? It isn't due to a recent change in NetSurf. NetSurf's cookie handling hasn't changed for ages. I suspect that the LloydsTSB site has tightened up its cookie handling. However, the fact remains that NetSurf's cookie handling in this particular case is wrong and needs to be fixed. My comment on the above linked bug report tells you how to work around the issue for the timebeing. John.
Re: Tried To Build R9615
On Sun, 2009-10-11 at 16:44 +, Chris Young wrote: On Sun, 11 Oct 2009 00:15:13 +0100, Rob Kendrick wrote: On Sat, 10 Oct 2009 23:45:21 +0100 Mark Williams mark.willi...@ntlworld.com wrote: You do not have a suitable Haru PDF library installed. Either use the pre-patched one from our subversion repository (our patch has been accepted, IIRC, but they have not yet made a release containing it) or turn it off via Makefile.config (see Makefile.defaults for what key to set.) I did notice the other day that disabling PDF export still tries to call functions in libharu. There must be some #ifdefs missing somewhere, but I didn't make a note at the time. I don't see that with either the RISC OS or GTK targets here. John.
Re: Libharu Subversion My Foot
On Sun, 2009-10-11 at 17:30 +0100, Mark Williams wrote: I can't find it anywhere. No matter how deeply I dig in Google. You have read Docs/BUILDING-GTK, I take it? If you have, you'll have encountered the following section, which I hope is self-explanatory: Libhpdf - NetSurf can use Haru PDF to enable PDF export and printing in GTK. This is currently enabled by default, and cannot be auto-detected by the Makefile. If you wish to disable it, do so by creating a Makefile.config file. Haru PDF can be obtained from http://libharu.org/, although we currently depend on features that none of the official released versions does have. The current development versions of libharu are fine and we anticipate the libharu 2.2 release will be fine for NetSurf usage. A recently taken snapshot of one of those libharu development versions can be found at: svn://svn.netsurf-browser.org/trunk/libharu John.
Re: Global history in RISC OS
On Tue, 2009-10-06 at 14:31 +0100, Dr Peter Young wrote: Please, where does RISC OS NetSurf keep its global history, and in what form? I have searched in vain so far. Choices:WWW.NetSurf.URL I've been able to search the hotlist by loading the HTML in Boot.Choices.WWW.NetSurf and then using NetSurf's search facility, and there are occasions in which it would be useful to be able to do this with the global history, if that's possible. The database of visited URLs is not in HTML format. John.
Re: BBC weather site using hige amounts of memory.
On Tue, 2009-09-29 at 19:14 +0100, Michael Drake wrote: In article cd3479a250.pnyo...@pnyoung.ormail.co.uk, Dr Peter Young pnyo...@ormail.co.uk wrote: I imagine that this is the fault of what seems to be a thoroughly bloated site rather than of NetSurf, but I'd like an expert opinion, particularly as to whether I should submit a bug report. Loading http://news.bbc.co.uk/weather/forecast/2174?area=GL52 We have known cache problems at the moment, which causes NetSurf to eat up a lot of memory. In any case, a bug report with sufficient details never does any harm and can often do some good. :) In this particular case, there is little point in reporting it on the bug tracker. We already know what the problem is, even if the fix is decidedly non-trivial. More tickets on the tracker just adds to our workload when triaging them. If, however, you encounter NetSurf warning about a lack of memory when it's using very little (i.e. its main dynamic area is smaller than 128MB), then we do want such reports please. John.
Re: Google Summer of Code Roundup
On Sun, 2009-09-20 at 16:10 +0200, Xavier Tardy wrote: I think it's important to keep alive important projects like the Netsurf browser for RISC OS. Then please do something about it. Of course if this is not even seen as important by the Netsurf team, then let's forget it. I have explained the NetSurf team's position multiple times in this thread alone. Please go and read those emails, instead of making assumptions about what we do or do not believe. It's likely the case that more development time has been wasted by this thread than would have been needed to make the RISC OS frontend work with the pending core changes. Oh, and I'm *still* waiting for someone who isn't a member of the NetSurf development team to thank our GSoC students for their hard work this summer. John.
Re: crash at PN's survey / big logfile vs bug report
On Mon, 2009-09-21 at 00:08 +0100, Jim Nagel wrote: and its crashing takes a further couple of minutes. 9-megabyte logfile results, with message asking me to submit a bug report and attach logfile. but 9M is far larger than the maximum you can send with a bug report. what should i do? Zip it and send it to me. John.
Re: Google Summer of Code Roundup
On Thu, 2009-09-17 at 00:08 +0100, aw29...@gmail.com wrote: In message 1253105573.5804.236.ca...@duiker John-Mark Bell j...@netsurf-browser.org wrote: That doesn't solve the problem -- the RISC OS frontend has no active maintainer. I've been taking sometime out and modifying the RISC OS frontend for the last 2.5 years. I no longer have time to do that. You said the RISC OS frontend has had almost no development for the last 2.5 years. Now you're saying you've been modifying it for 2.5 years. These two statements are not incompatible. Given the tens of thousands of lines of code I've written for NetSurf in the relevant timespan, the odd change here and there to stop the RISC OS frontend from breaking as a result of other changes is utterly insignificant. It also does not constitute development of the RISC OS frontend -- just preserving the status quo. Things have now reached the stage where major changes are beginning to be made to the rest of NetSurf that require active maintenance of the frontends (and the time commitment that implies). The RISC OS frontend does not have active maintenance, which is the problem. Do you have time to do this or is something else more appealing about the other platforms? Why should I have to make time to do anything? As I explained to you on Drobe some weeks ago, the amount of time I have available for NetSurf will reduce somewhat in the future due to other commitments. Therefore, I have to decide where the time I have available for NetSurf is best spent and focus upon those things. In practice, this means that I'll be focussing on the core engine. The RISC OS frontend loses out and does not receive the minimal maintenance it has had for the last 2.5 years. This has nothing to do with NetSurf's availability for other platforms -- if the Amiga frontend (for example) was not maintained, we'd be having the same discussion about that. Fortunately for Amiga 4 users, Chris is very effective at fixing issues as they come to light. I don't think that it's too much to ask that someone, or a group of people, come forward and do the same for the RISC OS frontend. John.
Re: Google Summer of Code Roundup
On Thu, 2009-09-17 at 00:15 +0100, Steve Fryatt wrote: I think the main problem here is that the majority seem happy to sit back, wring their hands and state how terrible this all is -- then wait for someone else to step forward to take on the work (or even demand that someone else does). As such, it seems to have been the usual suspects who have offered to help: those already doing other things. Certainly if I picked up NetSurf now, something else would have to give, which in all probability would just move the complaints from this list to another RISC OS forum a month or so down the line. Quite. This has been RISC OS' elephant in the room for the last 5 years or so -- there simply aren't enough active developers to do all the things that need to be done. I don't have a solution to this problem, but do think it's about time people acknowledged that it exists and review their expectations in that light. Returning to NetSurf, we really do appreciate the offers of help that we've received so far and fully understand that real life gets in the way :) John.