Re: Please test the latest build

[Chris Young]

Hi Michael

AmigaOS 4 build 5380 appears to be working well.

Regards
Chris


On Thu, 3 Nov 2022 at 10:09, Michael Drake  wrote:

> Hello,
>
> Please could you test the latest builds (5377 or later) from:
>
>   https://ci.netsurf-browser.org/builds/
>
> Particularly please test the Amiga OS4 and RISC OS builds as
> we have updated the versions of 3rd party libraries that we
> build against. Amiga OS4 builds have had about 3 years worth
> of updates and RISC OS builds about 1.5 years.
>
> In addition we have patched the version of libcurl we build
> against which should make HTTPS connections much faster for
> the RISC OS and Amiga OS4 builds.
>
> Best regards,
>
> --
> Michael Drake https://www.codethink.co.uk/
> ___
> netsurf-users mailing list -- netsurf-users@netsurf-browser.org
> To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org
>
___
netsurf-users mailing list -- netsurf-users@netsurf-browser.org
To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org

Re: Possible to get it out of reverse video?

[Chris Young]

>
>
> I'm struggling to post to this list, so hopefully this will work.
>
> Harriet Bazley wrote:
> > Unfortunately I'm pretty sure Netsurf doesn't offer any mechanism to
> > force HTML pages to change their colour scheme.  It doesn't offer all
> > that many customisations anyway (you can't even tell it not to load
> > images, so far as I'm aware).   I think it was mainly written to be
> > small and fast, rather than highly flexible.
>
> There is a media query which allows a web browser to pick up the preferred
> colour scheme (light mode/dark mode) from the OS (or the browser itself).
> NetSurf does not implement this, so everything defaults to light mode.
>
> That would help with sites which support dark mode (via the media query or
> otherwise), however still leaves the rest of the web in light mode.
>
> You can change the default CSS but it is only going to help if the page
> itself doesn't specify colours.
>
> If the page specifies colours then they will be used by NetSurf - there's
> no way of overriding them (beyond light mode/dark mode - and even then
> you're still using colours specified by the page).  I'm not sure how you'd
> even go about implementing this.
>
___
netsurf-users mailing list -- netsurf-users@netsurf-browser.org
To unsubscribe send an email to netsurf-users-le...@netsurf-browser.org

Re: Please test webp image format

[Chris Young]

On 07/12/2019 13:49, Michael Drake wrote:

Hello,

Recently we added the webp library to our SDK. NetSurf Builds from
our CI should now have webp support.

So far we've only tested on Linux. Please could users of other
platforms visit

https://developers.google.com/speed/webp/gallery1

and let us know if it's working?



Not working on AmigaOS 4/PPC, just fails to recognise them at all (once 
I'd remembered to disable my old WebP datatype).



Chris

Re: date in search results

[Chris Young]

On 16 February 2018 12:54:04 GMT+00:00, Jim Nagel  
wrote:
>What causes Google (or other searchers) to display (or not display) a 
>DATE in search results?
>
>For instance, search for "Terran site:iconbar.com" -- some of the 
>items listed show a date and some don't.
>

Do a search for "Google structured data".

Chris

Re: Experimental treeview search feature

[Chris Young]

Hi Michael

On Mon, 18 Sep 2017 23:29:47 +0100, Michael Drake wrote:

> It would be helpful if people could put it thoroughly through its
> paces.  Any feedback gratefully received.

I've finally had a chance to test it out here (Amiga frontend).  It is
working well, except for a couple of minor issues;

1. When there are a lot of search results, the treeview is scrolling
to the end of the list.  This is confusing because it means the search
string is no longer in view.  I think it should be remaining at the
top.

If you have enough entries to fill up the window you can test this out
easily by typing a single common letter into the search field.

2. If the window is closed with a search in progress, when the window
is re-opened the search string does not display but the results are
still present.  I think this is due to how I implemented the windows
as they are disposed of completely and re-created on the Amiga
frontend, whereas on others the state is remembered and re-used.

I guess the treeview cleanup code is clearing the search string but
not the underlying results?  Or forgetting the search string?  I'm not
sure what the correct thing for it to do is, but the state of the
treeview (which folders are open/closed) is normally remembered, so it
should probably be remembering the search string.

Chris

Re: Form submission timeouts

[Chris Young]

On 09/08/2017 17:21, Harriet Bazley wrote:

Over the last couple of days, Netsurf has unexpectedly started timing-out
when you attempt to send messages of more than a certain length via the
www.fanfiction.net website, either via the user forums or the PM system.
PMs have an official site limit of 8000 characters (and I've successfully
sent messages of up to this length before) and so far as I recall the forum
post size limit is 9,999 characters.

Currently I just get a time-out error after about twenty seconds if I
attempt to send messages of over 1,485 characters via a PM or a post of
more than 1,343 characters to a forum.  This is quite limiting when
attempting to reply to other users -- especially as the site rules prohibit
sending more than one post in succession!


Request Timeout

This request took too long to process, it has been timed out by the server.
If it should not be timed out, please contact the administrator of the
website to increase the 'Connection Timeout' time allotted.


I've looked at Netsurf's logs, but the failed attempts aren't showing any
error messages other than a 20-second gap.   There were site-wide problems
with the forums recently, but this particular issue doesn't seem to be
affecting anybody else.

Do these specific byte lengths suggest anything to anyone?



Maybe similar to http://bugs.netsurf-browser.org/mantis/view.php?id=2252 ?
Set suppress_curl_debug:0 in Choices and have another look at the log file.

Chris

Re: UTF-8 processing

[Chris Young]

On 14 July 2017 12:10:47 BST, Bernard Boase  wrote:
>Looking in detail at some recent HTML email attachments (received in 
>Messenger Pro), Netsurf's rendering of them seems to have a dependence 
>on the coding of the  tag.
>
>When this tag is present and includes:
>
> content="text/html charset=utf-8"
>
>any non-ASCII characters are shown as the three bytes of their UTF-8 
>encoding, whereas if the two attributes are separated by semicolon:
>
> content="text/html; charset=utf-8"
>
>the rendering (using Unicode font DejaVu) looks correct.
>
>Is this a known problem?

I don't know if it's a known problem, but certainly the first form is 
incorrect. The semicolon is required to separate the parameters.

If NetSurf doesn't know the encoding it will assume ASCII/ISO-8859-1, or maybe 
default character encoding of the OS.

Chris

Re: github crash

[Chris Young]

On Sun, 12 Feb 2017 13:48:31 GMT, Dave Higton wrote:

> In message <000646be.01f994909...@smtp.freeola.net>
>   Peter Slegg  wrote:
> 
> >The Atari Freemint OS moved to github and they posted this link to
> >an Issue:
> >
> >https://github.com/freemint/freemint/pull/20
> >
> >When I open it with NS, it downloads about 14k in about 3-4mins but
> >then shows the error below and crashes out.

This is an odd one.  I suspect that error is due to NetSurf processing
an incomplete file, as the slow speed suggests it may not have
downloaded fully (NetSurf should recognise this and you'd get a
timeout error, so not sure why that isn't happening)

I tested here on the closest thing I have to the Atari build - the
Amiga version built for OS3/68k.  This also has problems, it gets as
far as fetching, processing and then the entire computer locks up.
The Amiga build for OS4/PPC works fine.  This suggests that there may
be some 68k-specific problem in NetSurf or one of the libraries we
use.

I've been playing around with a 68k optimised version of OpenSSL
and I get slightly different behaviour with the non-optimised version.
This may point to OpenSSL being the problem, or it could be an
illegal memory access somewhere where the slight change in where code
is located is enough to make it crash rather than freeze. I don't know
how to go about debugging this.

> I just tried it on CI#3984 on RISC OS.  The page loads reasonably
> quickly and there's no crash.  The only problem I see is that the
> text within  tags is rendered as black rectangles,
> with no visible or selectable text.

On OS4 I'm getting dark grey text (probably a quirk of the font
engine) on a black background, so the text is there but the background
is the wrong colour.  Looking at Firefox the background should be
light grey - my suspicion is that the light grey is being created by
using the alpha channel, which I believe NetSurf ignores for
rendering.

Chris

Re: Javascript ON when it's OFF

[Chris Young]

On 16 December 2016 22:49:06 GMT+00:00, Richard Porter  
wrote:

>Have you got a test page that tells you if javascript is on or off 
>e.g. Javascript is NOT enabled. ?

http://www.javatester.org/javascript.html

Chris

Re: history and hotlist not saved; CSS site

[Chris Young]

On 25 November 2016 09:57:04 GMT+00:00, Jim Nagel  
wrote:

>I'll put in a formal feature suggestion (via Sourceforge?) 

http://bugs.netsurf-browser.org

Chris

Re: Netsurf on Amiga 68k

[Chris Young]

> 1) It keeps using he chip RAM to display pages and that is a
> problem...even if for some reason I did manage to upgrade my CHIP
> RAM to 2 MB from 1 MB...it is insufficient to view any website with
> such limited RAM. Is it possible page viewing consume FAST RAM and
> not CHIP RAM at all?

No, certain things have to be in Chip RAM for the chipset to access
them.  The tile rendering area is one of these, and that's why it's
set to a small size by default.

You can try saving some Chip RAM by using "simple refresh" windows
(Prefs=>Display) or setting Cache native versions to "None" (Rendering
tab)

> 2) I keep receiving SSL Certification error message and if I click
> accept it freezes and the page does not load and it keeps popping
> hundreds more of the certification ssl error and I am forced to
> restart Amiga. If I reject it...it freezes or the page is not viewed

Not seen this one, I've have a look.

> Oh..a third problem...when I exit netsurf it sometimes crashes with
> 04 at the end.

Have heard about this but not seen it myself and without any debug
info I can't do much about it.

Please can you raise this on the bugtracker with a log attached:
netsurf -v >ram:ns.log

That might give a bit of a clue as to what is making it crash, and
gives some visibility of the problem.  I can add more logging if it
doesn't narrow it down enough.

Bugtracker is at http://bugs.netsurf-browser.org

Ta
Chris

Re: Netsurf on Amiga 68k

[Chris Young]

On Tue, 30 Aug 2016 11:27:02 -0700, Fahed al daye wrote:

> In that case...is it possible then you can send me the latest version
> of netsurf by link? I want to try this again.

http://cy2.uk/netsurfos3

That's always the latest build until CI is up and running.

Chris

Re: Netsurf on Amiga 68k

[Chris Young]

On Mon, 29 Aug 2016 17:13:08 -0700, Fahed al daye wrote:

> The redraw_tile_size_x and y made no difference.

Please reply to the list, not directly to me.

If you visit about:config does it show the correct values for these
options?  Did you set them whilst NetSurf was running? (it won't pick
them up until the next launch, and may overwrite the file in the
meantime)

Chris

Re: Netsurf on Amiga 68k

[Chris Young]

On Sun, 28 Aug 2016 11:45:23 -0700, Fahed al daye wrote:

> I have this annoying issue with the current Netsurf for the Amiga 68k
> in that lots of websites returns with this annoying error "Unable to
> fetch document." My understanding this is the cause because of
> timeout.

There are various timeout settings that can be set in your Choices
file, I doubt tha's actually the problem but without a log file
(netsurf -v >ram:ns.log) it's impossible to know.

> Also another issue when
> viewing pages, it open them in tile 100 by 100 and this is annoying
> and unpleasurable is there anyone this can be fixed?

redraw_tile_size_x and redraw_tile_size_y in Choices
(users/default/choices) will set the tile sizes.  This is detailed in
the netsurf.guide under Options file.

Chris

Re: Failure when receiving data from the peer

[Chris Young]

On Sat, 9 Jul 2016 18:23:57 +0100, Vincent Sanders wrote:

> On Sat, Jul 09, 2016 at 03:45:08PM +, Peter Slegg wrote:
> > 
> > When I try to open a link to this page:
> > 
> > http://www.ccrexplorers.com/forumdisplay.php?f=17
> > 
> 
> I have retrived this page with and without javascript on gtk and
> framebuffer frontends without problem so i must conclude the issue
> lies with the Atari frontend.

It's working here too but I did notice that the page downloads *very*
slowly, so maybe it is just timing out?

Try increasing curl_fetch_timeout and/or max_retried_fetches

Chris

Re: Text drag & drop

[Chris Young]

On 29 April 2016 14:43:00 BST, Vincent Sanders <vi...@netsurf-browser.org> 
wrote:
>On Wed, Apr 27, 2016 at 06:13:18PM +0100, Chris Young wrote:
>> On Wed, 27 Apr 2016 16:47:15 +0100, Richard Torrens (lists) wrote:
>> 
>> > Somewhere between 3492 and 3507 the ability to mark text, drag and
>drop
>> > elsewhere has been lost. Also the find text does not mark the
>finds.
>> > 
>> > This was very useful! I hope it's  temporary thing?
>> 
>> Marking text isn't working correctly on the Amiga version either. 
>Put
>> it on the bugtracker if you haven't already.
>> 
>
>Apologies for this Chris. I have been making some core changes
>improving the font layout API and messed up converting some
>frontends. I fixed the others but you got to Amiga before I had
>chance. Thanks for that

No problem. I saw what you did with the RISC OS version so I figured the fix 
would be the same. :)

>I have also recently changed the warn_user API. The core have very,
>very few usages of this left and mostly returns nserror from API
>letting the frontends report errors as they see fit. However this
>means frontends now provide the warning API through the miscellaneous
>operations table. Frontend code that should generate user warnings
>(amiga_warn_user for the amiga frontend) should be called directly
>
>To avoid breakage I kept the old warn_user() API in utils/utils.[ch]
>while I carefully remove usages. Rather than risk me breaking your
>frontend again might it be possible for you to remove the use of this
>API?

Yes, no problem.

I'm having some problems accessing this email account from home, so whilst I'm 
struggling typing on my phone, can you kick off a rebuild of the ppc-amiga 
toolchain please? I added some includes I now need for the version in master.

Thanks
Chris

Re: Text drag & drop

[Chris Young]

On Wed, 27 Apr 2016 16:47:15 +0100, Richard Torrens (lists) wrote:

> Somewhere between 3492 and 3507 the ability to mark text, drag and drop
> elsewhere has been lost. Also the find text does not mark the finds.
> 
> This was very useful! I hope it's  temporary thing?

Marking text isn't working correctly on the Amiga version either.  Put
it on the bugtracker if you haven't already.

Chris

Re: bad formatting of news page

[Chris Young]

On 30 March 2016 00:20:52 BST, Jim Nagel  wrote:
>What is it about the formatting of my local newspaper's web pages?
>
>Netsurf shows the main column of text so narrow that every line 
>requires horizontal scrolling.  Typical example:
>http://www.centralsomersetgazette.co.uk/Baily-s-costs-dream-revive-2-000-jobs-Morland-s/story-28054944-detail/story.html
>
>I find the least painful way to read the story is to save the whole 
>page as text and then read the textfile.
>
>Has the website programmer made some inept setting in the stylesheet 
>(CSS), or is Netsurf making some mistake or interpreting something too 
>literally?
>
>I presume browsers such as Firefox show what the designer intended.

Is it the same as this? http://bugs.netsurf-browser.org/mantis/view.php?id=2426

Chris

Re: Javascript features list

[Chris Young]

On Sat, 20 Feb 2016 00:41:08 +0100, Ole Loots wrote:

> Am Freitag, den 19.02.2016, 20:11 + schrieb Peter Slegg:
> 
> > I am looking at a survey form and wondered if buttons should work ?
> 
> 
> When I testet JS builds for Atari three weeks ago, onclick for buttons
> worked well.

It's interesting that it does appear to work on buttons, but not on
 links.

Chris

Re: Backgrounds fill stopped working?

[Chris Young]

On Sat, 13 Feb 2016 16:08:57 + (GMT), Bob Latham wrote:

> Is it me or has 
> 
> 
> 
> stopped working? All of the pages on my Intranet according to Netsurf have
> gone white.

Just tried it here - I have an intranet page which specifies a
background image in CSS which is working.  When I change it to specify
it in the body tag, it doesn't appear.
I will hazard a guess that this is related to the recent CSS hints
work.  Please raise it on the bugtracker.

Chris

Re: Big push on testing needed

[Chris Young]

On 26 January 2016 23:58:51 GMT+00:00, John Rickman Iyonix 
 wrote:
>Dave Higton  wrote
>
>> In message <57e4a64755.iyoj...@rickman.argonet.co.uk>
>>   John Rickman Iyonix  wrote:
>
>>>As far as I know javascript should ignore html comments and the
>>>javascript validator does not flag them as errors
>>>
>>> http://www.javascriptlint.com/online_lint.php
>
>> My reference suggests that an HTML comment is /not/ a legal
>Javascript
>> comment.  Perhaps you should open a discussion with the author of the
>> above application.  I would, of course, be very interested to know
>the
>> conclusion!
>
>It doesn't really matter now as I have changed all the comments to 
>"proper" JavaScript style. I don't remember where I got the idea that 
>html style was acceptable in JS, but the article in this link says 
>they are ok:-
>  http://www.javascripter.net/faq/comments.htm
>
>I will investigate further.

They must be OK to some extent, because the old advice was to encapsulate 
everything within 

Re: Very slow page rendering

[Chris Young]

On Sat, 9 Jan 2016 19:43:53 + (GMT), Peter Slegg wrote:

> >>>http://git.netsurf-browser.org/netsurf.git/tree/atari/gemtk/guiwin.c
> >>>
> >>>This page takes abut 20mins to download and render, Highwire browser
> >>>takes about 6sec.
> >
> >> I just tried it with CI#3254 on an Iyonix.  Took about 24 sec.
> >
> >> If you're not on a very recent CI build, I would recommend you get
> >> one.  Very slow rendering was fixed a couple of months ago.
> >
> >About 43s here on RPC with standard ADSL connection. CI#3250. Firefox
> >on the macbook pro takes 2s over the same connection.
> 
> I am using a build from today and I know it is never going to be
> lighting quick on an M68060 but my point is that the speed seems
> unduly slow over an adsl link that normally downloads at 350k/s.
> 
> I have reported before that Netsurf on the Atari seems slow at
> downloading pages so it might just be an issue with how Netsurf is
> using the OS ?  It's not my area of expertise.
> 
> Is it because there are a lot of files ?
> Is it a cache issue ?
> 
> Long delays can often be observed on other pages. The git url is
> a useful test case because it is fairly stable and repeatable.
> 
> No criticism, I am hoping this might help the devs find some speed
> improvements.

There is a lot of CSS on those git pages, I believe some CSS caching
is being looked into to reduce the amount of processing required.

In my dabblings with building a 68k version of NetSurf for AmigaOS 3,
I've noticed that the text layout and rendering code plays a big part
in the speed NetSurf loads pages.  On AmigaOS 3, rendering outline
font glyphs is painfully slow, and I think this is down to the 68k.

I notice the Atari code has three font rendering engines; VDI,
FreeType and Internal.  It look like FreeType and Internal are
enabled, and FreeType is the default.

Try setting the following in your Choices file:
atari_font_driver:internal

See if that makes any perceivable difference.

Chris

Re: Google crash

[Chris Young]

On Mon, 9 Nov 2015 23:23:25 +, Vincent Sanders wrote:

> I think i fixed this, can people try and let me know?

Working for me now, thanks Vince.

Chris

Re: simple html pages slow to render

[Chris Young]

On Sat, 07 Nov 2015 20:10:45 GMT, David Pitt wrote:

> (20.17) content/content.c:386 content_destroy: content 0x4e5f42a0 
> file:///%3CBootDisc$Root%3E/Work/Internet/Browsers/html/favicon.png
> 
> (41.38) content/fetchers/curl.c:834 fetch_curl_done: done 
> http://www.gaydon.org.uk/css/pcminutes.css
> 
> (41.39) content/fetchers/curl.c:872 fetch_curl_done: Unknown
> cURL response code 28

> This is missing :-
> 
> http://www.gaydon.org.uk/css/pcminutes.css

A missing file shouldn't cause a timeout though - just a response from
the server containing error 404.

Chris

Re: simple html pages slow to render

[Chris Young]

On Sat, 7 Nov 2015 15:08:01 +, David Pitt wrote:

> "Chris Young", on 7 Nov, wrote:
> 
> > On Sat, 07 Nov 2015 11:35:26 GMT, John Rickman Iyonix wrote:
> > 
> > > The Gaydon online parish magazine is a single HTML file which references
> > > a common CSS file.
> > > 
> > > http://www.gaydon.org.uk/mags.html
> > > 
> > > It takes on average 31 seconds to load and display a magazine using
> > > NetSurf #3055 on an Iyonix. Javascript on or off makes no difference.
> > > 
> > > Using NetSurf 3.2 the page load is instantaneous.
> > > 
> > > Not sure if this counts as a bug but I have submitted a report to the
> > > tracker.
> > 
> > Loading in 0.8 seconds here.  Can you attach a log to the bug report,
> > otherwise it's unlikely the problem will be identified?
> 
> The 31secs bit on #3055 occurs on downloading a magazine, not on downloading
> the page above. I can confirm this on my Iyonix running OS5.22 and a
> Raspberry Pi running OS 5.23 (03Nov15). It's all good on NetSurf 2.9 on Mac
> VRPC at 0.2s.

OK, I'd missed that the page mentioned wasn't the one with the
problem.  They all load in 2s or less here, so it might be a RISC OS
specific issue.

Chris

Re: simple html pages slow to render

[Chris Young]

On 7 November 2015 17:24:31 GMT+00:00, David Pitt  wrote:

>
>We RISC OS users think that NetSurf is a RISC OS browser when in fact 
>it is multi-platform. Indeed it can even run on an Amiga.
>

 Yes, I know ;)

Chris

Re: simple html pages slow to render

[Chris Young]

On Sat, 07 Nov 2015 23:40:55 GMT, Richard Porter wrote:

> On 7 Nov 2015 Chris Young wrote:
> 
> > A missing file shouldn't cause a timeout though - just a response from
> > the server containing error 404.
> 
> 404 is page not found.

It's file not found.  Any missing file on a web server will result in
a 404, but if it's a linked file like CSS you won't see the error
(it'll just be treated as a missing file, as you say).

In any case, it looks like that's not the reason for the timeout.

Chris

Re: PLI

[Chris Young]

On Fri, 28 Aug 2015 11:03:31 -0600, Anthony J. Bentley wrote:

 Dave Higton writes:
  On Fri, 28 Aug 2015 12:50:00 +0100 Richard Torrens wrote:
  
   Doing a list of genealogical notes, I wanted to space them out so there
   was a blank line between each record. So I replaced each LI by PLI.
  
   This does not work in Netsurf - but does in Firefox and Chrome
  
   Is this a bug?
  
   The page in question is
   www.torrens.org.uk/Genealogy/Torrens/DataScot/sasines.html
  
  p opens a paragraph.  You need to close it again afterwards with /p.
 
 No, in HTML p does not require a closing tag except in specific
 circumstances. The end of a li is not one of those circumstances.

Also, I don't think there's any guarantee that a new paragraph
actually inserts a blank line, and starting a new paragraph in the
middle of a list seems like an odd thing to do.

I always use br, same as Rob suggested.

 As for the original question, I would probably use CSS to add padding or
 margin at the bottom of each li.

That's probably the correct modern way to do it.

Chris

Re: Saving Choices

[Chris Young]

On 11 August 2015 11:16:38 BST, Brian Jordan brian.jord...@btinternet.com 
wrote:
In article c45a69f154.ga...@wra1th.plus.com,
   Gavin Wraith ga...@wra1th.plus.com wrote:
 I have noticed that Disable JavaScript appears not to be  saved
 in Choices$Write.WWW.NetSurf.Choices. Is that right?
 Using NS 3.4 #2876

That seems to be the case.

From the Jenkins changes log:

#2872   (9th August)

2. Permanently enable display of the 'enable javascript UI options'

Does this mean that on other platforms there is now a javascript on/off
button or menu option which does away with the need to permanently save
this preference? If so maybe this won't work in the RISC OS version
until
the UI is updated. Only guessing.

That's not relevant to your issue. In the Amiga frontend the Enable 
JavaScript toggles only displayed if NetSurf was built with JavaScript 
enabled. Now the JavaScript engine has changed, the conditionals weren't 
working, so I took the decision to remove them. There's no particular reason to 
disable JavaScript in the build as it doesn't involve building external 
libraries.

The setting should be saved. I believe it is an intentional feature that it is 
enabled on startup whilst the new JavaScript engine is integrated, so once that 
is complete it will start working again.

Chris

Re: Crash with #2860

[Chris Young]

On Sun, 19 Jul 2015 09:57:04 +0100, David Pitt wrote:

 NetSurf opens a local home page here on startup. Reboot the machine, start
 NetSurf, quit NetSurf then restart NetSurf, crash. This starts with #2859
 and also happens with #2860, however #2858 seems OK. If NetSurf is restarted
 without opening a home page then the crash happens on clicking on its
 iconbar icon.

Actually, try 2862, that's more likely to fix it.

Chris

Re: Crash with #2860

[Chris Young]

On Sun, 19 Jul 2015 09:57:04 +0100, David Pitt wrote:

 NetSurf opens a local home page here on startup. Reboot the machine, start
 NetSurf, quit NetSurf then restart NetSurf, crash. This starts with #2859
 and also happens with #2860, however #2858 seems OK. If NetSurf is restarted
 without opening a home page then the crash happens on clicking on its
 iconbar icon.

That'll be my fault then.

Try build 2861.  If that doesn't work I'll have a proper look when I
have some time to test.

Chris

Re: Crash with #2860

[Chris Young]

On 19 July 2015 16:03:07 BST, george greenfield 
george.greenfi...@tiscali.co.uk wrote:

 
 
Working fine here too, thanks to Chris.

Actually John-Mark fixed it, not me. :)

Chris

Re: BBC News

[Chris Young]

On 30 March 2015 15:24:09 BST, Tony Moore old_coas...@yahoo.co.uk wrote:
According to http://www.bbc.com/news/blogs-the-editors-32016166
published by the BBC, on 23 March 2015

   Our new responsive design [of the BBC News website], which we've
   just launched for desktop computer, aims to make sure the site looks
   great whichever device or screen size you are on - mobile, tablet or
   desktop.

However, http://www.bbc.co.uk/news , displayed by NetSurf #2655, on a
RiscPC, seems to be a version for mobile, and looks far from 'great'.

Do others see the same mess?

Yes, saw this on the day it launched. I was going to whinge about it on the 
blog but couldn't get the page to load, and then promptly forgot.

I don't know how the site picks the layout, if it's using the user-agent field 
then it must be assuming NetSurf is a mobile browser.

Chris

Re: Disc cache updates

[Chris Young]

On Sat, 14 Mar 2015 07:44:46 GMT, Andrew Pinder wrote:

 In message 20141201160839.gh10...@kyllikki.org
  on 1 Dec 2014 Vincent Sanders vi...@netsurf-browser.org wrote:
 
  The disc cache has recently been updated to track the speed of write
  operations.
 
  As a result of this performance tracking The browser will now detect
  if a system cannot sustain a write speed of one Megabit (120kilobytes/
  second) the cache will disable itself and display a warning.
 
 I'm finding this on my ARMini.  Is this common?

I get it on my SAM440 very frequently.  I'd be very surprised if the
write bandwidth was that low as it's a fast SATA drive.

Curiously I've never seen the message on my old EEE, which has the
slowest SSD write times known to man (it's slower than any HDD I've
encountered).

I suspect the OS has a lot to do with it, especially when writing lots
of small files there's likely to be significant overhead.

Chris

Re: Windows Autobuild

[Chris Young]

On Tue, 6 Jan 2015 00:10:59 +1000, Brendan Stephenson wrote:

 I've noticed two errors with the builds listed here:
 http://ci.netsurf-browser.org/builds/windows/
 
 1. All jpeg images load, but refuse to display (a blank box is shown instead).
 2. Text wraps mid word.
 
 I just want to confirm if these are issues common only to the builds
 in question, or if they are part of a larger problem.

The Windows frontend is unfinished.  I've never seen (2) with it -
although I only fire it up occasionally for testing - but (1) is
certainly prevalent.

If you have any patches I'm sure they will be welcome.

Chris

Re: Image Link issue

[Chris Young]

On Sat, 13 Dec 2014 15:26:52 + (GMT), Peter Slegg wrote:

 http://www.express.co.uk/news/nature/546649/Geminids-meteor-shower-live-stream-best-places-to-watch
 
 On the page above there are image links down the right-hand
 side.
 
 The top one is shown with the wrong width.
 
 The numbered ones are ok but when the mouse is moved over the
 ones further down the link shown is the same for all the images
 and when it is clicked the same page is opened regarless of the
 image.
 
 This is with Atari build no. 2427
 
 Can someone confirm this with other versions ?

Getting the same here with the Amiga frontend.  Also the top banner
cuts off on the right, ie. LIFE  STYLE  PUZZLES  P.
There are bits of banner behind the text of the top items in each
block of pictures, which looks wrong too - one of those is
apparently Stuart Winter.  There's two lots of text over the top
of him which has appropriate ALT text, but the links both point to his
column - which is all about birds and nothing to do with the text
(which is asteriods, and weather bombs).

I suspect it all stems from some CSS we either don't support yet or
isn't being handled correctly.  Put it on the big tracker.

Chris

Re: RISC OS Save-as PDF file

[Chris Young]

On Mon, 10 Nov 2014 23:02:19 +, Vincent Sanders wrote:

 On Mon, Nov 10, 2014 at 08:04:02PM +, John Williams wrote:
  
  I noticed recently (from another posting) that (some) other platforms seem
  to still have Save as PDF which was implemented under RISC OS and later
  removed.
 
 The export as PDF was implemented as a core feature using a library
 called libharu. unfortunately the library has not had a new release
 since version 2.2 (several years ago) and NetSurf requires fixes that
 have never been released.
 
 This issue combined with the original author abandoning the work and
 the feature becoming broken has resulted in PDF export being
 configured off for some time now. 

Further to that, I believe it also relies on the core printing API,
which has also been disabled for a long time pending (IIRC) some work
on duplicating contents.

Chris

Re: Cache on Atari build

[Chris Young]

On 5 October 2014 01:20:48 BST, Ole Loots o...@monochrom.net wrote:

Unpacking the package makes sure there is an folder called res/cache,
AFAIK.
I don't know how the netsurf core behaves when the folder is not found.


I think the front end is supposed to ensure it exists before enabling caching. 

Chris 

Re: Boxconvert

[Chris Young]

On Thu, 21 Aug 2014 23:12:19 +0100, Richard Porter wrote:

 I got a warning from NetSurf - Boxconvert
 
 What is boxconvert?

It means NetSurf hasn't been able to parse an HTML box (I'm not
entirely sure what the definion of a box is).  I suspect the plain
English version of this error has somehow gone missing.

Can you consistently create the error on a particular page? Do you
have a log file?

Chris

Re: Disc cache worth it?

[Chris Young]

On Mon, 23 Jun 2014 17:58:52 +0100, Peter Young wrote:

 If I load, for instance, http://www.bbc.co.uk/news/ as the first site 
 of a session, it loads maybe a little faster, but then I get 
 intermittent hourglass activity for sometimes up to thirty seconds, 
 during which I can't do anything else. There are several other sites, 
 for instance Wikipedia home page, which do the same. And the next day 
 the same happens.

I'm sure Vince will correct me here, but I believe NetSurf saves the
cache files to disk when they are downloaded (as opposed to when they
are evicted from the memory cache), so you will get a delay as NetSurf
gets busy saving the files.

If you've gone back to the same site though most of the files should
be loading from disk and I wouldn't expect any additional delay.  You
will need to post a log file so we can get a handle on exactly what is
causing the pauses.

 Looking in !Cache, which is in !Boot.!Resources, I find that in the 
 Caches.Default.NetSurf directory there are currently 1933 files, 
 totalling 22449384 bytes. Is this to be expected, as I don't use 
 NetSurf a huge amount?

Yes.  That's only 22MB.  The default limit is 1GB.

 I've already excluded this directory from my 
 daily backup, which has been taking a lot longer since I started using 
 !Cache.

There's no point in backing up these files.

Chris

Re: Disc cache worth it?

[Chris Young]

On Mon, 23 Jun 2014 21:13:17 +0100, Peter Young wrote:

 Thanks, Chris. I've a couple of busy days coming up, with one early 
 start, but I'll try to save a log after a new day's start-up, and give 
 some timings of the hourglass activity on the BBC news site. Where or 
 to whom should I post the log?

The best thing to do is raise a bug report and attach it to that.

Chris

Re: Disc cache worth it?

[Chris Young]

On Mon, 23 Jun 2014 22:43:52 +0100, Peter Young wrote:

  Thanks, Chris. I've a couple of busy days coming up, with one early
  start, but I'll try to save a log after a new day's start-up, and give
  some timings of the hourglass activity on the BBC news site. Where or
  to whom should I post the log?
 
  The best thing to do is raise a bug report and attach it to that.
 
 OK, but is it really a bug? I'll do as you suggest, anyway.

If it's not working as it should, it's a bug. :)

There may be something that can be done to improve matters, such as
delaying disk cache writes until the browser isn't busy, or only
writing out when the item is removed from the memory cache.  It may be
that it is a true bug and is writing when it shouldn't be, or there is
some performance problem which is nothing to do with disk I/O.

Chris

Re: BoxConvert?

[Chris Young]

On 8 June 2014 12:18:49 BST, Bob Latham b...@mightyoak.org.uk wrote:
In article
out-539399f2.md-1.4.17.chris.yo...@unsatisfactorysoftware.co.uk,
   Chris Young chris.yo...@unsatisfactorysoftware.co.uk wrote:
 On Sat, 07 Jun 2014 19:58:41 +0100, Richard Torrens (lists) wrote:

  After a bit more experimenting, it is caused by the line:
  
  Pa href=http://!--#echo var=HTTP_HOST --!--#echo
  var=DOCUMENT_URI --Top of page/a

 Ah, right.  I know exactly what the problem is and what causes it.

 NetSurf now rejects invalid URLs.  By invalid I mean, anything
 containing characters that are not permitted either in old-style
ASCII
 names, or in IDNA, plus some other validity checking mostly related
to
 IDNs.

 Previously NetSurf would percent-encode the host and return that, so
 you'd end up with a garbage URL.  Now it rejects the URL outright. 
 However, it appears the layouting can't cope with that.  I did some
 extensive searching when I made this change initially and was unable
 to track down why this was causing the layouting to fail - ideally, a
 broken URL shouldn't cause termination of the layouting process, it
 should only make the problem link go nowhere.

 I considered this low risk at the time as it is incredibly rare to
 encounter a link to a URL which doesn't conform to the rules.

 However, obviously it does need to be fixed, so if you can raise it
on
 the bugtracker that'd be grand.

Good luck with that.

I'm plagued with this BocConvert as well as told in my thread Username
and Password problem.

Daniel showed some interest and asked for information and I've tried
everything to get it to him and after (in total) a couple of hours
trying
I've given up.

The log from my Netsurf seems to say that it bombs out BoxConvert
because
it encounters the character 5f presumably hex or 5F. I have searched
through the relevant html on my intranet and no 5F anywhere and
certainly
not in a page url.

Mine is triggered not by the web page as such but by the security
username
and password on items inside one folder on my Cherokee driven Intranet.
   

This started to happen for me with build 1950 and has persisted ever
since
but builds up to and including 1949 did not have the problem. I would
not
be surprised if the problem in this thread started with also 1950.

As regards reporting the problem, I found this an obstacle course and
no
doubt.

Firstly, I was asked for a log file from netsurf which after some
reading
and testing I managed to generate. I possibly wrongly emailed it to the
mailing list. I was told it was being help for the next moderator as it
was a little too big.

So I emailed Daniel to his email address with the file, no reply.

He also asked me for something in cgi. Err what?

And the config file from Cherokee for which I've asked what's the file
name and where it may be in the structure but had no reply.

The I thought maybe I should look into this bug tracker whatsit, that
when
the problems started.

Registration kept bombing out or doing nothing and I had to use a PC to
get this done. I then logged in using the ARMiniX and Netsurf filled in
a
report and then found in Netsurf I could not attach a log file.

I logged in on the pc, filled min the forms all again, attached the
file,
hit send and it bombed out again.

At this point I walked away pretty wound up if I'm honest.

Apologies for the excessive quoting. Can you send the log to me?

Chris

Re: Username and password problem

[Chris Young]

On Sun, 08 Jun 2014 19:27:15 +0100, Bob Latham wrote:

 Chris Young has asked for a copy of the log and I've sent it to him zipped.

It's the same cause as Richard's problem.

I'll see if I can find some time to look into it again, however it
would be very helpful if somebody could create a bug report for this,
as it makes it easier to manage and ensures it doesn't get forgotten.

Thanks
Chris

Re: BoxConvert?

[Chris Young]

On Sat, 07 Jun 2014 16:50:01 +0100, Richard Torrens (lists) wrote:

 What does the message BoxConvert mean? It's obscure to say the least!

It means NetSurf wasn't able to convert a box :)

I don't fully understand how it works, but it is to do with layouting.

 I now (3.2 (Dev CI #1959)) get it on every local www page I run. Which is
 a pain - as I cannot then check a www page before uploading.
 
 I suspect it may be because I use SSI on my www pages: they contain, for
 example

I wouldn't have thought so.  The SSI portion is an HTML comment if it
isn't getting parsed by the server, so it should be getting ignored.

a) When did this start?
b) Can you post one of the faulty pages here, or upload it in a way in
which it can be downloaded without the SSI being parsed?

Chris

Re: BoxConvert?

[Chris Young]

On 7 Jun 2014 17:32:35 +0100, Chris Young wrote:

 On Sat, 07 Jun 2014 16:50:01 +0100, Richard Torrens (lists) wrote:
  I now (3.2 (Dev CI #1959)) get it on every local www page I run. Which is
  a pain - as I cannot then check a www page before uploading.

 a) When did this start?
 b) Can you post one of the faulty pages here, or upload it in a way in
 which it can be downloaded without the SSI being parsed?
c) Log file?

Somehow missed (c).

Chris

Re: BoxConvert?

[Chris Young]

On Sat, 07 Jun 2014 19:58:41 +0100, Richard Torrens (lists) wrote:

 After a bit more experimenting, it is caused by the line:
 
 Pa href=http://!--#echo var=HTTP_HOST --!--#echo
 var=DOCUMENT_URI --Top of page/a

Ah, right.  I know exactly what the problem is and what causes it.

NetSurf now rejects invalid URLs.  By invalid I mean, anything
containing characters that are not permitted either in old-style ASCII
names, or in IDNA, plus some other validity checking mostly related to
IDNs.

Previously NetSurf would percent-encode the host and return that, so
you'd end up with a garbage URL.  Now it rejects the URL outright. 
However, it appears the layouting can't cope with that.  I did some
extensive searching when I made this change initially and was unable
to track down why this was causing the layouting to fail - ideally, a
broken URL shouldn't cause termination of the layouting process, it
should only make the problem link go nowhere.

I considered this low risk at the time as it is incredibly rare to
encounter a link to a URL which doesn't conform to the rules.

However, obviously it does need to be fixed, so if you can raise it on
the bugtracker that'd be grand.

Chris

Re: Accents in URLs

[Chris Young]

On Mon, 10 Mar 2014 17:10:23 + (GMT), Tim Hill wrote:

 Just came across this URL which opens in 'other' browsers but not in RISC
 OS NetSurf here. Okay, I'll admit to using an old version (3.1 #1298)
 but no doubt I'll be told if later versions do cope. ;-)
 
 http://dahlström.net/svg/favicon/favicon.html

Just a note that with the latest CI build you will now be able to
follow these links on web pages.  You should also be able to type them
in, however I've not tested that on RISC OS, so if it doesn't work,
please raise it on the bugtracker.

Chris

Re: Resolver failure on VRPC on the Mac

[Chris Young]

On Sat, 19 Apr 2014 15:57:21 +0100, Rob Kendrick wrote:

 On Sat, Apr 19, 2014 at 12:07:19PM +0100, David Pitt wrote:
  NetSurf version '2.9 (27th Febuary 2012)'
  
  (0.1) content/fetchers/curl.c fetch_curl_register 167: curl_version
  libcurl/7.20.1 OpenSSL/0.9.8n zlib/1.2.3.3 c-ares/1.6.0
  
  
  NetSurf version '3.1 (Dev CI #1809)'
  
  (0.02) content/fetchers/curl.c fetch_curl_register 163: curl_version
  libcurl/7.34.0 OpenSSL/1.0.1g zlib/1.2.8 c-ares/1.10.0
 
 It could indeed be that the newer c-ares (the asynchronous DNS resolver
 we use) does not like whatever is listening on 127.0.0.1 for DNS
 queries.

Possibly not relevant, but I've never managed to get c-ares working on
AmigaOS either.  It has been a while, but IIRC it just failed to
resolve anything, so similar symptoms.  I've always used Curl's
internal resolver (and more recently, the threaded version of the
same), which works perfectly.

Perhaps we could make a test build of NetSurf which uses the internal
threaded resolver instead of c-ares, to see if that works?

Chris

Re: beautiful wikipedia design

[Chris Young]

On Mon, 07 Apr 2014 11:28:41 +0100, Gavin Wraith wrote:

  Netsurf #1785 crashes on
  http://www.fastcodesign.com/3028615/the-beautiful-wikipedia-design-tha
  t-almost-w
  as
 
  It doesn't here.
 
 Actually, it was not going to this site directly that crashed NetSurf
 but clicking the link from Ars Technica in Ars Editor's Picks.

The address of the page containing the link and a log file from when
it crashes, would be exceedingly helpful.

Chris

Re: Accents in URLs

[Chris Young]

On Mon, 10 Mar 2014 17:36:17 +, Rob Kendrick wrote:

 On Mon, Mar 10, 2014 at 05:10:23PM +, Tim Hill wrote:
  
  Just came across this URL which opens in 'other' browsers but not in RISC
  OS NetSurf here. Okay, I'll admit to using an old version (3.1 #1298)
  but no doubt I'll be told if later versions do cope. ;-)
  
  http://dahlström.net/svg/favicon/favicon.html
 
 punycode.  And I don't think we support it, no. 

I've been having a look at this, largely for my own curiousity.
I've managed to add some sort of half-working punycode, but it has
raised a question:
Are nsurls supposed to be UTF-8?
I don't think they are, which means the urlencoding it does might not
necessarily be consistent.  Moreover, the scheme and host are getting
urlencoded when they shouldn't be (although this is easy to fix).

Chris

Re: Amazon e-mail field

[Chris Young]

On Sun, 2 Mar 2014 17:10:35 + (GMT), Peter Slegg wrote:

 I noticed that the e-mail field on Amazon is displayed quite small.
 input id=ap_email name=email value= type=email size=30 
 maxlength=128 tabindex=1 autocorrect=off autocapitalize=off /

The problem here appears to be related to type=email, as without
that the size is correct.

Please can you raise it on the bugtracker.

Chris

Re: Netsurf and the Mantis bug tracker

[Chris Young]

On Sat, 08 Feb 2014 11:29:14 + (GMT), Brian Jordan wrote:

 I carefully completed the form on the Mantis site and submitted it,
 Mantis responded with an page indicating that the information I had
 supplied about the CI build was incorrect and that I should use my
 browser's back button to return to the form and correct the error. This I
 did only to be presented with a totally blank page from which to submit
 my report. Fortunately I had had the foresight to compose the report in
 Zap so didn't have the frustration of typing it all again so I completed
 the form and ensured the build information was correct and it was
 successfully submitted.

I'm not sure that is a bug, but more a feature request.  It is easily
reproduced by searching for something using Google, and then pressing
back after the search results are displayed.  The previous page
displays, but without the previous search term pre-filled.

You would also have seen the same effect had you clicked back after
submitting your bug report.

Chris

Re: Cut to clipboard

[Chris Young]

On Tue, 01 Oct 2013 14:16:24 +0100, Michael Drake wrote:

  Editing blunders should be handled by undo/redo.  NetSurf's textarea
  doesn't currently support this.
 
 Supported via keyboard shortcut in the latest build.  The shortcut depends
 on the front end.

   ++==+==+==+=+
   |   GTK  Framebuffer   |   RISC OS   |  AmigaOS 4  |
   ++==+==+==+==+==+
   | Undo   | Redo | Undo | Redo | Undo | Redo |
   |+--+--+--|--+--+
   | Ctrl+Z | Ctrl+Y or| F8   | F9   | RAmiga+Z | RAmiga+Y |
   || Ctrl+Shift+Z |  |  |  |  |
   ++==+==+==+==+==+

 The other front ends don't yet pass KEY_UNDO / KEY_REDO to the core.

AmigaOS frontend does now, see above.
I've used RAmiga+Z because that's what the Style Guide says should be
Undo, despite Workbench using Z for Clear selection.  I've removed the
shortcut from Clear selection as I don't think it needs one anyway
(the Style Guide didn't have an opinion on that or Redo).

Are there any can_undo()/can_redo() type functions?  I'd like to
disable the items when they can't be used, although it's not a major
problem.

Chris

Re: Sourceforge

[Chris Young]

On Fri, 05 Jul 2013 16:22:09 +0100, Jim Nagel wrote:

 While we're on about Sourceforge, could someone already familiar with 
 its workings please explain to us in the cheap seats what OpenID is 
 all about?  This page says OpenID eliminates the need for multiple 
 usernames across different websites.  Does that mean different 
 Sourceforge websites or what?

It's an authentication system that works on any site that supports it.
 The idea is that you can choose your own OpenID provider, and use
that to log in on any website (well, as long as that website supports
OpenID).  It hasn't taken off in a big way, but it largely operates
(from a user perspective at least) the same as the log in with
Facebook, Log in with Twitter, Log in with my Google Account etc
buttons a lot of sites have.

I quite like the one on Sourceforge, as they used to force password
changes quite often, and logging in with an OpenID circumvented that. 
However, unless they've changed it recently, you can't actually log in
with an OpenID on Sourceforge if you're using NetSurf.

Chris

Re: Testing required: visited link performance

[Chris Young]

On Fri, 17 May 2013 13:06:07 +0100, Michael Drake wrote:

 It was originally implemented several years ago, but disabled since it had
 a detrimental effect on performance.

IIRC it took minutes to display even the Google homepage when this was
initially enabled.  I'm happy to report it is not having any
perceivable impact on performance at all now.

System: SAM440EP (600MHz PowerPC)
URL file size: 50K

Site: http://www.google.co.uk
Build #1177: 0.6s
Build #1178: 0.6s

Site: http://news.google.co.uk
Build #1177: 6.6s
Build #1178: 6.8s

Site: http://www.digitalspy.co.uk
Build #1177: 5.1s
Build #1178: 5.2s

Site: http://www.netsurf-browser.org
Build #1177: 2.2s
Build #1178: 2.1s

The visited links don't display differently on Google because of the
annoying redirect links they use these days, which change on every
reload.  It is visibly working on the NetSurf homepage however -
except for the Wakefield show link, where I think the initial URL
isn't being remembered by NetSurf, as the link redirects immediately
to http://www.wakefieldshow.org.uk/index.php

Chris

Re: quitting netsurf

[Chris Young]

On Thu, 16 May 2013 10:07:37 GMT, Tony Moore wrote:

   #1172 loads in 16 secs, and quits in 26 secs. The closing log is
   here https://dl.dropbox.com/u/77062274/Log_1172
 
  Thanks, please try #1175.
 
 #1175 loads in 16 secs, and quits in 3 secs (even after visiting several
 websites).
 
 Problem solved - many thanks.

I must admit I noticed a long time ago that NetSurf took ages to quit,
but I hadn't recognised it as an actual problem (I thought it was just
saving cookie files, history, etc).  It now shuts down the instant I
close it, much better!

Chris

Re: Orpheus Home page

[Chris Young]

On 01/05/13 17:45, Tony Moore wrote:

On 1 May 2013, Daniel Silverstone dsilv...@netsurf-browser.org wrote:

The JS support in the current builds is very likely to be identical to
the 3.0 build.


http://javatester.org/javascript.html responds to NS 3.0 with a blank
Test Results window, whereas the response to the current 3.1 build is
verbose. I took this to indicate JS progress. I'm sorry if the assumtion
was not correct.


I think it is highly likely that NetSurf 3.0 has been built without 
JavaScript support.  This would explain the persistent disabling of 
JavaScript on startup.  When JavaScript is enabled, the blank page on 
JavaScript tester is a result of the browser thinking it supports 
JavaScript (so the noscript section gets ignored), but not supporting 
JavaScript (so the expected text does not get printed, as it uses 
JavaScript to print that text.


The only bug here is that the Disable JavaScript choices option should 
not be available in builds that do not support JavaScript.


Chris

Re: Upper character truncation - CSS handling

[Chris Young]

On Mon, 15 Apr 2013 08:27:10 +0100, Vincent Sanders wrote:

 I also note NetSurf falls back to the sans-serif font choice
 which should not be a factor.

That should probably be on the request tracker too, if it isn't
already.  I remember some discussion about implementing usage of
specific fonts quite some years ago.

Chris

Re: Text area refresh

[Chris Young]

On Fri, 5 Apr 2013 13:09:01 +0200, J. F. Lemaire wrote:

 On 5 April 2013 12:30, p.sl...@scubadivers.co.uk wrote:
 
  What I have noticed recently is that pasting into the text area with Ctrl-V
  creates 2 copies of the pasted text.
 
 Yes, I noticed this too. I suspect this is a frontend issue, however,
 not a core one, since each platform has its own mechanism to access
 the clipboard.

I'm not so sure.  Here the frontend handles RAmiga-V as a paste
shortcut, and that works wonderfully.  However, something (I'm
suspecting the core, as I haven't added any code for this) is trapping
Ctrl-V and handling that as a paste operation too.

You've neglected to mention what platform you are using, but what I
suspect is happening, is the frontend is handling Ctrl-V as a paste
operation, and that shortcut is also reaching the core, which handles
it again as a paste - hence the two pastes.

The core handling of Ctrl-V (and any other similar shortcuts) probably
needs to be tracked down and obliterated.

Chris

Re: Text area refresh

[Chris Young]

On Mon, 08 Apr 2013 19:54:52 +0100, Michael Drake wrote:

 The core handling of Ctrl-V is the only supported way of pasting.
 
 When you select Paste in a menu, the front end should be implementing it
 by passing Ctrl-V to the core.

Oh yes, I didn't realise KEY_PASTE = ASCII 22 = Ctrl-V (until I looked
it up)

I still think the double-pasting would be the frontend interpreting it
directly as a KEY_PASTE *and* passing the original key press as well.

Chris

Re: Incompatible site?

[Chris Young]

On 17 Mar 2013 13:14:39 +0100, Chris Young wrote:

 On Sun, 17 Mar 2013 11:49:40 + (GMT),  wrote:
 
  
  On Netsurf - all versions I have, including 3.0 (Dev# 975) Netsurf crashes
  out on trying to access
  www.itproportal.com/2013/02/19/10-tips-for-the-samsung-galaxy-s3/
 
 It's ok here if I go straight to the site.  If I then go to Google and
 try and navigate back to that site (with direct URL input) the whole
 thing freezes up.  That's on the Amiga version.

I'm getting:
assertion (box-width != UNKNOWN_WIDTH)  (box-height != AUTO) failed: file 
render/layout.c, line 5026

Log here: http://homepage.ntlworld.com/cdyoung/tmp/ns.log.gz

 I've been seeing crashes in dom_hubbub_parser_destroy recently (and
 not systematically reproducably) that may be related.

Looks like these aren't related, so I'll have to keep looking for that
one.

Chris

Re: favicon confusion

[Chris Young]

On Fri, 08 Mar 2013 16:25:07 GMT, John Rickman Iyonix wrote:

 Mystery solved - I develop my web pages offline using relative 
 reerences to files. NetSurf behaves differently in this mode with 
 files and does not pick up the common favicon.
 
 It behaves the same as you describe when accessing pages from the web 
 server.

If you are working locally, the root will be the root of the
volume/drive your pages are stored on, which probably isn't the same
as the root directory of the files you are going to upload.  Hence
NetSurf won't find favicon.ico locally unless you give it additional
hints in HTML HEAD using relative paths.

Chris

Re: BBC News websites

[Chris Young]

Richard Porter r...@minijem.plus.com wrote:
 Considering their
clucking
 over the raspberry pi you would think  no large organisation,
 different hands, no co-ordination! Richard, you could let them know
that
 apparently very large numbers of the little wonders have been sold
which
 might help us on legacy hardware!

Yes, will do. I've just got another reply from the Beeb:

Please note that the issue was once again escalated to Future Media 
and finding that your browser is NetSurf running under RISC OS 6.16 
running on a StrongARM powered RiscPC informed that this is old 
technology and may explain why you are being directed to the mobile 
version of the website.

Does NetSurf 2.9 under Raspbian (ie the Pi's default OS with - I think - the 
default/recommended web browser) have the same problem (it should)? Report that 
and maybe they will take notice rather than hiding under the old technology 
banner.

On the plus side, they didn't call it obsolete technology.

Re: Long URL annoyance

[Chris Young]

On Thu, 7 Mar 2013 20:32:36 +, Steve Fryatt wrote:

 The front-end doesn't seem to know anything about the status bar content
 (beyond that it's some text), so it might be pretty stupid.  Intelligence
 such as Send form to http://foo.com/...bar/wibble.php; would probably
 require some very educated guesswork, I suspect (unless there's a way to
 negotiate the length of the text with the core?).

There isn't.  We'd need something like the nsfont interface for status
bar text for the core to intelligently reduce the size, I guess.

 I see there's actually some tooltip code buried in the local history GUI for
 the RISC OS front-end (turn them on via the option in Interface).  Can a
 front-end find out from the core that the pointer is hovering over a link,
 form submission button or similar?

browser_window_get_contextual_content() can do this.  ccdata.link_url
is probably the main thing to check, that will get the URL at least,
although alt text won't be available.  It returns data on forms too,
although I'm not sure what.

Chris

Re: Long URL annoyance

[Chris Young]

On Wed, 06 Mar 2013 23:40:00 GMT, Simon Smith wrote:

 A fix which would mitigate matters is if the proportion of the lower scroll
 bar (or, even better, the absolute size in pixels) allocated to the
 horizontal scroll was configurable and/or persistent from window to window.

You can configure it by changing toolbar_status_size in your Choices
file.  It's in 100ths of a percent, eg. the default is 6667, which
equates to 66.67% of the horizontal space allocated to the status bar
part.

 I appreciate that some URLs are impossibly long, and only a multi-line URL
 display would ever be able to fully display them, but in the meantime there
 are several ways you could mitigate the problem. (e.g. a key shortcut that
 toggles full-length (or at least a longer) URL display, a hovering link
 display tooltip, using the !Help application to display the link, etc.)
 Personally I would be happy to have the URL given a whole line of its own,
 even though that would sacrifice some vertical screen space, because I feel
 this info is important enough to justify such a usage. And a full line would
 usually be sufficient space, whereas on my setup half a line often isn't.

A better fix might be for the end of the URL to be given priority (aka
right-justify), or some shortening technique where the domain and the
end of the URL are shown, but the middle is truncated with ellipsis if
it is above a certain length.  I'm not sure the latter is feasible
without some changes in NetSurf, but the former is easy to change in
the frontend code.  I can however see a problem with links which also
have alt text, as that tends to get appended to the URL in the status
bar and _might_ prevent the URL being shown at all if it is
particularly long.

Chris

Re: Failure to post bug report.

[Chris Young]

On Tue, 26 Feb 2013 20:00:35 GMT, Dave Higton wrote:

 Three different results with #949 json (JS enabled):
 
 1) georgesregisjazzband: causes a crash.
 
 2) document-records: renders OK.
 
 3) amiga.org: infinite fetching.

With the absolute latest and JS enabled, amiga.org fetches for ages
and then displays a message which simply says OK (reloading then
displays it correctly). At one point it crashed, I didn't save the
crashlog though.

Chris

Re: page from The Register stalls Netsurf

[Chris Young]

On Sun, 03 Feb 2013 16:13:10 GMT, Jim Nagel wrote:

 This page at The Register opens OK:
 http://www.theregister.co.uk/2013/01/28/florida_incident/
 
 But, near the bottom, clicking on the first link in Related stories
 (Customer service rep fired for writing game that mocks callers)
 causes Netsurf to hang -- it says Fetching, processing for ever.

This is JavaScript related.  It works fine with JavaScript disabled.

Chris

Re: Toolbar Customisation

[Chris Young]

On Thu, 24 Jan 2013 09:32:03 + (GMT), John Harrison wrote:

  Perhaps the correct information to provide to Ken, is that the toolbar
  on the GTK version of NetSurf is not cutomisable.
 
 OTOH, if GTK provides a context sensitive menu using a different button,
 then try that over the toolbar and see what happens.

It does, under the right mouse button, but no amount of clicking is
going to customise a toolbar that can't be edited.

Chris

Re: Toolbar Customisation

[Chris Young]

On Wed, 23 Jan 2013 14:51:34 GMT, Peter Young (no relation) wrote:

  Ken Ellis wrote on 21 Jan:
  I am trying to use Netsurf on a Raspberry Pi ...
 
  Essential point:  which OS are you using on your Pi?  Netsurf comes in
  versions for RiscOS, Linux and others.
 
 Good point, Jim lad, but if his help file mentions Select, Adjust and 
 Menu, then that suggests that Ken is using RISC OS :-)

Not necessarily.  I believe the GTK version (which I'm pretty sure is
what is being used) does not have its own help file, but instead links
to the NetSurf website.  The documentation area of the NetSurf website
is, as I'm sure you are aware, targetted at the RISC OS version of
NetSurf.

Perhaps the correct information to provide to Ken, is that the toolbar
on the GTK version of NetSurf is not cutomisable.

Chris

Re: On line passport application

[Chris Young]

On Sun, 13 Jan 2013 23:32:14 + (GMT), Tim Hill wrote:

 As usual, this seems to be a NetSurf+Javascript issue. On 'another
 platform' with a modern browser, the page linked to from
 passports.ips.gov.uk is
 https://passports.ips.gov.uk/epa1r1a/Rich/Formpage.aspx and not what
 Netsurf comes up with that you quote; it uses a javascript link. Even
 NetSurf's partial javascript is obviously not catered for.

The solution is actually to DISABLE JavaScript for it to work. 
When enabled it tries to go to
javascript:__doPostBack(\'ctl04\',\'\'), whatever that is supposed
to do is clearly not supported/working in NetSurf currently.

With JS disabled, NetSurf goes to the following page:
https://passports.ips.gov.uk/epa1r1a/NoScriptRich/Formpage.aspx
Which should work perfectly, being a non-JS version of the site.

Chris

Re: Javascript and Fonts

[Chris Young]

On Sun, 16 Dec 2012 11:11:57 GMT, Gavin Wraith wrote:

 So how
 do I flush the RUfl cache, and get it to start over when I am trying out
 a new set of fonts?

*delete scrap$dir.scrapdirs.scrapdir.RUfl_cache

(that's possibly out-of-date, but I don't think it has changed)

Chris

Re: RISC OS Javascript support

[Chris Young]

On Sun, 16 Dec 2012 13:45:46 +, Vincent Sanders wrote:

  1.7 doesn't need NSPR

 I must have misunderstood what teh DIST variable points at then? but
 thats perfectly possible ;-)

IIRC it needs NSPR if you build it as threaded, but you don't need to
do that.

 well I would be very interested in a diff against the distribution
 archive anyway. If I can get the NetSurf toolchain package building
 spidermonkey for your platform we can get the CI builds (and by
 extension future release builds) to generate output with JavaScript
 support for your platform

Attached.

Chris


js17.diff.gz
Description: Binary data


makefile.os4-lib.gz
Description: Binary data

Re: RISC OS Javascript support

[Chris Young]

On Thu, 13 Dec 2012 13:22:43 GMT, John Rickman Iyonix wrote:

 Great news! Does the Javascript need to be switched O?

It is on by default, however if you've been running a non-JS build and
have changed NetSurf's settings, you may find it has saved
use_javascript:0 into your Choices file.

Remove that line or change it to a 1 and JS will be enabled.

If Steve is still around he should be able to add that to the GUI
quite easily.

Chris

Re: NetSurf Developer Workshop

[Chris Young]

On Wed, 07 Nov 2012 22:16:41 + (GMT), Chris Gransden wrote:

 In article 20121107215800.GB2863@somnambulist.local,
Daniel Silverstone dsilv...@netsurf-browser.org wrote:
 
  Until and unless we can get libmozjs ported to RISC OS, it'll be jsoff there
  all the time :-(
 
 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.

Sounds like the problem I was getting, which this patch fixes:
http://git.netsurf-browser.org/netsurf.git/patch/?id=24c43ea6a210cd1b98e911d989de400ef343e7bb

NetSurf won't build with 1.70 (yet), and me and Ole can't get 1.80 to
work on big-endian hosts, but if that is an endian problem it's not
going to be relevant for RISC OS/ARM.  Beyond 1.80 you're into C++ and
NSPR territory, whch adds porting complexity and increasingly bloaty
executables.

Chris

Re: More graceful handling of failed image renders?

[Chris Young]

On 04/11/12 16:16, Martin Bazley wrote:

NetSurf redraws its pages in rectangular sections.  This process
involves decoding any image files, such as JPEGs, which may be present
in that section.

At the moment, if an error occurs (e.g. JPEG data is invalid or
otherwise fails to decode), the redraw of the entire section is
cancelled.  This is undesirable, because it leads to the appearance of
large blank white spaces on a page with invalid images on it.

You can see an example of the problems this causes here:

http://www.corporateskull.com/

There are two invalid JPEGs in the sidebar there (or at least JPEGs
which RISC OS doesn't seem to like; Linux renders them OK).  These are
the link to 'Skull Art' just beneath the (working) 'Corporate Skull
Guest Artist Week' image, and the Ink Outbreak avatar in the top-left of
the box beneath that.

If you slowly scroll down the page using NetSurf on RISC OS, you'll see
a thick white horizontal bar blotting out the image in the blog post to
the left of the Skull Art picture, as well as the Skull Art picture
itself failing to render.

If you press F12 and then Return with the faulty JPEG showing, most of
the window goes white!

NetSurf may not be able to render an image, but could it at least not
trash the rest of the page because of it?


Sounds as if this might be a manifestation of bug #3435099

Chris

Re: Google + -Your Browser is no longer supported

[Chris Young]

On Thu, 25 Oct 2012 21:55:32 +0100, John Rickman Iyonix wrote:

 I am trying to find a cloud based application for sharing files across 
 several systems.
 
 These are: Iyonix RISC OS, HTC phone Android, 10 Tablet Android, 
 Raspberry Pi linux and RISC OS, and an Intel machine running Lubuntu.
 
 Google + or rather Google Drive is one possiblility, Dropbox and 
 EverNote are two others.
 
 While not expecting to be able to save and retrieve files owing to the 
 lack of java Script, I was hoping at least to be to be able to see a 
 list of the files put up from Linux. But if the message is to be 
 believed Google Drive tests for specific Browser ids.

I like MyDrive (http://www.mydrive.net).  It's WebDAV-based so a nice
open standard that can potentially work on any OS.  I know of virtual
WebDAV filesystems for Windows, Linux, MacOS, Android and AmigaOS
(unfortunately I've been unable to locate one for RISC OS), and for
others there is a web interface.

MyDrive's web interface works fine under NetSurf for browsing and
downloading files, however it is impossible to upload anything.

Chris

Re: RISC OS: Squeezed executable

[Chris Young]

On Sun, 14 Oct 2012 14:31:48 +0100, Rob Kendrick wrote:

  I'm at a loss to understand exactly what 'bitmaps' NetSurf could have
  'modified' in the course of the session, or where it's 'saving' them to.
  All I know is that this is a horrendously inefficient operation!
 
 They're probably the webpage thumbnails used in the history viewer.  It
 might be nice to have an option to disable this.

Is this a RISC OS-only feature, as thumbnails get lost here after
quitting?

Chris

Re: Could someone look at

[Chris Young]

On Mon, 08 Oct 2012 00:12:54 +0100, lists wrote:

 Could someone else take a look at:
 
 http://www.audon.co.uk/usb_scopes/gcr101.html
 
 Here it looks like the page is about 200 yards wide.
 
 Netsurf 2.9 (27th Feb 2012)

http://www.amigans.net/ has gone wide recently too, but I suspect
that's a different problem, as it is fine in NetSurf 2.9 and not so
fine in NetSurf 3.0 (revision 9cb4d0, a couple of days old)

Chris

Re: Could someone look at

[Chris Young]

On Tue, 09 Oct 2012 22:54:56 +0100, Chris Newman wrote:

   http://www.amigans.net/ has gone wide recently too, but I suspect
   that's a different problem, as it is fine in NetSurf 2.9 and not so
   fine in NetSurf 3.0 (revision 9cb4d0, a couple of days old)
 
  It is fine on my set up,
 
 Also fine here   Dev C1 #417  on  RiscPC 700  Adjust 4.39 with Unipod

Actually it's fine here on latest CI#417, should have checked that
first, sorry...

Chris

Re: Request for feedback

[Chris Young]

On Sat, 06 Oct 2012 16:45:03 +0100, John-Mark Bell wrote:

 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.

It just appears to terminate without warning, which is odd (and I
don't see what would cause this?)

I'll try with one of my own builds in case it's something specific to
CI.

(18.803219) desktop/browser.c browser_window_go_post 849: bw 0x56d7dda8, url 
http://www.danvk.org/wp/dragtable
(18.809603) utils/url.c url_host_is_ip_address 124: WARNING: called with 
non-host 'www.google.co.uk/'
(18.830195) desktop/browser.c browser_window_go_post 970: Loading 
'http://www.danvk.org/wp/dragtable'
(18.839547) content/fetchers/curl.c fetch_curl_setup 365: fetch 0x56c67698, url 
'http://www.danvk.org/wp/dragtable'
(18.848337) utils/url.c url_host_is_ip_address 124: WARNING: called with 
non-host 'www.danvk.org/wp/dragtable'
(19.611637) content/fetchers/curl.c fetch_curl_done 820: done 
http://www.danvk.org/wp/dragtable
(19.614008) content/fetchers/curl.c fetch_curl_process_headers 1178: HTTP 
status code 301
(19.617387) content/fetchers/curl.c fetch_curl_process_headers 1189: 
FETCH_REDIRECT, 'http://www.danvk.org/wp/dragtable/'
(19.623230) content/fetchers/curl.c fetch_curl_abort 681: fetch 0x56c67698, url 
'http://www.danvk.org/wp/dragtable'
(19.627327) content/fetchers/curl.c fetch_curl_setup 365: fetch 0x56c661d0, url 
'http://www.danvk.org/wp/dragtable/'
(19.631494) content/fetchers/curl.c fetch_curl_stop 702: fetch 0x56c67698, url 
'http://www.danvk.org/wp/dragtable'
(20.401906) content/fetchers/curl.c fetch_curl_process_headers 1178: HTTP 
status code 200
(20.406648) content/content.c content__init 85: url 
http://www.danvk.org/wp/dragtable/ - 0x56c64fe0
(20.409188) content/content.c content_add_user 576: content 
http://www.danvk.org/wp/dragtable/ (0x56c64fe0), user 0x6efda040 0x590a83a8
***Command 'netsurf' returned with unfreed signals 3FE0!

Re: Request for feedback

[Chris Young]

On 6 Oct 2012 19:19:09 +, Chris Young wrote:

 On Sat, 06 Oct 2012 16:45:03 +0100, John-Mark Bell wrote:
 
  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.
 
 It just appears to terminate without warning, which is odd (and I
 don't see what would cause this?)
 
 I'll try with one of my own builds in case it's something specific to
 CI.

It does appear to be CI-related, as my own build from git is perfectly
happy on that page.

I'm using gcc 4.4.3.

Chris

Re: Request for feedback

[Chris Young]

On Fri, 05 Oct 2012 13:13:20 +0100, Michael Drake wrote:

 Thanks, should be fixed in the latest build.

The latest builds seem to have broken forms.  The core select menu is
now crashing on demand (eg. selecting a language from Google's
advanced search page), and submitting a text field caused a similar
crash (although I haven't checked yet as to whether that is
reproducible).

Chris

Re: Request for feedback

[Chris Young]

On Fri, 05 Oct 2012 14:59:54 +0100, Michael Drake wrote:

  The latest builds seem to have broken forms.  The core select menu is
  now crashing on demand (eg. selecting a language from Google's
  advanced search page),
 
 Should now be fixed.

Yes, working, thanks!

  and submitting a text field caused a similar crash (although I haven't
  checked yet as to whether that is reproducible).
 
 I've not seen that, so any details would be helpful.

My mistake, it's not when submitting, it's when trying to put a CR in.

For example, adding some text to the example text field at the URL
below, and pressing return just freezes here:
http://www.w3schools.com/tags/showit.asp?filename=tryhtml_textarea

Chris

Re: Performance on Debian Linux

[Chris Young]

On Sun, 26 Aug 2012 18:55:07 +0200, Ole wrote:

  Looks like it is not good to try about:blank. It still has network
  access involved:
 
 (0.575378) content/fetchers/curl.c fetch_curl_setup 365: fetch
 0x8cd9458, url 'http://www.google.com/favicon.ico'
 
  Please try about:config.

Unlikely to make any difference; that favicon is fetched due to the
web search provider functionality.

I was meaning to look at caching favicons within URLdb a while back -
I think there may be benefits to that (we can call up favicons for
treeviews for instance).  They'd still need to be refreshed of course,
but I don't think the search provider needs to grab the absolute
latest favicon - the last time it was fetched would be fine.

That would solve this issue anyway.  A disk cache would also solve it,
but I'm not sure if that would allow us to associate bookmarked or
history URLs with favicons?

Chris

Re: Performance on Debian Linux

[Chris Young]

On Fri, 24 Aug 2012 12:09:17 -0700, Wendell P wrote:

 Netsurf 2.9 Debian distribution on Debian Wheezy.
 
 On my machine, Chrome and Firefox each take 3s to boot, while Netsurf
 takes 6s. Is this typical or do you think there is something wrong?

Sounds a bit slow.

 Also, I sometimes get:
 
 Netsurf in running out of memory.
  Please free some memory and try again.
 
 How do you free some memory?

The usual way?  Actually that's a fairly generic error NetSurf throws
up which can be a bit of a red herring.  What are you trying to do
when the error appears?  Is it reproduceable?  In which case, please
provide a test case so we can replicate (or at least see if it is
still an issue on the latest development version).

To diagnose both these issues a log file will be handy (start NetSurf
with nsgtk -v).

Chris

Re: scrolling jerky

[Chris Young]

On Fri, 27 Apr 2012 08:23:18 +0100, Roger Darlington wrote:

 On 25 Apr 2012, Michael Drake  wrote:
  In article be1b188652.roger...@rogerarm.freeuk.com,
 Roger Darlington roger...@freeuk.com wrote:
 
  Scrolling very jerky on the vertical slider
 
  Please see my response to your previous scrolling thread:
 
  http://www.mail-archive.com/netsurf-users@netsurf-browser.org/msg03976.html
 
  Your google URL is the scrolling frames issue.
 
 Thanks for that Michael. I have increased the Cache from 20MB to 90MB, 
 but there is still a very noticeable difference between scrolling the 
 two versions of the very same page (neither of which get anywhere near 
 90MB it must be said...)

Cache won't help, the issue is that core scrolls aren't optimised,
so if you scroll a frame the entire contents of that frame will be
redrawn - even if it is only scrolled a pixel.  Conversely, if you
scroll using the window scrollbar, the platform code handles the
scroll.  Usually the platform code is optimised, and will shift the
area and just redraw the newly-exposed bit.

Clearly NetSurf would benefit from some scrolling optimisation in the
core, but I'm not sure if it is as easy as telling the frontend code
to move a particular area and then redraw the newly exposed area.
(not least because frontends don't currently have any concept of move
a particular area)

Regards
Chris

Re: r13362 'seriour error' bug

[Chris Young]

On Mon, 02 Jan 2012 08:33:20 GMT, Peter Young wrote:

 I'm still using r13355, as none of the builds since then seem to have 
 affected the RISC OS version.

r13358 will affect the RISC OS version too (despite the log notes)

Chris

Re: Problems with NS and Orpheus Squirrel Mail

[Chris Young]

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.

I'll have another go later as I entered via a different path to what I
would normally use.  Might be me being stupid rather than NetSurf not
working.

Chris

Re: Problems with NS and Orpheus Squirrel Mail

[Chris Young]

On Sun, November 6, 2011 9:03 pm, Chris Young wrote:
 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.

 I'll have another go later as I entered via a different path to what I
 would normally use.  Might be me being stupid rather than NetSurf not
 working.

That route definitely works as I just checked with Firefox (I'm typing
this message there).  So I have a different problem with SquirrelMail on
NetSurf, which I'm guessing is cookie-related.

Chris

Re: Search oddity on r13075 (plus at least 1 version prior)

[chris . young]

 In article 498c4e2752.pnyo...@pnyoung.ormail.co.uk,
Dr Peter Young pnyo...@ormail.co.uk wrote:

 However, 13077 says Disable Sprite support -
 DataTypes can't handle these, so could that be fouling something up
for RISC OS?

 It actually says:

 Disable Sprite support - DataTypes can handle these.

 which is rather different!

Yep, it's an Amiga-specific build change affecting the non-RISC OS Sprite
loader.  There is no impact on RISC OS whatsoever.

 Sprites still load.

They do on the Amiga build too, I've just switched over to using the
fallback decoder (DataTypes).  For those who are interested, DataTypes are
an OS plug-in system for various filetypes - the Sprite DataType uses the
same base code as NetSurf's non-RISC OS built-in support.  Basically I'm
saving a few kilobytes by taking it out of the main executable, as it is
not something which is commonly used here.

Chris

Re: Fatal error when page render clashes with a mail/news fetch

[Chris Young]

On Thu, 20 Oct 2011 13:04:30 +0100, Harriet Bazley wrote:

 On 20 Oct 2011 as I do recall,
   cj  wrote:
 
  Has anyone else noticed that if a mail fetch initiates while NetSurf
  is rendering a page, then NS exits with a fatal error?

 What I've observed has been fatal errors reported from Netsurf while
 Firefox is rendering a page, resulting in both apps quitting the iconbar
 (and sometimes FileCore in Use errors).

There is also a crash that occurs when the homepage fails to fetch
(not involving any other non-NetSurf software!).  Due to the
randomness I've not yet been able to get a log nor systematically
reproduce it, but it might be related.

Chris

Re: puzzling email

[Chris Young]

On Tue, 18 Oct 2011 22:35:12 +0100, David J. Ruck wrote:

 On 18/10/2011 19:46, Chris Young wrote:
  On Tue, 18 Oct 2011 09:43:01 +0100, John-Mark Bell wrote:
 
  What is happening is that Pluto fails to rewrite the image reference in
  the HTML document, so NetSurf attempts to fetch a file that doesn't
  exist, and instead builds an error page, which then generates the
  BadType report.
 
  Should NetSurf perhaps be returning a bad image image instead of an
  HTML error, if the request which causes the error to be generated was
  expecting an object of type image/* ?
 
 But if I understand the problem correctly, what is happening here is 
 that with ImageFS active the file ADFS::$.Somewhere.Image/BMP is no 
 longer a file, but a directory containing an sprite file. Therefore 
 NetSurf isn't just getting a broken link or bad image format, but is 
 being given a URL of a directory which is invalid HTML.

I was referring to the case where ImageFS wasn't being used, and
NetSurf was legitimately reading corrupt or missing files.  If you are
running patches that transpose files into directories containing other
files, then quite frankly I'm surprised it works at all.

Chris

Re: puzzling email

[Chris Young]

On Mon, 17 Oct 2011 22:19:56 +0100, Richard Porter wrote:

 On 17 Oct 2011 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.
 
 OK. But an html message or web page shouldn't contain BMP images. I 
 expect that Internet Explorer might open them

Actually BMP files are pretty well supported by web browsers -
presumably mostly because favicons started out using BMP or ICO files.
Wikipedia has a reasonable table although the BMP column is incomplete
(NetSurf entry needs filling in!):
http://en.wikipedia.org/wiki/Comparison_of_web_browsers#Image_format_support

I'd agree that BMP is an odd choice though...

 I'm not surprised that NetSurf barfed at it.

NetSurf's BMP support is very complete/compatible from what I've seen.
It's certainly much better than the rubbish which comes with AmigaOS
:) (fortunately easily replaced with libnsbmp code)

Chris

Re: password for HTTPS site leads to endless loop

[Chris Young]

On Thu, 13 Oct 2011 18:21:14 +0100, Jim Nagel wrote:

 This is an URL file I use -- a doubleclick on it always worked until 
 last week to get directly to the Forwarders page, but now Netsurf 
 tells me it cannot resolve host server6.spellings.net.
 
 https://user:pwd@server6.spellings.net:2083/frontend/x3/mail/addfwd 
 .html
 
 Netsurf r13033 tells me Could not resolve host: server6.spellings.net 
 (Domain name not found).  Yet I can ping that host successfully.

The first digit of the port is getting overwritten by an extraneous
colon for some reason.  It ends up as (I rationalised the
credentials):

https://user:p...@server6.spellings.net::083/frontend/x3/mail/addfwd.html
^
Then it tries to resolve server6.spellings.net: as the hostname.

Chris

Re: Gif rendering - r12786

[Chris Young]

On Thu, 15 Sep 2011 12:19:32 +0100, David Pitt wrote:

 In message a1f60c1352.wra...@wra1th.plus.com
   Gavin Wraith ga...@wra1th.plus.com wrote:
 
  David Pitt has found the reason for the poor display of text in
  the gif. It was down to the css style it was displayed in and
  the consequent rescaling. By removing a line 'width: 40%;'
  which had crept in by copying inappropriately, good behaviour
  has been restored.
 
 I don't think that is quite right. The 'width' setting might have been 
 a web design error but in my view, which may or may not be entirely 
 correct, there is none the less a scaling issue within NetSurf.

Agreed.  I'm using a completely different frontend with different
scaling code, and could still see a couple of the artifacts - so there
is something odd going on there.

Chris

Re: Gif rendering - r12786

[Chris Young]

On Wed, 14 Sep 2011 20:35:03 +0100, Gavin Wraith wrote:

 I recently had occasion to make a scrolling banner for somebody's
 website. I used TextEffX 1.52 to create a sprite (408x56 pixels
 16M colours, no mask, no palette) then SAMP 1.21 to produce more 
 sprites of the same size for the animation, then Intergif 6.18 to
 create the animated gif. The result looked just like the original
 sprite when viewed with FireFox 7 on a Windows XP machine, but
 rather different when viewed with NetSurf r12786 on an Iyonix 
 (RO 5.16). Does this mean a bug in Libnsgif?

A link to the file in question would be incredibly useful if you would
like this checked/fixed!

Chris

Re: HTML Drop-down menu on Netsurf

[Chris Young]

On Fri, 01 Jul 2011 16:44:53 +0100, Tim Hill wrote:

 I guess we will need to write a pop-out or drop-down menus in the bits of
 CSS that NetSurf does understand but I suspect it isn't possible and it
 needs the bits it doesn't do (yet).

It won't work without dynamic layout changes, currently scheduled for
NetSurf 4.0.

Chris

Re: NetSurf for Mac OS X

[Chris Young]

   I have just tried NetSurf 2.7 for Mac and it downloads and
   unzips but won't run. The error message is  Base style sheet
   failed to load. It then quits.

For info, the base stylesheet being refered to is
resource:default.css.  I'm not sure where resource: searches on Mac
OS, but there should be a res or resources directory where you
installed NetSurf with that file in.  If it is in there and readable,
check it is filetyped to CSS with an associated MIME type of text/css.
 If NetSurf can't find the file, can't read it, or thinks it is
something other than text/css then it outputs the base stylesheet
error.

I'm not au fait with how this all fits together on Mac OS or the Mac
OS version of NetSurf, and the source code in those parts may as well
be heiroglyhs :)  It does look like there is a bit of logging though,
so if you can run NetSurf with a -v switch then it will output a load
of text which will possibly point to where the problem lies.  Redirect
that to a file and post it somewhere and we'll have a look at it.

My hunch (and the most likely cause) is that the MIME type of
default.css is not being detected correctly.

Chris

Re: Cookie deletion issues

[Chris Young]

On Wed, 04 May 2011 22:12:12 +0100, Martin Bazley wrote:

 Can anyone offer advice or replications on other platforms?  Is it a
 bug?

I've just tested on AmigaOS and the deleted cookie jumping to the
bottom of the list issue occurs here too, so it looks like a core
bug.  I've not experienced any flickering on any of the treeviews, so
I think that one is RISC OS specific.

Chris