|
NetSurf 2.6 Released |
|
nunfetishist (12:13 23/9/2010) swirlythingy (17:03 23/9/2010) nunfetishist (17:48 23/9/2010) ad (20:42 23/9/2010) nunfetishist (22:46 23/9/2010) Phlamethrower (22:58 23/9/2010) nunfetishist (09:02 24/9/2010) trevj (09:50 24/9/2010) nunfetishist (09:56 24/9/2010) Monty (14:46 24/9/2010) jmb (18:40 24/9/2010) Acornut (08:05 25/9/2010) nunfetishist (11:04 25/9/2010) stevef (23:02 25/10/2010) CJE (11:56 26/10/2010) stevef (18:24 26/10/2010) swirlythingy (22:15 26/10/2010) nunfetishist (22:30 26/10/2010) swirlythingy (12:42 27/10/2010) swirlythingy (13:20 27/10/2010) nunfetishist (15:52 27/10/2010) swirlythingy (16:52 27/10/2010) swirlythingy (17:15 27/10/2010) arawnsley (23:09 27/10/2010) swirlythingy (16:39 28/10/2010) arawnsley (16:58 28/10/2010) flibble (17:25 28/10/2010) swirlythingy (07:02 29/10/2010) Lampi (23:54 28/10/2010) swirlythingy (07:01 29/10/2010)
|
|
Rob Kendrick |
Message #115446, posted by nunfetishist at 12:13, 23/9/2010 |
Today's phish is trout a la creme.
Posts: 525
|
The NetSurf developers are happy to announce the immediate availability of NetSurf 2.6. This release contains many bug fixes and improvements.
It is available to download from http://www.netsurf-browser.org/
Here is a change log detailing the important changes in this release:
Core / All ----------
* Improvements to float positioning. * Fix absolute positioned inlines. * Improve handling of percentage margins and paddings on floats. * Fix several memory leaks. * Rationalised memory allocation functions. * Updated SSL root certificates. * Fix handling of PNGs with zero data. * Moved input handling into content handlers. * Simplified browser window module. * Fix table borders specified in em/ex units. * Improved CSS content handling. * Don't try to draw zero size images. * Simplified and fixed save complete feature. * Reduce frequency of cache clean attempts. * Improve and optimise rendering of borders in HTML. * Fix URL comparison issue. * Improve object handling in HTML. * Fix absolute positioned root element. * Rewritten file: fetch handler for local URLs. * Much improved directory listings for local file: paths. * Improve handling of binary data in plain text renderer. * Fixed cache expiry problems.
* LibCSS library (CSS parser and selection engine): + Allow stylesheets to be used in multiple contexts simultaneously. + Fix possible stylesheet reuse crash.
RISC OS-specific ----------------
* Increased maximum allowed value for memory cache size option. * Update 'about' page to include licences for linked components. * Fix layout to paper width when printing. * Fix screen redraw after printing. * Enable sideways print option.
GTK-specific ------------
* Avoid potential use of NULL pointer. * Fix for old versions of GTK. * Improve internationalisation support. * Improve handling of bitmap opacity. * Fix redraw issues with non-HTML content. * Cleaned up menu generation and handling.
BeOS/Haiku-specific -------------------
* Fix build. * Fix mimetype acquisition for local files. * Improve handling of bitmap opacity. * Enable knockout rendering optimisation. * Improved polygon plotter accuracy. * Fix colours on copy to clipboard. |
|
[ Log in to reply ] |
|
Martin Bazley |
Message #115448, posted by swirlythingy at 17:03, 23/9/2010, in reply to message #115446 |
Posts: 460
|
Why is this in the Playpen? |
|
[ Log in to reply ] |
|
Rob Kendrick |
Message #115450, posted by nunfetishist at 17:48, 23/9/2010, in reply to message #115448 |
Today's phish is trout a la creme.
Posts: 525
|
Why is this in the Playpen? Why not? A quick glance over the fora choices didn't reveal an "announcements", or even "software". |
|
[ Log in to reply ] |
|
Andrew Duffell |
Message #115454, posted by ad at 20:42, 23/9/2010, in reply to message #115450 |
Posts: 3262
|
news[at]iconbar.com ? |
|
[ Log in to reply ] |
|
Rob Kendrick |
Message #115455, posted by nunfetishist at 22:46, 23/9/2010, in reply to message #115454 |
Today's phish is trout a la creme.
Posts: 525
|
news[at]iconbar.com ? I had assumed, given the release was some days ago, and posted to our mailing list and usenet, that expecting a TIB article update would be silly |
|
[ Log in to reply ] |
|
Jeffrey Lee |
Message #115456, posted by Phlamethrower at 22:58, 23/9/2010, in reply to message #115455 |
Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot Hot stuff
Posts: 15100
|
news[at]iconbar.com ? I had assumed, given the release was some days ago, and posted to our mailing list and usenet, that expecting a TIB article update would be silly "some days ago" - you mean yesterday? |
|
[ Log in to reply ] |
|
Rob Kendrick |
Message #115458, posted by nunfetishist at 09:02, 24/9/2010, in reply to message #115456 |
Today's phish is trout a la creme.
Posts: 525
|
"some days ago" - you mean yesterday? I think you'll find that's some |
|
[ Log in to reply ] |
|
Trevor Johnson |
Message #115459, posted by trevj at 09:50, 24/9/2010, in reply to message #115458 |
Member
Posts: 660
|
"some days ago" - you mean yesterday? I think you'll find that's some "some time ago"? |
|
[ Log in to reply ] |
|
Rob Kendrick |
Message #115460, posted by nunfetishist at 09:56, 24/9/2010, in reply to message #115459 |
Today's phish is trout a la creme.
Posts: 525
|
"some days ago" - you mean yesterday? I think you'll find that's some "some time ago"? One uses the plural form for all numbers excluding 1 and -1. |
|
[ Log in to reply ] |
|
Edward Rogers |
Message #115461, posted by Monty at 14:46, 24/9/2010, in reply to message #115460 |
Member
Posts: 154
|
Thank you Rob & Netsurf team in general. What is the situation with the endangered RISC OS front-end at the moment? |
|
[ Log in to reply ] |
|
JMB |
Message #115462, posted by jmb at 18:40, 24/9/2010, in reply to message #115461 |
Member
Posts: 467
|
Thank you Rob & Netsurf team in general. What is the situation with the endangered RISC OS front-end at the moment? Precisely the same as it was in April. |
|
[ Log in to reply ] |
|
Blind Moose |
Message #115464, posted by Acornut at 08:05, 25/9/2010, in reply to message #115461 |
No-eye-deer (No Idea)
Posts: 487
|
So this has nothing to do with it then? |
|
[ Log in to reply ] |
|
Rob Kendrick |
Message #115468, posted by nunfetishist at 11:04, 25/9/2010, in reply to message #115464 |
Today's phish is trout a la creme.
Posts: 525
|
So this has nothing to do with it then? Steve spent some time looking at it, sure. And has even done some work on it. But we've not actually heard from him in ages, probably because he has more important things in his life (don't we all?). Thus, the situation has not changed. |
|
[ Log in to reply ] |
|
Steve Fryatt |
Message #115752, posted by stevef at 23:02, 25/10/2010, in reply to message #115468 |
Member
Posts: 18
|
Just spotted this, a little late, and thought perhaps a comment might be useful...
Steve spent some time looking at it, sure. And has even done some work on it. And it even managed to survive "that" branch being merged in a couple of weeks ago without falling over completely...
The situation from here is fairly simple: I've looked at the code, done some work on it, and will continue to do more if/when time allows (and the observant amongst you may have spotted that there were changes made to the RISC OS frontend yesterday). I know roughly what still needs to be done treeview-wise, and also have some other ideas that I'd like to implement should I get chance.
If anyone (further up this thread) is looking for a more cast-iron guarantee along the lines of "NetSurf's future for RISC OS is guaranteed", then you're in the wrong market. No-one in their right mind would ever make such a guarantee, particularly not for free software that's developed "for fun". The best you're going to get is that at present, it still hasn't stopped building for RISC OS (not actually so small an achievement, given that we've passed the original hurdle now with only relatively minor breakage). And the source code won't just vanish, anyway.
But we've not actually heard from him in ages Oops... I probably need to visit #netsurf again at some point. I am still around, though -- honest. |
|
[ Log in to reply ] |
|
Chris Evans |
Message #115759, posted by CJE at 11:56, 26/10/2010, in reply to message #115752 |
CJE Micros chap
Posts: 228
|
Thanks for the update Steve. I may have got a hold of the wrong end of the stick, but I'm sure I was told or read that quite a few things were expected to break when certain build changes were made, so it is good to hear "only relatively minor breakage" has taken place, Is what the 'minor breakage' is, documented anywhere in an end user understandable way? Or can you explain it here? |
|
[ Log in to reply ] |
|
Steve Fryatt |
Message #115764, posted by stevef at 18:24, 26/10/2010, in reply to message #115759 |
Member
Posts: 18
|
I may have got a hold of the wrong end of the stick, but I'm sure I was told or read that quite a few things were expected to break when certain build changes were made That was the case when the announcements were made: when I started playing with the RISC OS front-end at the start of the year, the treeview branch didn't even build for our platform. The fact that in the end it didn't break so badly doesn't mean that the announcements were wrong -- just that later on, someone did some work to reduce the breakage.
Is what the 'minor breakage' is, documented anywhere in an end user understandable way? Or can you explain it here? There's a fairly user-hostile list (that's really for the benefit of anyone working on the front-end) at http://wiki.netsurf-browser.org/Todo/RISC_OS_frontend
The more friendly summary is "various bits and pieces of the GUI in and around the Hotlist, Global History and Cookies windows". It's still usable (mostly), but "different". |
|
[ Log in to reply ] |
|
Martin Bazley |
Message #115766, posted by swirlythingy at 22:15, 26/10/2010, in reply to message #115764 |
Posts: 460
|
Speaking of treeview, I feel I must mention this bug report I recently filed. (I submitted it to SourceForge as the developers tend to make sarcastic comments when you don't, but since you're here and talking about breakage in treeview I might as well point it out.) |
|
[ Log in to reply ] |
|
Rob Kendrick |
Message #115767, posted by nunfetishist at 22:30, 26/10/2010, in reply to message #115766 |
Today's phish is trout a la creme.
Posts: 525
|
I submitted it to SourceForge as the developers tend to make sarcastic comments when you don't And I'm sure the wording of that bug report is somehow meant to make the developers more hospitable to you. Although I am unsure how.
If you find something you consider urgent, you know where both the source code and previous versions are.
I can tell you with complete confidence that your issue is not urgent to us. |
|
[ Log in to reply ] |
|
Martin Bazley |
Message #115769, posted by swirlythingy at 12:42, 27/10/2010, in reply to message #115767 |
Posts: 460
|
And I'm sure the wording of that comment is somehow meant to make the users more hospitable to submitting bug reports. Although I am unsure how.
I could download the entire source tree, if you like... but surely it's not that much effort to remove two characters from the 'must be encoded' list?
I can tell you with complete confidence that your issue is not urgent to us. 'Us' being...? |
|
[ Log in to reply ] |
|
Martin Bazley |
Message #115770, posted by swirlythingy at 13:20, 27/10/2010, in reply to message #115769 |
Posts: 460
|
OK, so I'm no C programmer, but I believe what needs to happen is this:
In desktop/tree_url_node.c, lines 403 and 404 (in function tree_url_node_callback):
res = url_escape(text, 0, false, NULL, &escaped_text);
needs to change to:
res = url_escape(text, 0, false, ":/", &escaped_text);
Would his majesty like me to commit it to SVN too? |
|
[ Log in to reply ] |
|
Rob Kendrick |
Message #115771, posted by nunfetishist at 15:52, 27/10/2010, in reply to message #115769 |
Today's phish is trout a la creme.
Posts: 525
|
I could download the entire source tree, if you like... but surely it's not that much effort to remove two characters from the 'must be encoded' list? The entire source tree's pretty small for such a project. I'm sure it's not beyond your grasp.I can tell you with complete confidence that your issue is not urgent to us. 'Us' being...? Anybody who cares and has commit access. You are, and always have been, welcome to formulate a patch and submit it to the developer list for review. All the snide comments you've made have done is make us even less interested in fixing your problems. |
|
[ Log in to reply ] |
|
Martin Bazley |
Message #115772, posted by swirlythingy at 16:52, 27/10/2010, in reply to message #115771 |
Posts: 460
|
I could download the entire source tree, if you like... but surely it's not that much effort to remove two characters from the 'must be encoded' list? The entire source tree's pretty small for such a project. I'm sure it's not beyond your grasp. Well, after writing that post (and before writing the next one, which you totally ignored) I did indeed download the source tree, only to discover that it was compressed in a format not supported by RISC OS. You're really doing your best to discourage this, aren't you...?
I could have checked out via svn, except that would have required me to (a) read through the instructions on how to do so, (b) install svn, and (c) wait for hours while it chugged through the checkout process, all to get access to a couple of files which I am unlikely ever to want or need again.
I worked out what was needed by downloading likely-looking files from the web interface, establishing which bits did what and following the #include links to establish the source of the escape function call. Not bad, I think, for someone who has never programmed in C before.
I can tell you with complete confidence that your issue is not urgent to us. 'Us' being...? Anybody who cares and has commit access. Ah, "us" == "me". Thought so.
You are, and always have been, welcome to formulate a patch and submit it to the developer list for review. That would be the one I'm not subscribed to and am unlikely ever to have cause to post to again, then. And I've never used IRC in my life.
For heaven's sake, it's a four character change, to fix a major bug, and you're the one with commit access. (Excluding, of course, all the other developers on this forum, who are likely to be more helpful.)
All the snide comments you've made have done is make us even less interested in fixing your problems. I think I can feel a snide comment about a pot and a kettle coming on... |
|
[ Log in to reply ] |
|
Martin Bazley |
Message #115773, posted by swirlythingy at 17:15, 27/10/2010, in reply to message #115772 |
Posts: 460
|
EDIT: Now resolved.
Oh, well, isn't that utterly bloody charming? Messenger Pro 6.00 contains a new bug which causes any attempt to create a new mailing list - under any circumstances - to fail with a 'Messenger is running out of memory' error. Because my two main Internet applications obviously weren't nearly broken enough already.
So, long story short, doesn't look like I'm subscribing to the development list any time soon.
[Edited by swirlythingy at 19:04, 27/10/2010] |
|
[ Log in to reply ] |
|
Andrew Rawnsley |
Message #115774, posted by arawnsley at 23:09, 27/10/2010, in reply to message #115773 |
R-Comp chap
Posts: 600
|
Phew, had me worried there for a minute Martin! Was about to send you your second lot of "source tree" of the day! Glad you fixed whatever it was (would appreciate an email if there's something I could tweak to avoid other users having the same problem). |
|
[ Log in to reply ] |
|
Martin Bazley |
Message #115776, posted by swirlythingy at 16:39, 28/10/2010, in reply to message #115774 |
Posts: 460
|
Turned out I misunderstood the 'Folder' field in the 'Add list' dialogue - I assumed it was a pathname, not a filename, so left it blank and put the leafname in the 'Name' field. The thing which convinced me this was an actual bug and not merely PEBKAC was that clicking the action button produced the unexpected error "Messenger has run out of memory", which was not in any way relevant to the problem - I assumed something had gone wrong with a list expansion routine somewhere!
Would you really have sent me a source tree? Seems a bit drastic, specially given Messenger's status as a commercial product!
PS: If you're looking at that bit of code, it would be nice to amend it so it automatically creates a folder if it doesn't already exist. |
|
[ Log in to reply ] |
|
Andrew Rawnsley |
Message #115777, posted by arawnsley at 16:58, 28/10/2010, in reply to message #115776 |
R-Comp chap
Posts: 600
|
Well, I have a feeling opening the source up to users who have skills could well be beneficial in the future. It's just a case of how best to balance a commercial application with a more open attitude to source. I believe that's an issue for many in the software industry these days. |
|
[ Log in to reply ] |
|
Peter Howkins |
Message #115779, posted by flibble at 17:25, 28/10/2010, in reply to message #115777 |
Posts: 892
|
Yes, discovering how to find people to program for free so you can keep selling things is definitely an area that affects many in the RISC OS software industry. |
|
[ Log in to reply ] |
|
James Lampard |
Message #115781, posted by Lampi at 23:54, 28/10/2010, in reply to message #115772 |
Posts: 190
|
... I did indeed download the source tree, only to discover that it was compressed in a format not supported by RISC OS. Which format is it? |
|
[ Log in to reply ] |
|
Martin Bazley |
Message #115783, posted by swirlythingy at 07:01, 29/10/2010, in reply to message #115781 |
Posts: 460
|
Which format is it? I used the tarball link, which produced an error when I tried to extract it. I later discovered, quite by chance, having for some reason decided to start reading the NetSurf development wiki, and having just happened to glance at the GSoC introduction page on a whim, where it was briefly mentioned in passing, that it actually was broken. |
|
[ Log in to reply ] |
|
Martin Bazley |
Message #115784, posted by swirlythingy at 07:02, 29/10/2010, in reply to message #115777 |
Posts: 460
|
Well, I have a feeling opening the source up to users who have skills could well be beneficial in the future. Just as well you didn't waste your time on me, then. |
|
[ Log in to reply ] |
|
|