4:22 p.m.
OK another small tidy along the road to making the plotter API less mad
This one removes the clg call altogether. It was used in four places,
all of which set the clip rectangle just before clg was called!
This patch replaces these four calls with their direct equivalent fill
calls and removes the clg call from the plotters. Removes another
hundred lines of redundant code and makes the plotter API simpler.
amiga/plotters.c | 1 -
beos/beos_plotters.cpp | 23 -----------------------
desktop/knockout.c | 16 +---------------
desktop/plotters.h | 2 --
desktop/save_pdf/pdf_plotters.c | 7 -------
framebuffer/framebuffer.c | 1 -
gtk/gtk_plotters.c | 7 -------
gtk/gtk_print.c | 7 -------
render/html_redraw.c | 2 +-
render/textplain.c | 2 +-
riscos/plotters.c | 20 --------------------
riscos/print.c | 9 +++------
riscos/save_draw.c | 11 -----------
riscos/window.c | 2 +-
14 files changed, 7 insertions(+), 103 deletions(-)
--
Regards Vincent
http://www.kyllikki.org/
5:55 p.m.
In article <20090630202211.GA18658@derik>,
Vincent Sanders <vince(a)kyllikki.org> wrote:
This one removes the clg call altogether. It was used in four
places,
all of which set the clip rectangle just before clg was called!
Tested on RISC OS, seems to work fine.
Michael
--
Michael Drake (tlsa) http://www.netsurf-browser.org/
6:18 p.m.
I know I call ami_clg in a couple of places manually, IIRC the reason
was that when clicking on a link to certain content (SVG, I'm sure
there were others too) didn't clear the previous page so you ended up
with the SVG superimposed on top of the previous page.
I can easily replace these calls, but there may be another issue there
that needs looking at.
Chris
6:39 p.m.
On Tue, 2009-06-30 at 23:18 +0100, Chris Young wrote:
I know I call ami_clg in a couple of places manually, IIRC the
reason
was that when clicking on a link to certain content (SVG, I'm sure
there were others too) didn't clear the previous page so you ended up
with the SVG superimposed on top of the previous page.
I can easily replace these calls, but there may be another issue there
that needs looking at.
I guess there's 2 options here:
1) Expect that the frontend clears the window contents when a complete
redraw is performed.
2) The relevant content redraw functions clear the background of the
area they're going to draw.
Either way, I think this is orthogonal to removing the CLG hook.
Thoughts?
J.
1:36 p.m.
On Tue, 30 Jun 2009 23:39:33 +0100, John-Mark Bell wrote:
On Tue, 2009-06-30 at 23:18 +0100, Chris Young wrote:
> I know I call ami_clg in a couple of places manually, IIRC the reason
> was that when clicking on a link to certain content (SVG, I'm sure
> there were others too) didn't clear the previous page so you ended up
> with the SVG superimposed on top of the previous page.
>
> I can easily replace these calls, but there may be another issue there
> that needs looking at.
I guess there's 2 options here:
1) Expect that the frontend clears the window contents when a complete
redraw is performed.
2) The relevant content redraw functions clear the background of the
area they're going to draw.
Either is fine by me. Option 2 may be more sensible as it takes
responsibility away from the frontend.
Chris
4:33 p.m.
New subject: render/html.c html_convert 450: imagemap extraction failed error
Hello
i get this error only when want show the BBC page(link on startsite)
after that come a out of memory message.To find the correct line i do this
call into debugger.
const char *messages_get(const char *key)
{
if (strcmp(key,"NoMemory")==0){STOP}
return messages_get_ctx(key, messages_hash);
}
so i am sure its the right place.memory is enough here.the build have no MNG
support.Is this maybe a problem ?
I dont know what type of image that is.see log file of last lines before
error.
'http://news.bbc.co.uk/sol/shared/img/timezone/z.zzz'
please help whats wrong.netsurf do not crash but it dont show the page.i
have update render dir to newest files but problem is same.
content/fetchcache.c fetchcache_go 249: url
http://newsimg.bbc.co.uk/shared/img/v4/palette/orange.gif, status
TYPE_UNKNOWN
content/fetchcache.c fetchcache 143: url
http://newsimg.bbc.co.uk/shared/img/v4/palette/orange.gif
content/content.c content_add_user 1122: content
http://newsimg.bbc.co.uk/shared/img/v4/palette/orange.gif (0x11a19f28),
user
0x10d55d94 0x11604360 0xc6
content/fetchcache.c fetchcache_go 249: url
http://newsimg.bbc.co.uk/shared/img/v4/palette/orange.gif, status
TYPE_UNKNOWN
content/fetchcache.c fetchcache 143: url
http://newsimg.bbc.co.uk/shared/img/v4/panel_gradient.gif
content/content.c content_create 404: url
http://newsimg.bbc.co.uk/shared/img/v4/panel_gradient.gif -> 0x11c27888
content/content.c content_add_user 1122: content
http://newsimg.bbc.co.uk/shared/img/v4/panel_gradient.gif (0x11c27888),
user
0x10d55d94 0x11604360 0xc7
content/fetchcache.c fetchcache_go 249: url
http://newsimg.bbc.co.uk/shared/img/v4/panel_gradient.gif, status
TYPE_UNKNOWN
content/fetchers/fetch_curl.c fetch_curl_setup 324: fetch 0x11c27d58, url
'http://newsimg.bbc.co.uk/shared/img/v4/panel_gradi
ent.gif'
content/fetchcache.c fetchcache 143: url
http://newsimg.bbc.co.uk/shared/img/v4/footer_blocks_grey.gif
content/content.c content_create 404: url
http://newsimg.bbc.co.uk/shared/img/v4/footer_blocks_grey.gif -> 0x11c3b088
content/content.c content_add_user 1122: content
http://newsimg.bbc.co.uk/shared/img/v4/footer_blocks_grey.gif (0x11c3b088),
user 0x10d55d94 0x11604360 0xc8
content/fetchcache.c fetchcache_go 249: url
http://newsimg.bbc.co.uk/shared/img/v4/footer_blocks_grey.gif, status
TYPE_UNKNOW
N
content/fetchers/fetch_curl.c fetch_curl_setup 324: fetch 0x11c3b658, url
'http://newsimg.bbc.co.uk/shared/img/v4/footer_bloc
ks_grey.gif'
content/fetchcache.c fetchcache 143: url
http://news.bbc.co.uk/sol/shared/img/timezone/z.zzz
content/content.c content_create 404: url
http://news.bbc.co.uk/sol/shared/img/timezone/z.zzz -> 0x11c3d488
content/content.c content_add_user 1122: content
http://news.bbc.co.uk/sol/shared/img/timezone/z.zzz (0x11c3d488), user
0x10d
55d94 0x11604360 0xc9
content/fetchcache.c fetchcache_go 249: url
http://news.bbc.co.uk/sol/shared/img/timezone/z.zzz, status TYPE_UNKNOWN
content/fetchers/fetch_curl.c fetch_curl_setup 324: fetch 0x11c3d958, url
'http://news.bbc.co.uk/sol/shared/img/timezone/z.zz
z'
render/html.c html_convert 450: imagemap extraction failed
Regards
7:29 p.m.
New subject: render/html.c html_convert 450: imagemap extraction failed error
On Thu, 2009-07-02 at 22:33 +0200, Bernd Roesch wrote:
Hello
i get this error only when want show the BBC page(link on startsite)
after that come a out of memory message.To find the correct line i do this
call into debugger.
If you have a debugger, presumably you can single-step imagemap_extract
to find out what fails?
const char *messages_get(const char *key)
{
if (strcmp(key,"NoMemory")==0){STOP}
return messages_get_ctx(key, messages_hash);
}
so i am sure its the right place.memory is enough here.
There may be sufficient free memory, but if the allocation calls fail,
NetSurf doesn't know this. The imagemap functions fail *only* when
malloc/calloc/realloc return NULL.
the build have no MNG support.Is this maybe a problem ?
Disabling support for specific content types won't affect this.
I've tidied up the imagemap code and stopped it trying to call
calloc(0, ...), which was probably your problem.
J.
8:50 a.m.
New subject: render/html.c html_convert 450: imagemap extraction failed error
Hello John-Mark
On 02.07.09, you wrote:
I've tidied up the imagemap code and stopped it trying to call
calloc(0, ...), which was probably your problem.
Yes thanks, calloc 0 was the problem.your new file work and also when i
change the 0 to 1 in to places in old file work.
there is calloc use from netbsd.and this fail on 0.but i am not sure what
unix specification say to this.if all Unix accept 0, i change AOS 68k lib
to be compatible.
* Copyright (c) 1990 The Regents of the University of California.
* All rights reserved.
.....
#if defined(LIBC_SCCS) && !defined(lint)
static char sccsid[] = "(a)(#)calloc.c 5.5 (Berkeley) 5/17/90";
#endif /* LIBC_SCCS and not lint */
#define _KERNEL
#include "ixemul.h"
#include <stdlib.h>
#include <string.h>
void *
calloc (size_t num, size_t size)
{
register void *p;
size *= num;
if ((p = (void *) syscall (SYS_malloc, size)))
bzero(p, size);
return(p);
}
void
cfree(void *p)
{
(void)syscall (SYS_free, p);
}
J.
Regards
9:39 a.m.
New subject: render/html.c html_convert 450: imagemap extraction failed error
On Fri, 2009-07-03 at 14:50 +0200, Bernd Roesch wrote:
Hello John-Mark
On 02.07.09, you wrote:
>
> I've tidied up the imagemap code and stopped it trying to call
> calloc(0, ...), which was probably your problem.
Yes thanks, calloc 0 was the problem.your new file work and also when i
change the 0 to 1 in to places in old file work.
there is calloc use from netbsd.and this fail on 0.but i am not sure what
unix specification say to this.if all Unix accept 0, i change AOS 68k lib
to be compatible.
It's perfectly acceptable for malloc/calloc to return NULL if asked for
an allocation of zero bytes.
J.
6:37 p.m.
On Tue, 2009-06-30 at 21:22 +0100, Vincent Sanders wrote:
OK another small tidy along the road to making the plotter API less
mad
This one removes the clg call altogether. It was used in four places,
all of which set the clip rectangle just before clg was called!
This patch replaces these four calls with their direct equivalent fill
calls and removes the clg call from the plotters. Removes another
hundred lines of redundant code and makes the plotter API simpler.
Fine by me. I'll sync the plotter wiki page.
J.
5088
days inactive
5091
days old
netsurf-dev@netsurf-browser.org
9 comments
5 participants
participants (5)
-
Bernd Roesch -
Chris Young -
John-Mark Bell -
Michael Drake -
Vincent Sanders