4:37 a.m.
Hi
I downloaded and built the NetSurf on Fedora 8 after clearing out a few
problems with help of rjek and jmb. After which I read through the
layout.ccode and understood how it works.. the structures, functions
used etc..
Now that I have gained familiarity with the Layout code and line breaking
algorithm, I am positive that I can successfully handle the issues with
borders, margins, padding, backgrounds, css support. I would like to know
what is expected out of this project.
So that I can look into those issues and disover ways to overcome the
problems.
And it would be great if I am provided with a list of url test cases to test
my code and see to it.. that the issues are solved.. I have started working
on it. I later plan to run the NetSurf on the RPCemu emulator with RISC OS 3
review the work done.
--
Yours
Manjeet
12:43 p.m.
On Wed, 2008-03-26 at 09:07 +0530, Manjeet Chayel wrote:
And it would be great if I am provided with a list of url test cases
to test my code and see to it.. that the issues are solved.. I have
started working on it. I later plan to run the NetSurf on the RPCemu
emulator with RISC OS 3 review the work done.
I have no real knowledge of the layout system, so you'll have to wait
for John-Mark or James to reply, I suspect. However, when using RPCemu,
you should at least try using RISC OS 3.7 at minimum - anything older is
supported by accident, I suspect.
Additionally, I can't condone using ROM images for RPCemu from anything
other than a real RISC OS box that you own.
In a few days, an "open-access" box will be available for accepted
students to use for testing purposes.
B.
12:28 a.m.
On Wed, 26 Mar 2008 09:07:11 +0530, "Manjeet Chayel"
<manjeet.chayel(a)gmail.com> wrote:
Now that I have gained familiarity with the Layout code and line breaking
algorithm, I am positive that I can successfully handle the issues with
borders, margins, padding, backgrounds, css support. I would like to know
what is expected out of this project.
I would be looking to see a sizeable number of the issues with the current
implementation fixed. Michael is best-placed to give a full run-down of all
the issues with inlines. We have test cases for a fair number of the
issues, though by no means all. You might like to start by creating a few
simple test cases for the issues we don't have tests for and investigating
how they're going wrong.
I seem to recall pointing you in the direction of layout_inline_container()
already (this is pretty well the top-level of inline handling, though
inline blocks complicate matters slightly). It's definitely the first place
to look in the code when working out how inline boxes are laid out.
I would also advise that you avoid the issue of the Unicode line-breaking
algorithm until you've got better knowledge of the existing code. There's
plenty to be doing in the meantime.
So that I can look into those issues and disover ways to overcome
the
problems. And it would be great if I am provided with a list of url test
cases to
test my code and see to it.. that the issues are solved.. I have
started
working on it.
Michael should know which testcases apply here. They should be in the
"haveproblems" directory of "netsurftest". Do "svn co
svn://source.netsurf-browser.org/trunk/netsurftest" to get this.
HTH,
John.
3:56 p.m.
In article <263f02df0803252037h2adfbd82nde60358363d1f8cd(a)mail.gmail.com>,
Manjeet Chayel <manjeet.chayel(a)gmail.com> wrote:
Hi
I downloaded and built the NetSurf on Fedora 8 after clearing out a
few
problems with help of rjek and jmb.
Excellent.
After which I read through the layout.ccode and understood how it
works.. the structures, functions used etc.. Now that I have gained
familiarity with the Layout code and line breaking algorithm, I am
positive that I can successfully handle the issues with borders,
margins, padding, backgrounds, css support.
OK, great. You should become familiar with the relevant parts of the
CSS 2.1 specification. Do you have much CSS/HTML experience already?
I would like to know what is expected out of this project.
The project will involve the following:
1. Understanding the HTML layout and display code. (layout.c and
html_redraw.c)
2. Generating test cases for all the problems that NetSurf currently
exhibits. This will involve looking at what the specification says,
testing how NetSurf behaves and also testing how other standards
compliant browsers behave. (E.g. Firefox, Opera & Safari/Konqueror).
3. Updating the code to fix the issues that have been identified.
4. Checking for any regressions brought about by code changes. This will
require quite a good test suite to be created in "2".
5. It would be good to document the layout and line breaking code more
explicitly.
So that I can look into those issues and disover ways to overcome
the
problems. And it would be great if I am provided with a list of url test
cases to test my code and see to it.. that the issues are solved.. I
have started working on it.
There are already several test cases for inline issues in netsurftest/ in
the SVN repository. Off the top of my head, the main problems we have at
the moment include:
+ Inlines with borders have broken and disjointed borders when there is
also padding present.
+ Vertical alignment is currently unsupported. Normally baselines should
be aligned. We also need subscript and superscript vertical alignment.
Images as well as text need to be vertically aligned.
+ There is some issue with incorrect line heights. I think it has
something to do with font geometry or various font sizes being used. It
can cause lines of text to overlap. John-Mark (jmb) and James (zamez)
know more about this.
+ Background colours for inlines do not fill the full area when there is
padding present.
+ Background image position is not correctly supported for inlines. Sites
such as Wikipedia place non-repeating images in the right padding of
link text. NetSurf currently places the background image at the left hand
side of the inline, under the text.
+ The white-space property is not supported. At the moment we can't do
no-wrap and stuff.
+ NetSurf's current line breaking goes wrong for situations like:
<span>1<sup>st</sup></span>. NetSurf will currently break the
line
between "1" and "st", but it shouldn't, because there is no
space in
the source.
+ Text that is inside a block, with no inline children around the text,
incorrectly inherit the background colour of the block. I made a test
case for this the other day:
http://www.smoothartist.com/netsurf/inline-background.html
I'll check that into NetSurf's SVN later.
There are more issues which I haven't mentioned. However, fixing the
background colour / image issues and the border problem will make a big
improvement. These often cause NetSurf's display to look untidy and ragged.
Implementing the white-space and vertical-align properties are also very
clear-cut goals.
There are already test cases for some of the issues above in SVN.
I later plan to run the NetSurf on the RPCemu emulator with RISC
OS 3 review the work done.
I would not worry about this. The layout engine is part of the platform
independent core, so it will behave the same on RISC OS as on GTK NetSurf.
There may be a GTK NetSurf specific line height bug, but I wouldn't bother
about that for now.
Also, we plan to get a RISC OS box set up, that GSoC students will be able
to use remotely. This is nearly ready. It will show you what the RISC OS
version's features are (bookmarks, downloads, etc) which aren't present on
GTK NetSurf.
Michael
--
Michael Drake (tlsa) http://www.netsurf-browser.org/
12:56 p.m.
On Sat, Mar 29, 2008 at 8:26 PM, Michael Drake <tlsa(a)netsurf-browser.org>
wrote:
In article
<263f02df0803252037h2adfbd82nde60358363d1f8cd(a)mail.gmail.com>,
Manjeet Chayel <manjeet.chayel(a)gmail.com> wrote:
> Hi
> I downloaded and built the NetSurf on Fedora 8 after clearing out a few
> problems with help of rjek and jmb.
Excellent.
> After which I read through the layout.ccode and understood how it
> works.. the structures, functions used etc.. Now that I have gained
> familiarity with the Layout code and line breaking algorithm, I am
> positive that I can successfully handle the issues with borders,
> margins, padding, backgrounds, css support.
OK, great. You should become familiar with the relevant parts of the
CSS 2.1 specification. Do you have much CSS/HTML experience already?
I have been doing projects in CSS/HTML for past 2 years. I think thats quite
a experience :)
> I would like to know what is expected out of this project.
The project will involve the following:
1. Understanding the HTML layout and display code. (layout.c and
html_redraw.c)
Have looked throught the code.. understood what the modules do.. how they
interact. the structures used etc.
2. Generating test cases for all the problems that NetSurf currently
exhibits. This will involve looking at what the specification says,
testing how NetSurf behaves and also testing how other standards
compliant browsers behave. (E.g. Firefox, Opera & Safari/Konqueror).
Using Netsurf for browsing. Which gives helps out in determing the behaviour
of netsurf on various scenrios. And developing some test cases to for
current problems..
3. Updating the code to fix the issues that have been identified.
4. Checking for any regressions brought about by code changes. This will
require quite a good test suite to be created in "2".
5. It would be good to document the layout and line breaking code more
explicitly.
> So that I can look into those issues and disover ways to overcome the
> problems. And it would be great if I am provided with a list of url test
> cases to test my code and see to it.. that the issues are solved.. I
> have started working on it.
There are already several test cases for inline issues in netsurftest/ in
the SVN repository. Off the top of my head, the main problems we have at
the moment include:
+ Inlines with borders have broken and disjointed borders when there is
also padding present.
+ Vertical alignment is currently unsupported. Normally baselines should
be aligned. We also need subscript and superscript vertical alignment.
Images as well as text need to be vertically aligned.
+ There is some issue with incorrect line heights. I think it has
something to do with font geometry or various font sizes being used. It
can cause lines of text to overlap. John-Mark (jmb) and James (zamez)
know more about this.
+ Background colours for inlines do not fill the full area when there is
padding present.
+ Background image position is not correctly supported for inlines. Sites
such as Wikipedia place non-repeating images in the right padding of
link text. NetSurf currently places the background image at the left hand
side of the inline, under the text.
+ The white-space property is not supported. At the moment we can't do
no-wrap and stuff.
+ NetSurf's current line breaking goes wrong for situations like:
<span>1<sup>st</sup></span>. NetSurf will currently break the
line
between "1" and "st", but it shouldn't, because there is no
space in
the source.
+ Text that is inside a block, with no inline children around the text,
incorrectly inherit the background colour of the block. I made a test
case for this the other day:
http://www.smoothartist.com/netsurf/inline-background.html
I'll check that into NetSurf's SVN later.
Learning these problems one by one.. And what could be the possible reason
for such a behavior.
There are more issues which I haven't mentioned. However, fixing
the
background colour / image issues and the border problem will make a big
improvement. These often cause NetSurf's display to look untidy and
ragged.
Implementing the white-space and vertical-align properties are also very
clear-cut goals.
There are already test cases for some of the issues above in SVN.
> I later plan to run the NetSurf on the RPCemu emulator with RISC
> OS 3 review the work done.
I would not worry about this. The layout engine is part of the platform
independent core, so it will behave the same on RISC OS as on GTK NetSurf.
There may be a GTK NetSurf specific line height bug, but I wouldn't bother
about that for now.
Also, we plan to get a RISC OS box set up, that GSoC students will be able
to use remotely. This is nearly ready. It will show you what the RISC OS
version's features are (bookmarks, downloads, etc) which aren't present on
GTK NetSurf.
Michael
--
Great!! A RISC OS box :)
Michael Drake (tlsa) http://www.netsurf-browser.org/
--
Yours
Manjeet
5733
days inactive
5737
days old
netsurf-dev@netsurf-browser.org
4 comments
4 participants
participants (4)
-
John-Mark Bell -
Manjeet Chayel -
Michael Drake -
Rob Kendrick