[Edbrowse-dev] 3.6.0

[Edbrowse-dev] 3.6.0

[Karl Dahlke]

As the windows port marches on to success, we may want to pause,
and finish everything currently in progress, and test, and debug,
and then call it 3.6.0.
This brings in some questions,
which I am in no way qualified to answer.

linux:
We want to keep this easy and clean for our distributers, debian, ubuntu, bsd, etc.
Will they git and package a version of tidy5, specifically 5.1.15,
as prerequisite to packaging edbrowse?
In other words, a package dependency?
Or will edbrowse build and distribute with tidy linked in as a static library,
so there are no additional package dependencies?

windows:
Same question about tidy, but also questions about mozjs.
Given the additional complexity of mozjs in windows, should we package
edbrowse and edbrowse-js separately, the latter depending on the former?
Or is it just as easy to bring them both in together.
How can casual windows users, who may not be experts in windows
installation processes, download and install edbrowse?

Chris will maintain his static binaries on edbrowse.org,
are there meaningful .exe files he can build and maintain for windows users?

Are there other bugs we need to fix, small enhancements we need to make?
In particular, (Kevin), are we miles away from jQuery,
wherein that should wait for 3.6.1, or would a few more small changes get us there?

Couple open issues with tidy but I'm working around them
so maybe they too should wait.

We should all browse around various sites on the net,
looking for seg faults, or mangled pages,
or other various bugs.
I imagine most of us do that in the course of our daily work and play.

Looking forward to 3.6.0.

Karl Dahlke

Re: [Edbrowse-dev] 3.6.0

[Kevin Carhart]

> Are there other bugs we need to fix, small enhancements we need to make?
> In particular, (Kevin), are we miles away from jQuery,
> wherein that should wait for 3.6.1, or would a few more small changes get us there?

The recent progress is significant, but I don't think it's going to be 
ready for 3.6.0.  Here's a crude gauge.  We're working through this 
javascript routine called jQuery.support, where they have bundled together 
a long series of tests of different mandatory behaviors.  And it tries 
them one after the next.  We've done several.  I'm currently working on a 
method called cloneNode, which roughly speaking is found at around line 75 
of 232 lines.  So I think our recent good work is about 33% of the way 
through the supports routine!  However, it could be even further than 
that, because when I scan lower, I notice a lot of use of the big node 
manipulations which are already done thanks to Karl. There's also empty 
space and comments, so it's not like the remaining 150 lines is dense with 
tricky DOM problems.

So it's coming, but it's a little more than a few small changes, I think.

By the way, recently I spent a few days learning the codebase better and 
tracing C calls from place to place.  I had hopped around in an episodic 
way, but hadn't gotten the big picture.  So now that I got to know the 
jseng-moz.cpp, readMessage, writeMessage, and the EJ_ commands, I like it!
Thank you for writing this, Karl, Adam & Chris.

Kevin

Re: [Edbrowse-dev] 3.6.0

[Adam Thompson]

[-- Attachment #1: Type: text/plain, Size: 1905 bytes --]

On Mon, Oct 26, 2015 at 11:30:21PM -0700, Kevin Carhart wrote:
> 
> >Are there other bugs we need to fix, small enhancements we need to make?
> >In particular, (Kevin), are we miles away from jQuery,
> >wherein that should wait for 3.6.1, or would a few more small changes get us there?
> 
> The recent progress is significant, but I don't think it's going to be ready
> for 3.6.0.  Here's a crude gauge.  We're working through this javascript
> routine called jQuery.support, where they have bundled together a long
> series of tests of different mandatory behaviors.  And it tries them one
> after the next.  We've done several.  I'm currently working on a method
> called cloneNode, which roughly speaking is found at around line 75 of 232
> lines.  So I think our recent good work is about 33% of the way through the
> supports routine!  However, it could be even further than that, because when
> I scan lower, I notice a lot of use of the big node manipulations which are
> already done thanks to Karl. There's also empty space and comments, so it's
> not like the remaining 150 lines is dense with tricky DOM problems.
> 
> So it's coming, but it's a little more than a few small changes, I think.

Excellent, thanks for doing this work.

> By the way, recently I spent a few days learning the codebase better and
> tracing C calls from place to place.  I had hopped around in an episodic
> way, but hadn't gotten the big picture.  So now that I got to know the
> jseng-moz.cpp, readMessage, writeMessage, and the EJ_ commands, I like it!
> Thank you for writing this, Karl, Adam & Chris.

Karl and Chris have to take the credit for this setup.

Given that we can't release until a tidy5 package,
I say we continue fixing things and then do the bug fix phase at that point.
I guess I'm really hoping for some sort of jquery support by that stage.

Cheers,
Adam.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

Re: [Edbrowse-dev] 3.6.0

[Kevin Carhart]

Thanks Adam.
We'll see what happens but the thing with this DOM support
is that it takes a long time, but just imagine!   When it's
up, and pages are finding what they're looking for a greater
percent of the time....  I think it's going to be reeeeeally nice.

Kevin

Re: [Edbrowse-dev] 3.6.0

[Chris Brannon]

Karl Dahlke <eklhad@comcast.net> writes:

> We want to keep this easy and clean for our distributers, debian, ubuntu, bsd, etc.
> Will they git and package a version of tidy5, specifically 5.1.15,

Most distros won't package directly from git.  If it's tagged as a
release, that's a different story.  So we should wait for
tidy-html5 to tag their next release before we tag one.
I can't really speak for Debian specifically, but I think we have a
Debian developer following this list.
>
> How can casual windows users, who may not be experts in windows
> installation processes, download and install edbrowse?

Hopefully through static binaries.  But I'll need someone to build
them.  I don't have access to Windows.  Well, I suppose I could get it
if I really needed it, but I don't know a thing about the development environment.

> We should all browse around various sites on the net,
> looking for seg faults, or mangled pages,

You know what's cool?  I haven't had a segfault with edbrowse in a long
time, and pages are looking pretty good as well.

-- Chris