Board.KolibriOS.org

Mario_r4 wrote:2) Не смог запустить Netsurf, даже выставив Qemu памяти 512 МБ. При этом его даже через CPU прибить нельзя.

Need to check why Netsurf hung filesystem.

1. download http://kolibri-n.org/files/KolibriN9Beta.7z
2. start system
3. run /kolibrios/network/netsurf/netsurf or /cd0/1/kolibrios/network/netsurf/netsurf (it doesn't matter)
=> in both choices there will by filesystem hand http://prntscr.com/3ccbac

Leency wrote:

Mario_r4 wrote:2) Не смог запустить Netsurf, даже выставив Qemu памяти 512 МБ. При этом его даже через CPU прибить нельзя.

Need to check why Netsurf hung filesystem.

1. download http://kolibri-n.org/files/KolibriN9Beta.7z
2. start system
3. run /kolibrios/network/netsurf/netsurf or /cd0/1/kolibrios/network/netsurf/netsurf (it doesn't matter)
=> in both choices there will by filesystem hand http://prntscr.com/3ccbac

because NetSurf wants writable filesystem.
Copy NetSurf from cd0 to tmp0 helps.

Writable filesystem for what? Anyway its not OK that system hangs when Netsurf is loading. Maybe there occurs writing loop when writing fails?

Leency wrote:Writable filesystem for what?

For the log files.

Нашел порт в Syllable https://sites.google.com/site/syllables ... rk/netsurf

NetSurf 3.1 released 26 Apr 2014

http://www.netsurf-browser.org/

Hello!
Sorry for the delay.

I'm currently working on getting a multi-transfer (basically, parallelly fetching multiple files for multiple GET requests) working with Netsurf. Earlier, I got to the point of fixed redirections which enabled the Google search results to load, but there was a bug with the code. Also, the older code seemed to be centered around single one-at-a-time GET transfers, so with advice from hidnplayr and soUrcerer, I've begun work on this from scratch. I should hopefully be able to get it done soon and use http.obj after removing parts of curl that are in the code.

Stay tuned in

Regards,
Ashish

Getting the compiler and linker errors out of the way, looks like most parts of the file will need to be rewritten (such as all the curl_* functions that are used) in order to keep the netsurf structure intact.

The file curl.c seems to also handle parts of cache fetching so this should be useful when we are dealing with cookies and cache later.

I currently plan to maintain the handles in a linked list (possibly doubly) and then use this ring to deal with the ones which have finished and the ones which need more time to finish.

Regards,
Ashish

Hello.

I've read through most of the curl man pages both available on my machine as well as on their website and replaced much of the functionality offered by curl for HTTP related tasks. I've added replacements for most libcurl functions. The compiler errors/warnings have been resolved, just fixing the linker errors regarding some stubs that soUrcerer wrote when he was working on the port.

All the parts for the multi interface currently use http.obj for everything except POST data, which I am working on. The next update should have a binary for testing this interface out so I can fix any bugs that come up

Regards.

Hi

The curl interface is almost replaced. There's a binary also available on my github repo regarding netsurf kos if you are interested. I didn't post it yet because some work remains to finalize the multi interface. I'll post a binary soon . Netsurf stopped crashing for now, seems like it's stable.

Regards,
Ashish

ashmew2 wrote:There's a binary also available on my github repo regarding netsurf kos if you are interested.

I hope you upload all files to SVN in the future.

Mario_r4 wrote:I hope you upload all files to SVN in the future.

Yeah definitely. That's the plan. The only reason to do this on a separate git repo rather than the main SVN repo was to not litter the SVN with too many small changes

Regards,
Ashish

Hi!

Netsurf has seen some progress in the last week. The sites load, (google.com loads!) , but the content does not seem to load. This is because enough information from the fetcher interface is not being passed to the CACHE (HLCACHE and LLCACHE) in Netsurf. Once this is done, all the content on a website such as .png files, .ico files and other images/resources should be able to load using Netsurf's content handlers. Seems like it's almost fixed, but there are a minor hiccups which are preventing it from working properly.

I'm currently fixing the content handlers.

Regards,
Ashish

Hi!

I have fixed the content handlers for handling images etc and it seems to be loading images. Also added FETCH_HEADER callbacks for all header lines so that the browser knows whats going on. Some work remains in order to get the fetcher completely working. I plan to start soon with taking care of the 100% CPU usage which seems to be a framebuffer related issue.

Regards,
Ashish

Hi!

Looks like netsurf can load most web pages now, seems pretty stable to me although you can definitely witness an occasional crash. Please report if you face any crashes etc.
Also attaching the binary so you can run it and test

PS - The 100% CPU Usage still exists, it should hoppefully be better once we use the native UI for KolibriOS instead of the Netsurf framebuffer. I'm working on it.

Regards,
Ashish