-- It installs fine, then upon arriving at the welcome page, none of the links are able to fetch a page.
-- I tried the top right search bar, again same thing, any search criteria such as goolge, bbc, riscos, unable to fetch page.
-- If I was to type, www.google.co.uk in the address bar, the page does load fine.
-- If I search within google, its search results appears fine, but clicking on any of the links, unable to fetch page.
-- If I set frogfind as the default home page upon loading netsurf, (however it did lockup the entire computer requiring a restart) the page loads fine and I can search and click of any of its links, the website page itself loads fine albeit in a very simple text format without any fancy graphics.
-- I also tried the debug tool from: https://www.cy2.uk/ns_os4_dbg.zip and captured the ns.log for you for further review but I am not sure how to add the full ns log into the "" button or </> button at top of the reply button?
Hmm, OK, I need to go back to AmiSSL but with this new patch then!
Curl uses gethostbyname_r in the threaded and other resolver, so with this patch it should stop it using it regardless.
I'd like to get the threaded resolver working again though.
@boabster You can use something like pastebin, but at this stage unless you're seeing some other problem it's not going to tell me anything I don't already know.
Not sure if it helps in any way, but i stumbled upon this while looking for information regarding (un)supported features of OpenSSL/AmiSSL.
This is what i get while configuring ScummVM, which uses OpenSSL
Checks for ENet...
Checking for sendmsg... yes
Checking for recvmsg... yes
Optional checks for ENet...
Checking for getaddrinfo... yes
Checking for getnameinfo... yes
Checking for gethostbyaddr_r... no
Checking for gethostbyname_r... no
Checking for poll... no
Checking for fcntl... yes
Checking for inet_pton... yes
Checking for inet_ntop... yes
Checking for msghdr.msg_flags... yes
Checking for socklen_t... yes
gethostbyname_r is not supported (as it seems) amongst others.
So, this seems to be a general problem of our OpenSSL/AmiSSL port?
I think some people might be conflating 2 different issues. The gethostbyname_r issue only relates to libcurl resolving hostnames (Chris's latest Netsurf build is now linked against a libcurl that has that function disabled). Although libcurl uses either OpenSSL/AmiSSL for fetching SSL content, any SSL issue is unrelated to the hostname resolving issue. gethostbyname_r is not present in the OpenSSL/AmiSSL source.
The reason your configure script doesn't find gethostbyname_r is probably because you're using an old SDK. That function wasn't present in the included newlib library/headers until SDK 53.34, and not in the included Roadshow headers until SDK 54.16.
Edited by MickJT on 2024/10/4 11:59:43 Edited by MickJT on 2024/10/4 12:04:40 Edited by MickJT on 2024/10/4 12:13:06 Edited by MickJT on 2024/10/4 12:18:43 Edited by MickJT on 2024/10/4 12:19:50
This build works very well at first, but NetSurf also loses its connection later on. If you wait long enough NetSurf will terminate later where it has lost the connection, you can see it in the video quite far at the end. Otherwise the build is very stable....
When it eventually timed out did you get an error message? Can't see in the video as it processes the window close so any error doesn't stay on screen.
When it eventually timed out did you get an error message? Can't see in the video as it processes the window close so any error doesn't stay on screen.
There is no error message when NetSurf terminates after the time-out. If I then restart NetSurf, everything is fine again and I can visit all my websites until the connection is lost again.
I can reproduce the whole thing every time....
MacStudio ARM M1 Max Qemu//Pegasos2 AmigaOs4.1 FE / AmigaOne x5000/40 AmigaOs4.1 FE
for me, the Google search seems to eventually give a list of items, but NetSurf appears to go to sleep if I click on any of the entries.
I thought the program was hanging every time and the only way I could recover was to reset my system.
However, this last time I just waited to see if anything would happen. After about 5 to 10 minutes, the window updated with the article I was wanting to view and NetSurf was awake again.
Isn't the threaded resolver working ok now (with _r disabled)? Did Netsurf's libcurl have the threaded resolver enabled previously, years ago (with different patches) ?
I thought netsurf_os4_amissl_no_r.lha had the threaded resolver enabled but just not using _r. I didn't personally test it though.
@redfox
Whereas the OpenSSL one https://www.cy2.uk/netsurf_os4_no_r.lha
seemed to be much quicker.
Is that working without problems for you?
If that one has the threaded resolver enabled (but not using _r), then I would expect a version that uses AmiSSL should work fine too.
Edited by MickJT on 2024/10/5 7:58:54 Edited by MickJT on 2024/10/5 8:00:47
While both (10/1 & 10/4) builds still complain about the guigfx.lib at start up, both of them now are loading most of the pages linked on the NetSurf "home" page.
I say "most" because, as others reported, just clicking on the google link, froze up Netsurf with "loading" in the bottom corner.
I also found that while the ArsTechnica main page loaded, clicking on a link in that page also resulted in the same frozen "loading" state.
in both cases, I didn't see anything appear over the serial port reflecting a crash, etc.