Older blog entries for bagder (starting at number 174)

1 Aug 2003 (updated 1 Aug 2003 at 10:03 UTC) »

The well known tool GNU wget seems to have been left orphaned since the last year or so.

It is an interesting project to watch. How long can a well-known, well-used project be left to dry without anyone forking the source base and starting to improve it in a separate source tree?

People still mail in odd patches and bug reports to the mailing list, but there's no one with CVS access around.

The latest mail posted to the wget mailing list by the maintainer, Hrvoje Niksic, was posted back in December 2002.

Anyone out there interested in a 'HTTP authentication' library? Possibly only for NTLM, but I would imagine that it could do Digest and GSS-Negotiate too as they're not that simple to implement (at least it takes a lot of reading and testing).

A few weeks ago I added Digest authentication support to libcurl, Daniel Kouril added GSS-Negotiate authentication and then I challanged the programmer gods and added NTLM authentication too.

Now, there's a few other HTTP-oriented open source tools out there that I believe also would like to have these authentication methods supported. Why not gather a crowd and see if we can work up an API and then we could move the now written code to use that, and I could make the auth stuff available for others as well... (the libcurl source code is MIT licensed and thus not restricted to GPL, non-GPL, BSD or non-BSD projects)

It would also make the code better, as we all would strive on improving this.

If you think this is a good idea, well, mail me.

25 Mar 2003 (updated 25 Mar 2003 at 16:21 UTC) »

This is supposedly some kind of an open source/free software community. Yet 75% of the members claim to be involved in no projects at all, with another 10% involved in only one project...

I do get mildly annoyed by the fact that with the recent Redhat betas of 8.1 or 9 or whatever they'll end up calling it, curl fails to build with OpenSSL. Again, Redhat steps forward and causes havok.

This is because they've decided that we should all use 'pkg-config --cflags openssl' to figure out that we need the include path setup to point out the kereberos headers in order to be able to successfully include <openssl/ssl.h>...


Today my 'postmaster' got mail from "spamcop.net". He was utterly confused and mailed me and asked why he got this weird mail...

It turned out that spamcop.net is a blatantly stupid complain-to-the-spammers system, which tracks down the origins of URLs in "spams" and mails the postmasters of the supposedly offending spammer sites.

Now, the "spam" in this case was a mail in which someone described his travel by train in Peru, and there he added a URL to my page with a few pictures from the city of Aguas Calientes near Machu Picchu:


So, I'm an alleged spammer because I keep a page that one person liked to much he mailed a bunch of people pointing it out.

Well, I figure someone pulled some strings to make this happen and someone can of course have made a perfectly normal human error here, but it sure ended up very strange in my side of the world..

Now, let's get the cURL release complete and away. Stay tuned friends, it'll soon pop up in your freshmeat-syndicated box of recent announcements.

So McAfee apparently makes an anti-virus program called Virex, that is using libcURL, and it graciously mention me in the about window.

This has had the minor side-effect that users of Virex find my email address there and mail me for requests and questions regarding Virex.

I hadn't even heard of it when the first mails started to show up in my mailbox this weekend...! ;-)

This is actually one of the more pleasent kinds of problems, at least if it doesn't grow out of proportions.

jhermann, I can only agree with your sentiments on the sourceforge public CVS server status as of lately. It seems it is more dead than alive these days... :-/

Just wanted to mention that Denise is still alive, and that I've started working on it again after it having been idle for a couple of months.

It does work somewhat rudimentary already and I hope to get the example program to build and run fine with asynch one name => ipv4 resolve this week.

I will need help from people to get DNS server information on various platforms, to be able to make a decent gethostbyname() (etc) clone.

I like the fact they went with trio in Glib.

glibc 2.2.93 is the one used in Redhat 8.0

When we use that and call gethostbyname_r() with a too small buffer, it no longer returns ERANGE. It now returns EAGAIN!!!

But, older versions (like the ones in redhat 6.0 or 7.3) of glibc obviously sometimes returns EAGAIN when the lookup fails due to normal resolve failures.

So, what the heck shall I check for when I want to know if I need to enlarge the buffer or if the name just doesn't resolve?


165 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!