Absolute URL from relative path

Question

I've got a base URL and some relative URIs in that document I want to have the absolute path of.

E.g. base = https://example.com/some/path.html?query=string and relative URIs in that document:

• index.html → https://example.com/some/index.html
• .. → https://example.com/
• ../../../../abc → https://example.com/abc
• abc/.. → https://example.com/some/
• //example.org/ → https://example.org/
• ftp://example.net/ → ftp://example.net/

In Java you have the class URL accomplishing that:

URL abs = new URL(new URL(basePath), relPath);

But strangely enough I can't find a plain C library or function realizing just that.

Is there a library proving that feature? Or better yet some small self-contained file one can use?

Answer 1

You can bet this has been written a thousand times in C already. For apache, for example.

Here are some pointers:

libSoup, the http library used by GNOME: http://developer.gnome.org/libsoup/unstable/SoupURI.html#soup-uri-new-with-base

Proposed for Boost libraries: http://cpp-netlib.github.com/

By Google themselves (part of Chrome?): http://code.google.com/p/google-url/

Yet another: http://uriparser.sourceforge.net/

W3C: http://www.w3.org/Library/src/HTParse

URL parsing in libcamel: http://www.google.com/codesearch#KhbZeNk3OGk/camel/camel-url.c

Some more URI parsing APIs, that all seem to not have relative URIs:

GLib, my favorite C library: http://developer.gnome.org/glib/unstable/glib-URI-Functions.html

libedataserver (from Evolution) http://developer.gnome.org/libedataserver/stable/libedataserver-e-url.html

GNet, a glib addon: http://developer.gnome.org/gnet/stable/gnet-uri.html

Answer 2

For Windows, you can use CoInternetCombineUrl from Urlmon.dll or UrlCombine from Shlwapi.lib. They do the same AFAIK.