This repository has been archived by the owner. It is now read-only.
Fetching contributors…
Cannot retrieve contributors at this time
47 lines (36 sloc) 2.21 KB
Due to the excessive amount of problem reports and long known (and unfixed)
problems and due to the non-trivial nature of internet protocols, this fork
_strongly recommends to avoid_ using this library in anything production or
in fact anything more serious than simple test code for educational purposes.
Please notice that even if something happens to work in some specific
scenario, it's highly likely it's not a stable solution, nor is it a secure
The only reason hbtip hasn't been deleted altogether is compatibility with
mainline (3.2) Harbour, xHarbour and existing projects, plus the fact there
are still some low-level functions that have have no better replacement yet
For internet protocol related tasks, the recommended and supported library
is _hbcurl_, which is a thin wrapper over libcurl's 'easy' API. libcurl is
a highly ubiquitious, very stable, actively and professionally developed,
secure communications library:
hbcurl's functionality covers _all_ hbtip protocol classes, and it does offer
support for several protocols besides these.
Usage examples found on the internet for other languages are most of the time
directly usable with minor modifications in Harbour. You can find several
working examples under the hbcurl tests directory as well. Live curl
command-lines can also be converted into compilable/adaptable API code using
the `--libcurl <targe-source-file>` curl option.
This fork's snapshot binaries for Windows come bundled with all required
static and dynamic libraries to use hbcurl, including OpenSSL and libssh2,
both 32 and 64-bit. See this page about feature support and other details:
If you opt to use hbtip anyway and find any problem with it (like very common
SMTP or FTP incompatibilities with various live servers found on the internet),
your best bet is to report these on mainline in the hope someone will address
them, or even better to address the problem yourself. Speaking of this fork,
I'm only interested in reports that address regressions compared to mainline,
plus I merge/port almost all mainline updates back into this fork.