Commit Graph

217 Commits

Author SHA1 Message Date
Jake Burkholder
7b628f43e0 Reduce WARNS level to 3 in the non-crypto case to fix make release on sparc64.
This may only be turning up now to changes in the cpp predefines for sparc64,
which may be turning on more strict alignment checking.

Approved by:	re
2002-12-04 04:34:59 +00:00
Dag-Erling Smørgrav
d5216a4fb7 Implement and document support for an HTTP_REFERER environment variable.
PR:		28171
Submitted by:	Andre Albsmeier <andre.albsmeier@mchp.siemens.de>
Approved by:	re (bmah)
MFC after:	1 week
2002-11-28 12:07:15 +00:00
Bill Fenner
d5c24aa856 Fix two typos. 2002-11-08 19:10:05 +00:00
Dag-Erling Smørgrav
a49b853d94 Use __func__ and break a long line. 2002-10-30 15:01:29 +00:00
Dag-Erling Smørgrav
7504527ed2 Fix a bug in fenner's _fetch_writev() patch (rev 1.36)
Submitted by:	fenner
2002-10-30 14:25:00 +00:00
Dag-Erling Smørgrav
66b8df5fdf Recognize the (incorrect) error code a MediaHawk server sends in reply to
unrecognized commands such as MDTM.

Requested by:	Stephen Roome <stephen_roome@pepcross.com>
2002-10-30 06:06:16 +00:00
Dag-Erling Smørgrav
75c3973ab8 Clarify my feelings towards fetch / libfetch. 2002-10-30 04:53:58 +00:00
Dag-Erling Smørgrav
bb13d0af67 Recommit the non-broken parts of 1.34 and 1.37.
Change the type and name of a variable introduced in 1.33.
2002-10-30 04:43:00 +00:00
Dag-Erling Smørgrav
cab8495092 Don't forget to set the error flag when _fetch_read() returns -1. 2002-10-30 04:42:01 +00:00
Warner Losh
a4a37038bb Reinstate revs 1.35-36 and 1.38. Revisions 1.34 and 1.37 were specifically
the root cause of the bus errors I was experiencing.

Submitted by:	fenner
Tested by:	obrien
Prompted by:	peter
2002-10-30 00:17:16 +00:00
David E. O'Brien
b68fbebd5a Fix `pkg_add -r' by backing out revs 1.34-1.38.
Revs 1.37-8 produce a bus error in some environments.
Revs 1.34-6 do not bus error, but write corrupted files.
2002-10-29 12:17:43 +00:00
Dag-Erling Smørgrav
cf2d89d98b libfetch is now WARNS5 clean in the non-SSL case. 2002-10-28 10:37:31 +00:00
Dag-Erling Smørgrav
32a4a82829 Fix an off-by-one error (> where >= should have been used) which caused
_fetch_writev() to incorrectly report EPIPE in certain cases.

Also fix a number of const warnings by using __DECONST(), plus a signed /
unsigned comparison by casting the rhs to ssize_t.

Submitted by:	fenner, Craig Rodrigues <rodrigc@attbi.com>
2002-10-28 10:19:03 +00:00
Bill Fenner
8497092d00 Up WARNS to 3 if not building with crypto.
Approved by:	des
2002-10-28 01:41:28 +00:00
Dag-Erling Smørgrav
1a5424b137 Slight amendment to rev 1.34: instead of considering any short read an
error, only report an error if no data was read at all (unless len was
0 to start with).  Otherwise, the final read of practically any transfer
will end in a fatal error.
2002-10-27 17:20:49 +00:00
Dag-Erling Smørgrav
2761348f78 Introduce _fetch_writev(), which is the conn_t version of writev(2). In
the SSL case, it is no different from the old _fetch_write(), but in the
non-SSL case it uses writev(2) to send the entire vector as a single
packet (provided it can fit in one packet).  Implement _fetch_write()
and _fetch_putln() in terms of _fetch_writev().

This should improve performance in the non-SSL case (by reducing protocol
overhead) and solve the problem where too-smart-for-their-own-good
firewalls reject FTP packets that do not end in CRLF.

PR:		bin/44123
Submitted by:	fenner
2002-10-27 16:11:21 +00:00
Dag-Erling Smørgrav
9f788e9c90 Eliminate two cases of undefined behaviour: total in _fetch_write() was
not initialized before use, and _http_growbuf() did not return a value
on success.

Reported by:	Peter Edwards <pmedwards@eircom.net>
MFC after:	2 weeks
2002-10-27 15:43:40 +00:00
Dag-Erling Smørgrav
e24f60e74f Back out the previous commit, and fix the bug rather than try to hide its
symptoms: make timeouts and short transfers fatal, and set errno to an
appropriate value (ETIMEDOUT for a timeout, EPIPE for a short transfer).

MFC after:	2 weeks
2002-10-27 15:08:21 +00:00
Nate Lawson
13cc1c8394 The FTP connection caching needs a better interface -- connections are
closed through _fetch_close() which is the only one who knows the connection
REALLY was closed (since ref -> 0).  However, FTP keeps its own local
cached_connection and checks if it is valid by comparing it to NULL.  This
is bogus since it may have been freed elsewhere by _fetch_close().

This change checks if we are closing the cached_connection and the ref is 1
(soon to be 0).  If so, set cached_connection to NULL so we don't
accidentally reuse it.  The REAL fix should be to move connection caching
to the common.c level (_fetch_* functions) and NULL the cache(s) in
_fetch_close().  Then all layers could benefit from caching.
2002-10-25 01:17:32 +00:00
David E. O'Brien
513829ec5e Use a warns setting we can catch regressions with. 2002-10-20 23:50:28 +00:00
Tim J. Robbins
9d649c1fd0 When recycling a cached connection, increment the reference count so that
the heap block does not get freed and reused. This should fix the
pkg_add -r crashes that have been happening for months.
2002-10-20 10:36:27 +00:00
Ruslan Ermilov
721d559c03 HTTP authentication got broken in rev. 1.49. 2002-10-03 10:42:19 +00:00
Peter Wemm
224af215a6 Zap now-unused SHLIB_MINOR 2002-09-28 00:25:32 +00:00
Alfred Perlstein
a6756ecc22 Fix an infinite loop when _fetch_read() can return 0 (if the
connection is broken), take this into account and return at this
point.
2002-09-20 21:50:57 +00:00
Bill Fenner
40cfbfd508 Make _fetch_connect() always set the error code.
Tell ftp that _fetch_connect() always sets the error code (http already knew)
2002-09-17 05:54:33 +00:00
Dag-Erling Smørgrav
190c185e22 Surround prototypes with __BEGIN_DECLS / __END_DECLS.
PR:		misc/40399
MFC after:	3 days
2002-07-22 16:11:39 +00:00
Hajimu UMEMOTO
f8fa093e82 Cope with 2292bis-01 getaddrinfo (no NI_WITHSCOPEID, always attach
scope identifier).

Approved by:	des
MFC after:	3 weeks
2002-07-02 11:09:02 +00:00
Dag-Erling Smørgrav
66ffb8a371 Reintroduce debugging code that somehow got lost in a previous revision. 2002-06-24 12:18:41 +00:00
Dag-Erling Smørgrav
a7a4510fd8 Fix incorrect library ordering. I thought I'd committed this already... 2002-06-21 09:56:38 +00:00
Dag-Erling Smørgrav
f8bd33a0ae Don't try to dereference conn when we know it's NULL. 2002-06-19 08:36:00 +00:00
Dag-Erling Smørgrav
f606d589b9 Add a reference count to struct fetchconn so we don't prematurely close and
free a cached FTP connection.
2002-06-11 11:27:28 +00:00
Ruslan Ermilov
4093807dd9 libfetch now depends on libcrypto and libssl. 2002-06-06 13:45:46 +00:00
Dag-Erling Smørgrav
3070f6cb06 Make SSL support conditional on NOCRYPT. 2002-06-05 21:35:35 +00:00
Dag-Erling Smørgrav
d9615d7da4 During buildworld, "regular" libraries are built before crypto stuff, so
libfetch can't depend on lib{crypto,ssl}.  Move the dependency to fetch
until we can figure out how to fix this.
2002-06-05 21:25:33 +00:00
Dag-Erling Smørgrav
111e251009 Add SSL support + slight cleanup.
Submitted by:	Henry Whincup <henry@techiebod.com> (in principle)
2002-06-05 12:46:36 +00:00
Dag-Erling Smørgrav
9601e333a8 Wrap everything in struct connection, and enforce timeouts everywhere
(except for DNS operations).  Always use funopen() for HTTP, to support
both timeouts and SSL.
2002-06-05 12:19:08 +00:00
Dag-Erling Smørgrav
4dc0da3f3e Add the necessary dependencies for SSL. 2002-06-05 11:38:19 +00:00
Dag-Erling Smørgrav
ccdd94bdd9 Rename struct cookie to struct httpio to avoid confusion (it's not an HTTP
cookie) and increase symmetry with equivalent FTP code.
2002-06-05 10:31:01 +00:00
Dag-Erling Smørgrav
d3b03a9006 Add comments to struct cookie. 2002-06-05 10:27:24 +00:00
Dag-Erling Smørgrav
3f4823c55d Fix a bug I introduced in the chunk decoder in the previous commit.. 2002-06-05 10:23:19 +00:00
Dag-Erling Smørgrav
dea29ca1d5 First step towards SSL support: wrap connections in a 'struct connection'
which contains the socket descriptor, the input buffer and (yet unused)
SSL state variables.  This has the neat side effect of greatly improving
reentrance (though we're not *quite* there yet) and opening the door to
HTTP connection caching.

This commit is inspired by email conversations with and patches from
Henry Whincup <henry@techiebod.com> last fall.
2002-06-05 10:05:03 +00:00
Mark Murray
4cd0119367 Do not use __progname directly (except in [gs]etprogname(3)).
Also, make an internal _getprogname() that is used only inside
libc. For libc, getprogname(3) is a weak symbol in case a
function of the same name is defined in userland.
2002-03-29 22:43:43 +00:00
Poul-Henning Kamp
06229ad2af Modernize my email address 2002-03-25 13:53:46 +00:00
Dag-Erling Smørgrav
5a51c23be3 Switch to a self-starting allocation scheme. 2002-02-05 22:15:16 +00:00
Dag-Erling Smørgrav
e19e6098b3 Reindent, and add parentheses to return statements. Some functions in
ftp.c and http.c now have exceedingly long lines due to deep nesting;
this will be corrected by reorganizing the code in a later revision.
2002-02-05 22:13:51 +00:00
Dag-Erling Smørgrav
551858f0b0 Eliminate a redundant loop. 2002-02-05 21:32:16 +00:00
Dag-Erling Smørgrav
e6f0a33e68 Check the return value from read() when reading the CR/LF at the end of a
chunk.

PR:		bin/33608
MFC after:	2 weeks
2002-01-20 19:53:12 +00:00
Dag-Erling Smørgrav
e0583e0c23 Mark uploads as O_WRONLY, not O_RDONLY.
PR:		misc/34043
MFC after:	2 weeks
2002-01-20 19:52:25 +00:00
Dag-Erling Smørgrav
d6811c6851 Remove long-obsolete README. 2002-01-01 16:27:26 +00:00
Dag-Erling Smørgrav
a8e9bd8750 In verbose mode, display the full error message from the server, stripping
it of HTML tags and comments.

PR:		bin/32989
MFC after:	1 week
2002-01-01 16:25:29 +00:00