]> git.ipfire.org Git - thirdparty/squid.git/log
thirdparty/squid.git
27 years agochange FSF address
wessels [Mon, 20 Jul 1998 23:18:42 +0000 (23:18 +0000)] 
change FSF address

27 years agoupdate
wessels [Mon, 20 Jul 1998 23:15:17 +0000 (23:15 +0000)] 
update

27 years agocopy request_t->flags when building a new request_t from the original
wessels [Mon, 20 Jul 1998 22:27:08 +0000 (22:27 +0000)] 
copy request_t->flags when building a new request_t from the original
one! (Arjan de Vet)

27 years agovery late update
wessels [Mon, 20 Jul 1998 10:57:19 +0000 (10:57 +0000)] 
very late update

27 years agoCheck for MORE than 4 digits when scanning input strings
wessels [Mon, 20 Jul 1998 09:51:29 +0000 (09:51 +0000)] 
Check for MORE than 4 digits when scanning input strings

27 years agodebugging
wessels [Sat, 18 Jul 1998 13:54:56 +0000 (13:54 +0000)] 
debugging

27 years agoupdate
wessels [Sat, 18 Jul 1998 13:54:17 +0000 (13:54 +0000)] 
update

27 years agotypo
wessels [Sat, 18 Jul 1998 13:43:04 +0000 (13:43 +0000)] 
typo

27 years agopretty major rewrite. Previously we wouldn't read from a socket
wessels [Sat, 18 Jul 1998 13:33:56 +0000 (13:33 +0000)] 
pretty major rewrite.  Previously we wouldn't read from a socket
again until the previously read chunk had been fully written to
the other side.  Now we read ahead on both sides, moving unwritten
chunks to the beginning of the copy buffers.

This also has consequences for the closing sequence.  When either side
closes, we are supposed to write any pending data to the other side
and then close the other side as well.  The code is now more symmetric,
but has to include all these ugly checks for valid filedescriptors.

27 years agotreat Cache-Control: revalidate just like proxy-revalidate
wessels [Sat, 18 Jul 1998 13:29:56 +0000 (13:29 +0000)] 
treat Cache-Control: revalidate just like proxy-revalidate

27 years agocachability for responses from authenticated requests should depend
wessels [Sat, 18 Jul 1998 13:28:03 +0000 (13:28 +0000)] 
cachability for responses from authenticated requests should depend
upon the presence of Cache-Control: Public, and not Cache-Control:
Proxy-Revalidate.

27 years agoalways explicitly set [Proxy-]Connection to be either Keep-Alive or Close
wessels [Sat, 18 Jul 1998 13:26:20 +0000 (13:26 +0000)] 
always explicitly set [Proxy-]Connection to be either Keep-Alive or Close

27 years ago- added a label param to debugObj to avoid mess in a log
rousskov [Fri, 17 Jul 1998 11:09:02 +0000 (11:09 +0000)] 
- added a label param to debugObj to avoid mess in a log

27 years ago- fixed type-casting warning for debugObj()
rousskov [Fri, 17 Jul 1998 10:57:20 +0000 (10:57 +0000)] 
- fixed type-casting warning for debugObj()

27 years ago- added debugObj(); same as debug() put prints "objects" that have an objPack()
rousskov [Fri, 17 Jul 1998 10:49:59 +0000 (10:49 +0000)] 
- added debugObj(); same as debug() put prints "objects" that have an objPack()
  method

27 years agoadded initial CARP code
wessels [Fri, 17 Jul 1998 07:02:22 +0000 (07:02 +0000)] 
added initial CARP code

27 years agoadd carp.o
wessels [Fri, 17 Jul 1998 06:48:21 +0000 (06:48 +0000)] 
add carp.o

27 years agoupdate
wessels [Fri, 17 Jul 1998 06:46:08 +0000 (06:46 +0000)] 
update

27 years agoadded --enable-carp
wessels [Fri, 17 Jul 1998 06:45:17 +0000 (06:45 +0000)] 
added --enable-carp

27 years agoAIX portability hacks
wessels [Fri, 17 Jul 1998 06:16:28 +0000 (06:16 +0000)] 
AIX portability hacks

27 years agoupdate
wessels [Fri, 17 Jul 1998 06:04:34 +0000 (06:04 +0000)] 
update

27 years agoFrom: Toomas Soome <tsoome@ut.ee>
wessels [Fri, 17 Jul 1998 06:02:46 +0000 (06:02 +0000)] 
From:    Toomas Soome <tsoome@ut.ee>
there is a bit better translation for Estonian ERR_READ_TIMEOUT error.
please update....

27 years ago* Do not warn about size mismatch on responses without a Content-Length:
wessels [Fri, 17 Jul 1998 06:00:51 +0000 (06:00 +0000)] 
* Do not warn about size mismatch on responses without a Content-Length:
header
* Documented all the options in usage().
* Added -t to build a tracefile suitable for later input

A tracefile has the format

METHOD URL BODY-FILE SIZE CHECKSUM
Where:
METHOD          Is the HTTP method to use, usually GET
URL             The URL
BODY-FILE       Is a file to send as request body (- for none)
SIZE            The object size of the returned object (- to ignore)
CHECKSUM        A simple checksum of the object (- to ignore)

tcp-banger already accepts this format as input (was in one of my
previous tcp-banger patches).

/Henrik

27 years agoSCO 3.2v4.2 doesn't have rint() -- mauri@mbp.ee
wessels [Fri, 17 Jul 1998 04:55:42 +0000 (04:55 +0000)] 
SCO 3.2v4.2 doesn't have rint() -- mauri@mbp.ee

27 years agomore const goop
wessels [Fri, 17 Jul 1998 04:47:24 +0000 (04:47 +0000)] 
more const goop

27 years agoconst goop
wessels [Fri, 17 Jul 1998 04:41:33 +0000 (04:41 +0000)] 
const goop

27 years agocall urlCanonical() in refreshCheck() if entry->mem_obj == NULL.
wessels [Fri, 17 Jul 1998 04:22:46 +0000 (04:22 +0000)] 
call urlCanonical() in refreshCheck() if entry->mem_obj == NULL.

Changed urlCanonical() interface and operation.  Writes the URL
into a xmalloc'd buffer (request_t->canonical).  Eliminated 'buf'
arg to urlCanonical().  client_side.c was the only place which used it,
so now we must xstrdup() there.

27 years agoneed to account for swap_hdr_sz offset when calculating on-disk bytes
wessels [Thu, 16 Jul 1998 11:53:06 +0000 (11:53 +0000)] 
need to account for swap_hdr_sz offset when calculating on-disk bytes

27 years agoupdate
wessels [Thu, 16 Jul 1998 10:27:46 +0000 (10:27 +0000)] 
update

27 years agoWe had a sort of a race condition problem. For object data in memory
wessels [Thu, 16 Jul 1998 10:24:49 +0000 (10:24 +0000)] 
We had a sort of a race condition problem.  For object data in memory
we were freeing up to the QUEUED offset instead of the DONE offset.
This resulted in a chunk of the object which was not in memory and
not yet written to disk.

27 years agocomments
wessels [Thu, 16 Jul 1998 09:46:47 +0000 (09:46 +0000)] 
comments

27 years agoadded no-netdb-exchange cache_peer option
wessels [Thu, 16 Jul 1998 08:39:59 +0000 (08:39 +0000)] 
added no-netdb-exchange cache_peer option

27 years ago- do not supply Age header if store entry does not have a valid timestamp
rousskov [Thu, 16 Jul 1998 06:29:53 +0000 (06:29 +0000)] 
- do not supply Age header if store entry does not have a valid timestamp

27 years ago- for multipart byterange responses we should not send Content-Length
rousskov [Thu, 16 Jul 1998 06:18:52 +0000 (06:18 +0000)] 
- for multipart byterange responses we should not send Content-Length
  until we can calculate it.

27 years agofix up some confusion/bugs with http->out.offset in clientSendMoreData().
wessels [Thu, 16 Jul 1998 05:59:07 +0000 (05:59 +0000)] 
fix up some confusion/bugs with http->out.offset in clientSendMoreData().
Now that wierd hacks for HEAD requests are gone, we have to change
where we increment the offset.   Now its incremented in two places,
for the headers and for the body.

27 years agoreplaced url.c::url_escape() with rfc1738_escape()
wessels [Thu, 16 Jul 1998 05:56:24 +0000 (05:56 +0000)] 
replaced url.c::url_escape() with rfc1738_escape()

27 years agoescape characters in ident
wessels [Thu, 16 Jul 1998 05:55:32 +0000 (05:55 +0000)] 
escape characters in ident

27 years agoupdate
wessels [Thu, 16 Jul 1998 04:30:38 +0000 (04:30 +0000)] 
update

27 years agoadd extern
wessels [Thu, 16 Jul 1998 02:35:48 +0000 (02:35 +0000)] 
add extern

27 years agowarnings
wessels [Wed, 15 Jul 1998 05:58:08 +0000 (05:58 +0000)] 
warnings

27 years agohad sample_start_time & sample_end_time reversed
wessels [Wed, 15 Jul 1998 05:20:33 +0000 (05:20 +0000)] 
had sample_start_time & sample_end_time reversed

27 years agotypos
wessels [Wed, 15 Jul 1998 05:16:16 +0000 (05:16 +0000)] 
typos

27 years agoadding
wessels [Wed, 15 Jul 1998 05:14:41 +0000 (05:14 +0000)] 
adding

27 years agoFrom: Henrik Nordstrom <hno@hem.passagen.se>
wessels [Wed, 15 Jul 1998 04:59:15 +0000 (04:59 +0000)] 
From:    Henrik Nordstrom <hno@hem.passagen.se>
Here is a patch that fixes FTP directory listings on some FTP servers /
directories.

* Always include .. first in directory listings, regardless if the
origin server lists one or not. This fixes both servers that don't list
.., and directories that have strange files/directories that get listed
before ..  (UNIX example: ".-test")

* Show a "empty" listing when the origin server returns a empty listing.
This is much more intiutive than returning a 0-bytes object (which gives
a "document contains no data" error box in most browsers). This is for
example seen on protected directories where the user is not allowed to
list the files (like bin and etc in most anon-ftp setups).

27 years agoFixed select-spin bug when an ICP reply actually gets queued
wessels [Wed, 15 Jul 1998 04:55:59 +0000 (04:55 +0000)] 
Fixed select-spin bug when an ICP reply actually gets queued
and we failed to execute the write callback

27 years agoremoved some OLD_CODE
wessels [Wed, 15 Jul 1998 04:36:46 +0000 (04:36 +0000)] 
removed some OLD_CODE

27 years agoupdate
wessels [Wed, 15 Jul 1998 04:30:02 +0000 (04:30 +0000)] 
update

27 years agoFixes to proxy-auth stuff, mostly to get the user name logged in
wessels [Wed, 15 Jul 1998 04:28:08 +0000 (04:28 +0000)] 
Fixes to proxy-auth stuff, mostly to get the user name logged in
access.log.

27 years agomissing storeComplete on need-proxy-auth replies
wessels [Wed, 15 Jul 1998 04:18:20 +0000 (04:18 +0000)] 
missing storeComplete on need-proxy-auth replies

27 years agoremove OLD_CODE
wessels [Wed, 15 Jul 1998 03:29:54 +0000 (03:29 +0000)] 
remove OLD_CODE

27 years agoStoreEntry->refresh was a bad idea. It would require changing
wessels [Wed, 15 Jul 1998 03:25:51 +0000 (03:25 +0000)] 
StoreEntry->refresh was a bad idea.  It would require changing
swap.state format and would change the format depending on if
USE_CACHE_DIGESTS were defined.

Now we use default refresh_pattern entries if the StoreEntry doesn't
have a MemObject (and therefore a URL).

27 years agoAdded time_t StoreEntry->refresh for cache digests.
wessels [Wed, 15 Jul 1998 02:54:53 +0000 (02:54 +0000)] 
Added time_t StoreEntry->refresh for cache digests.
Don't need to compile a lot of the store_digest code when USE_CACHE_DIGESTS
is not set

27 years agoChanged handling of cache hits. Swap in FIRST, then check IMS, refresh
wessels [Wed, 15 Jul 1998 01:56:39 +0000 (01:56 +0000)] 
Changed handling of cache hits.  Swap in FIRST, then check IMS, refresh
and other client-side processing.  This gives us mem_obj->url when
we call refreshCheck()

27 years agoupdate
wessels [Tue, 14 Jul 1998 23:28:21 +0000 (23:28 +0000)] 
update

27 years agounused vars
wessels [Tue, 14 Jul 1998 23:03:53 +0000 (23:03 +0000)] 
unused vars

27 years agopunctuation
wessels [Tue, 14 Jul 1998 22:17:55 +0000 (22:17 +0000)] 
punctuation

27 years agoMoved 'miss_access' ACL check to fwdStart()
wessels [Tue, 14 Jul 1998 12:12:58 +0000 (12:12 +0000)] 
Moved 'miss_access' ACL check to fwdStart()

27 years agoupdate
wessels [Tue, 14 Jul 1998 03:47:12 +0000 (03:47 +0000)] 
update

27 years agominor HTCP progress
wessels [Tue, 14 Jul 1998 03:37:44 +0000 (03:37 +0000)] 
minor HTCP progress

27 years agoallocate FwdState from mem pools
wessels [Tue, 14 Jul 1998 03:37:15 +0000 (03:37 +0000)] 
allocate FwdState from mem pools

27 years ago- replaced strncmp(token, "HTTP/", 5) with strncasecmp(...) -- Oskar Pearson
rousskov [Fri, 10 Jul 1998 01:21:00 +0000 (01:21 +0000)] 
- replaced strncmp(token, "HTTP/", 5) with strncasecmp(...) -- Oskar Pearson

27 years ago - Fixed snmp system Group; all objects are now returned.
glenn [Wed, 8 Jul 1998 23:36:17 +0000 (23:36 +0000)] 
    - Fixed snmp system Group; all objects are now returned.
        - Fixed snmp system Group sysDescr and sysContact.

27 years agochanged MemObject mem_hdr from a pointer to a struct
wessels [Tue, 7 Jul 1998 08:54:13 +0000 (08:54 +0000)] 
changed MemObject mem_hdr from a pointer to a struct

27 years ago- deleted some of the extensive http header debugging.
rousskov [Tue, 7 Jul 1998 01:49:35 +0000 (01:49 +0000)] 
- deleted some of the extensive http header debugging.

27 years agomaybe call fsync() before close
wessels [Tue, 7 Jul 1998 01:43:21 +0000 (01:43 +0000)] 
maybe call fsync() before close

27 years agomemInit before parseConfigFile
wessels [Tue, 7 Jul 1998 01:18:30 +0000 (01:18 +0000)] 
memInit before parseConfigFile

27 years ago- ignore range header in non-GET requests
rousskov [Sun, 5 Jul 1998 00:55:22 +0000 (00:55 +0000)] 
- ignore range header in non-GET requests

27 years agobody_buf/body_size sync fix
wessels [Sat, 4 Jul 1998 06:41:39 +0000 (06:41 +0000)] 
body_buf/body_size sync fix

27 years agobody_buf/body_size sync fix
wessels [Sat, 4 Jul 1998 06:34:07 +0000 (06:34 +0000)] 
body_buf/body_size sync fix

27 years agoCleaned up handling special-case partial replies like HEAD and Range
wessels [Sat, 4 Jul 1998 06:17:15 +0000 (06:17 +0000)] 
Cleaned up handling special-case partial replies like HEAD and Range
requests.  Added clientHttpRequest->flags.done_copying flag.

Fixed cache hits for HEAD requests on cached GET replies.

27 years agorequest_t->urlpath became a strbuf at some point
wessels [Thu, 2 Jul 1998 23:39:41 +0000 (23:39 +0000)] 
request_t->urlpath became a strbuf at some point

27 years agounescape special characters in names and passwords
wessels [Thu, 2 Jul 1998 22:53:39 +0000 (22:53 +0000)] 
unescape special characters in names and passwords

27 years agodebug, gindent
wessels [Wed, 1 Jul 1998 05:10:16 +0000 (05:10 +0000)] 
debug, gindent

27 years agobuffered reply fixes
wessels [Wed, 1 Jul 1998 05:03:17 +0000 (05:03 +0000)] 
buffered reply fixes

27 years agoline-at-a-time processing fixes. Bugs caused by ftp://ftp.proftpd.org/
wessels [Wed, 1 Jul 1998 01:32:38 +0000 (01:32 +0000)] 
line-at-a-time processing fixes.  Bugs caused by ftp://ftp.proftpd.org/

ftpParseListing() didn't properly retain ->data.offset between
calls.

proftpd was sending two control replies together, e.g.

150 opening connection
226 transfer done

ftpParseControlReply() was returning '226' only.  Now the control stream
is "buffered" and ftpParseControlReply() has been split into a reading
part and a processing part.

27 years ago- cleanup
rousskov [Tue, 30 Jun 1998 21:08:10 +0000 (21:08 +0000)] 
- cleanup

27 years agoentry might be NULL in peer_select
wessels [Tue, 30 Jun 1998 13:11:11 +0000 (13:11 +0000)] 
entry might be NULL in peer_select

27 years agolock store entry during peer selection phase, esp important for ICP
wessels [Tue, 30 Jun 1998 05:29:16 +0000 (05:29 +0000)] 
lock store entry during peer selection phase, esp important for ICP
because of delays.
check the cbdata status of the peer select callback in peerSelectTimeout()

27 years agoa little rearrangement. For CACHEOBJ and INTERNAL requests we don't
wessels [Tue, 30 Jun 1998 03:03:01 +0000 (03:03 +0000)] 
a little rearrangement.  For CACHEOBJ and INTERNAL requests we don't
need (want) to create fwdState because it won't be attached to
a server-side FD and won't be deallocated.

27 years agohttpSocketOpen unused code
wessels [Tue, 30 Jun 1998 01:29:01 +0000 (01:29 +0000)] 
httpSocketOpen unused code

27 years agodebug level
wessels [Mon, 29 Jun 1998 21:26:49 +0000 (21:26 +0000)] 
debug level

27 years agoMove pumpRestart check from http.c to forward.c
wessels [Mon, 29 Jun 1998 21:22:50 +0000 (21:22 +0000)] 
Move pumpRestart check from http.c to forward.c
Removed unused httpRestart() and httpTryRestart()
removed storeUnregister/storeClientListAdd calls for pump requests because
we need the pump code to stay attached to the client in case we need to
re-forward the request.

27 years ago4k buf leak
wessels [Sun, 28 Jun 1998 22:53:09 +0000 (22:53 +0000)] 
4k buf leak

27 years agowhoisStart
wessels [Sun, 28 Jun 1998 22:18:44 +0000 (22:18 +0000)] 
whoisStart

27 years agoDon't re-forward STORE_COMPLETE entries!
wessels [Sun, 28 Jun 1998 22:18:21 +0000 (22:18 +0000)] 
Don't re-forward STORE_COMPLETE entries!

27 years agochanged whoisReadReply() to call fwdFail() for re-forwarding
wessels [Sun, 28 Jun 1998 22:17:31 +0000 (22:17 +0000)] 
changed whoisReadReply() to call fwdFail() for re-forwarding
requests

27 years agoassert storeComplete only transitions from PENDING to COMPLETE states
wessels [Sun, 28 Jun 1998 13:52:59 +0000 (13:52 +0000)] 
assert storeComplete only transitions from PENDING to COMPLETE states

27 years agodebug, gindent
wessels [Sun, 28 Jun 1998 13:52:13 +0000 (13:52 +0000)] 
debug, gindent

27 years agohandle STORE_ABORTED case
wessels [Sun, 28 Jun 1998 09:24:34 +0000 (09:24 +0000)] 
handle STORE_ABORTED case

27 years agocall storeComplete BEFORE fwdUnregister
wessels [Sun, 28 Jun 1998 08:01:44 +0000 (08:01 +0000)] 
call storeComplete BEFORE fwdUnregister

27 years agore-forwarding fixes
wessels [Sat, 27 Jun 1998 03:54:48 +0000 (03:54 +0000)] 
re-forwarding fixes

27 years agoretry forwarding requests
wessels [Sat, 27 Jun 1998 03:24:55 +0000 (03:24 +0000)] 
retry forwarding requests

27 years agoadjust sensitivity of peer RTT calculations. Changed the average factor
wessels [Fri, 26 Jun 1998 10:23:33 +0000 (10:23 +0000)] 
adjust sensitivity of peer RTT calculations.  Changed the average factor
from 1000 to 50.  Added a call to gettimeofday() in comm_incoming().
Surely someone will protest that.

27 years agodebug
wessels [Fri, 26 Jun 1998 04:54:50 +0000 (04:54 +0000)] 
debug

27 years agowhitespacing
wessels [Fri, 26 Jun 1998 02:14:36 +0000 (02:14 +0000)] 
whitespacing

27 years agoupdate
wessels [Fri, 26 Jun 1998 02:09:36 +0000 (02:09 +0000)] 
update

27 years agoWe can now cache Set-Cookie replies. This header is filtered out
wessels [Fri, 26 Jun 1998 02:08:54 +0000 (02:08 +0000)] 
We can now cache Set-Cookie replies.  This header is filtered out
for cache hits on the client side.

27 years agostrip arguments AFTER a question-mark, but leave the question-mark in
wessels [Thu, 25 Jun 1998 05:04:59 +0000 (05:04 +0000)] 
strip arguments AFTER a question-mark, but leave the question-mark in
place

27 years agostop reading NON-GET requests until they get resolved or forwarded
wessels [Thu, 11 Jun 1998 11:32:19 +0000 (11:32 +0000)] 
stop reading NON-GET requests until they get resolved or forwarded

27 years agocompiler warnings
wessels [Thu, 11 Jun 1998 00:46:41 +0000 (00:46 +0000)] 
compiler warnings