]> git.ipfire.org Git - thirdparty/squid.git/log
thirdparty/squid.git
28 years ago1.2.alpha4
wessels [Fri, 27 Jun 1997 05:07:14 +0000 (05:07 +0000)] 
1.2.alpha4

28 years agouse calloc, not malloc
wessels [Fri, 27 Jun 1997 05:02:15 +0000 (05:02 +0000)] 
use calloc, not malloc

28 years agoHAVE_SNPRINTF check
wessels [Fri, 27 Jun 1997 04:43:51 +0000 (04:43 +0000)] 
HAVE_SNPRINTF check

28 years agoupdate
wessels [Fri, 27 Jun 1997 04:43:40 +0000 (04:43 +0000)] 
update

28 years agoMax Okumoto configuration file patch. Config goop will be generated
wessels [Fri, 27 Jun 1997 04:41:35 +0000 (04:41 +0000)] 
Max Okumoto configuration file patch.  Config goop will be generated
from a single file.  Still needs work.

28 years agogindent
wessels [Fri, 27 Jun 1997 04:35:35 +0000 (04:35 +0000)] 
gindent

28 years agoadding
wessels [Fri, 27 Jun 1997 04:29:30 +0000 (04:29 +0000)] 
adding

28 years agoupdate
wessels [Wed, 25 Jun 1997 02:24:24 +0000 (02:24 +0000)] 
update

28 years agocompiler
wessels [Wed, 25 Jun 1997 02:24:20 +0000 (02:24 +0000)] 
compiler

28 years agoObject might be validated between being added to validation list and
wessels [Sat, 21 Jun 1997 10:55:58 +0000 (10:55 +0000)] 
Object might be validated between being added to validation list and
when storeValidate is called.

28 years agoclean up access logging, remove lots of Cacheinfo garbage
wessels [Sat, 21 Jun 1997 08:38:03 +0000 (08:38 +0000)] 
clean up access logging, remove lots of Cacheinfo garbage
Need to relocate enums to an enums.h file, too many other wierd
dependencies.

28 years agoadding
wessels [Sat, 21 Jun 1997 08:36:47 +0000 (08:36 +0000)] 
adding

28 years agofix assertions
wessels [Fri, 20 Jun 1997 11:26:08 +0000 (11:26 +0000)] 
fix assertions

28 years agoremove some of that HTTPCacheInfo->crap
wessels [Fri, 20 Jun 1997 11:08:30 +0000 (11:08 +0000)] 
remove some of that HTTPCacheInfo->crap

28 years agoin ftpAbort, close the descriptors instead of trying to send QUIT
wessels [Fri, 20 Jun 1997 08:22:11 +0000 (08:22 +0000)] 
in ftpAbort, close the descriptors instead of trying to send QUIT

28 years agossize_t support
wessels [Fri, 20 Jun 1997 06:03:24 +0000 (06:03 +0000)] 
ssize_t support
snprintf source

28 years ago- memCopy needs to return a signed value to report errors, so we need
wessels [Fri, 20 Jun 1997 06:00:10 +0000 (06:00 +0000)] 
- memCopy needs to return a signed value to report errors, so we need
  to have 'ssize_t'
- Fixed FTP end-of-transfer confusion
- Rearranged assertions in icpSendMoreData to account for STORE_ABORT

28 years agoMucking about with storeAbort. Removed squid_error_entry() (need
wessels [Fri, 20 Jun 1997 04:51:48 +0000 (04:51 +0000)] 
Mucking about with storeAbort.  Removed squid_error_entry() (need
to replace it later!) and just call storeAbort directly now.

28 years agoPut storeDeleteBehind() call before storeCopy callback. Apparently
wessels [Wed, 18 Jun 1997 22:15:53 +0000 (22:15 +0000)] 
Put storeDeleteBehind() call before storeCopy callback.  Apparently
invoking the callback may release the StoreEntry, so either we need
to lock it or do the delete behind first.

28 years agocbdataLock/cbdataUnlock around comm_write
wessels [Wed, 18 Jun 1997 22:00:09 +0000 (22:00 +0000)] 
cbdataLock/cbdataUnlock around comm_write

28 years agoassociate storeRegisterAbort calls with comm_add_close_handler calls
wessels [Wed, 18 Jun 1997 10:04:13 +0000 (10:04 +0000)] 
associate storeRegisterAbort calls with comm_add_close_handler calls

28 years agogindent, debug
wessels [Wed, 18 Jun 1997 10:03:42 +0000 (10:03 +0000)] 
gindent, debug

28 years agodebugging and assertions
wessels [Wed, 18 Jun 1997 09:06:48 +0000 (09:06 +0000)] 
debugging and assertions

28 years agoremove assertion from store rebuild code. A swap log entry might be
wessels [Wed, 18 Jun 1997 09:06:22 +0000 (09:06 +0000)] 
remove assertion from store rebuild code.  A swap log entry might be
"newer" if we're in a slow rebuild.  In that case the swap log entry
should be ignored, assuming the existing entry was added during the
rebuild.

28 years agomore cbdata additions
wessels [Wed, 18 Jun 1997 07:43:42 +0000 (07:43 +0000)] 
more cbdata additions

28 years agoadding
wessels [Wed, 18 Jun 1997 06:20:07 +0000 (06:20 +0000)] 
adding

28 years agoJunk the 'callback_meta' idea
wessels [Wed, 18 Jun 1997 06:19:50 +0000 (06:19 +0000)] 
Junk the 'callback_meta' idea
Now use a hash table of callback datas.  A callback pointer is added when
the data is created, then callback functions lock and unlock it while
waiting for the operation to complete.

28 years agofurther HashID changes
wessels [Wed, 18 Jun 1997 01:00:02 +0000 (01:00 +0000)] 
further HashID changes

28 years agoforgot to lock object before validating
wessels [Wed, 18 Jun 1997 00:58:19 +0000 (00:58 +0000)] 
forgot to lock object before validating

28 years agoThe 'int status' arg of validate callbacks seems unused, nuke it
wessels [Wed, 18 Jun 1997 00:04:46 +0000 (00:04 +0000)] 
The 'int status' arg of validate callbacks seems unused, nuke it
In storeCleanup(), skip entries which are already validated, or which
   are pending for release
Lock/Unlock objects during validation
set RELEASE_REQUEST flag if delaying release because of store rebuild

28 years agohash.c cleanup
wessels [Tue, 17 Jun 1997 10:54:08 +0000 (10:54 +0000)] 
hash.c cleanup

28 years agoreload dup bugfix
wessels [Tue, 17 Jun 1997 09:44:35 +0000 (09:44 +0000)] 
reload dup bugfix

28 years agogeneric callback support
wessels [Tue, 17 Jun 1997 09:03:19 +0000 (09:03 +0000)] 
generic callback support

28 years agoadding
wessels [Tue, 17 Jun 1997 04:02:04 +0000 (04:02 +0000)] 
adding

28 years agoremove ->fd member from struct _ip_pending
wessels [Tue, 17 Jun 1997 04:01:42 +0000 (04:01 +0000)] 
remove ->fd member from struct _ip_pending
move some multicast code to multicast.[ch]

28 years agoNeed to set FTP_USE_BASE bit when we know its a directory so the BASE
wessels [Fri, 6 Jun 1997 12:16:19 +0000 (12:16 +0000)] 
Need to set FTP_USE_BASE bit when we know its a directory so the BASE
element gets added.

28 years agogindent
wessels [Wed, 4 Jun 1997 13:12:23 +0000 (13:12 +0000)] 
gindent

28 years agoupdate
wessels [Wed, 4 Jun 1997 13:11:09 +0000 (13:11 +0000)] 
update

28 years agocompiler
wessels [Wed, 4 Jun 1997 13:02:55 +0000 (13:02 +0000)] 
compiler

28 years agouse alarm(3) for timing
wessels [Wed, 4 Jun 1997 13:00:31 +0000 (13:00 +0000)] 
use alarm(3) for timing

28 years agoupdate
wessels [Wed, 4 Jun 1997 12:20:09 +0000 (12:20 +0000)] 
update

28 years agonew debug() macro; David Luyer
wessels [Wed, 4 Jun 1997 12:15:44 +0000 (12:15 +0000)] 
new debug() macro; David Luyer

28 years agoUSE_ASYNC_IO cleanup
wessels [Wed, 4 Jun 1997 11:50:26 +0000 (11:50 +0000)] 
USE_ASYNC_IO cleanup

28 years agofix code protected by USE_ICMP
wessels [Wed, 4 Jun 1997 10:32:58 +0000 (10:32 +0000)] 
fix code protected by USE_ICMP

28 years agoupdate
wessels [Wed, 4 Jun 1997 10:15:08 +0000 (10:15 +0000)] 
update

28 years agogot a coredump from storeAbort because ping_status == PING_WAITING.
wessels [Wed, 4 Jun 1997 02:09:45 +0000 (02:09 +0000)] 
got a coredump from storeAbort because ping_status == PING_WAITING.
I'm not sure why its bad to abort an object in that state, so that
check has been commented out.  Also I don't think its necessary to
set ENTRY_DISPATCHED anymore; there should be other checks in place
for aborted entries.   Also changed a bunch of debug_trap()'s to
assertions.

28 years agoThere were a couple of situations where we create new request_t structures
wessels [Wed, 4 Jun 1997 02:08:22 +0000 (02:08 +0000)] 
There were a couple of situations where we create new request_t structures
that didn't have the headers buf from the original request.  Fix was to
give httpBuildRequestHeader() one fewer argument because the headers
will be attached to the 'orig_request' arg anyway.

28 years agoupdate
wessels [Wed, 4 Jun 1997 00:03:40 +0000 (00:03 +0000)] 
update

28 years agodebug
wessels [Tue, 3 Jun 1997 01:56:44 +0000 (01:56 +0000)] 
debug

28 years agopurifying
wessels [Tue, 3 Jun 1997 01:55:58 +0000 (01:55 +0000)] 
purifying

28 years agofix failed assertion
wessels [Mon, 2 Jun 1997 23:19:25 +0000 (23:19 +0000)] 
fix failed assertion

28 years agochange CLIENT_ABORT_REQUEST into mem_obj->abort.callback()
wessels [Mon, 2 Jun 1997 11:39:43 +0000 (11:39 +0000)] 
change CLIENT_ABORT_REQUEST into mem_obj->abort.callback()

28 years ago- finish moving mem_obj->req_hdr to request->headers.
wessels [Mon, 2 Jun 1997 07:06:09 +0000 (07:06 +0000)] 
- finish moving mem_obj->req_hdr to request->headers.
- Misc debug cleanup
- Changed store_rebuilding/store_validating duality.  Now store_rebuilding
  is set until validation finishes.

28 years agonew storeCreateEntry has fewer args
wessels [Mon, 2 Jun 1997 05:23:46 +0000 (05:23 +0000)] 
new storeCreateEntry has fewer args

28 years ago- Moved IP lookups to commConnect stuff.
wessels [Mon, 2 Jun 1997 05:22:17 +0000 (05:22 +0000)] 
- Moved IP lookups to commConnect stuff.
- Added support for retrying connect().

28 years agogindent
wessels [Mon, 2 Jun 1997 05:21:37 +0000 (05:21 +0000)] 
gindent

28 years agostart removing mem_obj->req_hdr and moving it to request_t->headers
wessels [Mon, 2 Jun 1997 00:19:51 +0000 (00:19 +0000)] 
start removing mem_obj->req_hdr and moving it to request_t->headers

28 years agocleanup from lack of ftpget
wessels [Sun, 1 Jun 1997 10:23:08 +0000 (10:23 +0000)] 
cleanup from lack of ftpget

28 years agoremove ftpProcessReplyHeader
wessels [Tue, 27 May 1997 09:13:19 +0000 (09:13 +0000)] 
remove ftpProcessReplyHeader

28 years agoadding
wessels [Tue, 27 May 1997 08:54:27 +0000 (08:54 +0000)] 
adding

28 years agopurify fixes
wessels [Tue, 27 May 1997 08:48:50 +0000 (08:48 +0000)] 
purify fixes

28 years agocompiler
wessels [Mon, 26 May 1997 10:20:40 +0000 (10:20 +0000)] 
compiler

28 years agomostly done inlining all FTP requsts
wessels [Mon, 26 May 1997 10:04:52 +0000 (10:04 +0000)] 
mostly done inlining all FTP requsts

28 years agonever use mallocmap()
wessels [Sat, 24 May 1997 02:46:13 +0000 (02:46 +0000)] 
never use mallocmap()

28 years agoadd recursion detection to storeClientCopy()
wessels [Sat, 24 May 1997 02:45:58 +0000 (02:45 +0000)] 
add recursion detection to storeClientCopy()

28 years ago- the global 'store_rebuilding' is now a boolean. The clean/dirty indication
wessels [Sat, 24 May 1997 01:48:06 +0000 (01:48 +0000)] 
- the global 'store_rebuilding' is now a boolean.  The clean/dirty indication
  is associated with the individual swap directories.
- This fixes entries from a clean rebuild getting marked as not VALIDATED.
- storeSwapInCheck() was broken.  Moved it back into storeSwapInStart().

28 years agofix parseTimeLine() calls. The units arg needs to be the units in which
wessels [Sat, 24 May 1997 01:45:25 +0000 (01:45 +0000)] 
fix parseTimeLine() calls.  The units arg needs to be the units in which
the values are stored, not necessarily the defaults in the config file.

28 years agoentry/old_entry switching bug
wessels [Fri, 23 May 1997 23:21:17 +0000 (23:21 +0000)] 
entry/old_entry switching bug

28 years agoadd assertion
wessels [Fri, 23 May 1997 22:56:30 +0000 (22:56 +0000)] 
add assertion

28 years agoreinstall read hander on partial dnsserver replies
wessels [Fri, 23 May 1997 22:56:16 +0000 (22:56 +0000)] 
reinstall read hander on partial dnsserver replies

28 years agocompiler
wessels [Fri, 23 May 1997 11:33:18 +0000 (11:33 +0000)] 
compiler

28 years agoNeed to allocate mem_obj in storeSwapInStart() and not
wessels [Fri, 23 May 1997 11:30:36 +0000 (11:30 +0000)] 
Need to allocate mem_obj in storeSwapInStart() and not
storeSwapInStartComplete().

28 years agoIn storeComplete(), call InvokeHandlers() AFTER setting all the statuses.
wessels [Fri, 23 May 1997 11:20:57 +0000 (11:20 +0000)] 
In storeComplete(), call InvokeHandlers() AFTER setting all the statuses.

28 years agomemAppend can be void
wessels [Fri, 23 May 1997 11:15:57 +0000 (11:15 +0000)] 
memAppend can be void

28 years agofix incorrect assertion
wessels [Fri, 23 May 1997 05:16:40 +0000 (05:16 +0000)] 
fix incorrect assertion

28 years agoNeed two offsets in the storeClientCopy stuff. The lower offset is
wessels [Fri, 23 May 1997 04:53:59 +0000 (04:53 +0000)] 
Need two offsets in the storeClientCopy stuff.  The lower offset is
represents the amount of data the client has seen and successfully
dealt with.  The upper offset represents data we have seen, but not dealt
with.  For example, incomplete HTTP reply headers.

28 years agoVCB typedef fixes
wessels [Fri, 23 May 1997 04:19:49 +0000 (04:19 +0000)] 
VCB typedef fixes
added storeSwapInCheck(), born from storeSwapInStart().
For unvalidated entries, need to check the filemap bits.

28 years agomemFreeData() not used
wessels [Fri, 23 May 1997 04:18:22 +0000 (04:18 +0000)] 
memFreeData() not used

28 years agounused variable
wessels [Fri, 23 May 1997 04:16:52 +0000 (04:16 +0000)] 
unused variable

28 years agoadd keepalive option
wessels [Thu, 22 May 1997 23:29:08 +0000 (23:29 +0000)] 
add keepalive option

28 years agofix body size calculation
wessels [Thu, 22 May 1997 23:28:51 +0000 (23:28 +0000)] 
fix body size calculation

28 years agoMerging in NOVM diffs. Making functions NOVM compatible.
wessels [Thu, 22 May 1997 21:51:50 +0000 (21:51 +0000)] 
Merging in NOVM diffs.  Making functions NOVM compatible.

Big change is nuking storeRegister() and having storeClientCopy()
do the same thing.  storeClientCopy fills a buffer passed in from
the caller and makes a callback when there is data in the buffer.

28 years agoattempt pipelining requests
wessels [Sat, 17 May 1997 21:59:45 +0000 (21:59 +0000)] 
attempt pipelining requests

28 years agoWarning if ipcache queue_length reaches NDnsServersAlloc
wessels [Fri, 16 May 1997 13:45:06 +0000 (13:45 +0000)] 
Warning if ipcache queue_length reaches NDnsServersAlloc

28 years agolog swap_file_number in store.log
wessels [Fri, 16 May 1997 13:44:57 +0000 (13:44 +0000)] 
log swap_file_number in store.log

28 years ago- change file_map_create debug level
wessels [Fri, 16 May 1997 13:44:32 +0000 (13:44 +0000)] 
- change file_map_create debug level

28 years ago- don't storeAddSwapDisk() again on reconfigure, instead call
wessels [Fri, 16 May 1997 13:43:46 +0000 (13:43 +0000)] 
- don't storeAddSwapDisk() again on reconfigure, instead call
  [new] storeReconfigureSwapDisk()

28 years ago- rename 'reread/reinitialize' to 'reconfigure'
wessels [Fri, 16 May 1997 13:42:49 +0000 (13:42 +0000)] 
- rename 'reread/reinitialize' to 'reconfigure'

28 years ago- rename 'reread/reinitialize' to 'reconfigure'
wessels [Fri, 16 May 1997 13:42:38 +0000 (13:42 +0000)] 
- rename 'reread/reinitialize' to 'reconfigure'
- s/first_time/configured_once/

28 years agofix swapfileno wierdness with 'SwapOutStartComplete', just set
wessels [Fri, 16 May 1997 06:19:30 +0000 (06:19 +0000)] 
fix swapfileno wierdness with 'SwapOutStartComplete', just set
e->swap_file_number instead of using some intermediate variable.
Remove storeDescribeStatus()
Make storeSwapInStart() global
Simplify storeLockObject(); make someone call storeSwapInStart() if
thats what they want.

28 years agono debug() from sigin handlers
wessels [Fri, 16 May 1997 05:43:31 +0000 (05:43 +0000)] 
no debug() from sigin handlers

28 years agoassert(dirn < ncache_dirs)
wessels [Fri, 16 May 1997 05:43:14 +0000 (05:43 +0000)] 
assert(dirn < ncache_dirs)

28 years agovoid fooStart()
wessels [Fri, 16 May 1997 05:43:12 +0000 (05:43 +0000)] 
void fooStart()

28 years agogindent
wessels [Fri, 16 May 1997 05:43:10 +0000 (05:43 +0000)] 
gindent

28 years agoopt_debug_stderr
wessels [Fri, 16 May 1997 05:40:58 +0000 (05:40 +0000)] 
opt_debug_stderr

28 years agoxmemcpy
wessels [Fri, 16 May 1997 05:40:25 +0000 (05:40 +0000)] 
xmemcpy

28 years agochange key for locked+released entries
wessels [Fri, 16 May 1997 05:40:15 +0000 (05:40 +0000)] 
change key for locked+released entries

28 years agoMake the fooStart() functions all void
wessels [Fri, 16 May 1997 05:37:58 +0000 (05:37 +0000)] 
Make the fooStart() functions all void
Since storeLockObject no longer does swapping in, we don't need to split
fooStart() into two functions separated by the swappin callback.

28 years agos/debug_trap/assert/
wessels [Fri, 16 May 1997 05:36:29 +0000 (05:36 +0000)] 
s/debug_trap/assert/

28 years agoDont' flush write buffers on write error, UNLESS there is no callback handler.
wessels [Fri, 16 May 1997 05:35:59 +0000 (05:35 +0000)] 
Dont' flush write buffers on write error, UNLESS there is no callback handler.
The callback should check the status return code and take necessary measures,
the disk module will keep the data queued.

I'm pretty sure we don't need to seek on disk files since they are only
open for reading or writing, never both.
    Remove all seek calls
    Remove FD_AT_EOF wierdness
    Remove wierd dread_ctrl->cur_len member