Author: eduard-bagdasaryan
Author Date: 2025-04-23 12:59:03 UTC

Do not treat tiny virgin response buffer space specially (#2056)

A long series of commits sprinkled Squid code with special treatment of
virgin response buffer space equal to one or two bytes:

* 2005 commit 2afaba07 adds a 2-byte hack to HttpStateData. The hack
  avoids stalling transactions by never asking
  StoreEntry::delayAwareRead() to read 1 byte.

* 2006 commit 253cacc adds the same 2-byte hack to Ftp::Client.

* 2007 commit 478cfe99 adds a 2-byte hack to
  Icap::ModXact::virginConsume() so that ICAP transactions do not stall
  when HttpStateData hack prevents new virgin bytes from being read into
  a BodyPipe buffer that has just one or two space bytes left. The ICAP
  hack was probably one-byte more aggressive than necessary...

* 2011 commit 0ad2b63b adds +1 byte hacks to ServerStateData and
  ClientHttpRequest to work around the same StoreEntry::delayAwareRead()
  problem leading to excessive buffering with slow clients.

* A followup 2012 commit 4dc2b072 adjusts StoreEntry::delayAwareRead()
  to allow 1-byte reads, making _all_ of the above hacks unnecessary.
  However, that commit only removed +1 byte hacks added in 2011, missing
  HttpStateData and ICAP hacks added in 2005, 2006 and 2007, probably
  because that 2012 work focused on fixing the bug introduced by 2011
  commit (that was triggered by slow client-Squid communication).

We now remove the remaining known hacks that worked around
StoreEntry::delayAwareRead() inability to read 1 byte (i.e. the first
three hacks on the above list).

These changes also happen to fix transaction stalls with read_ahead_gap
set to 1 and probably some other rare use cases: Transactions could
stall because inBuf.spaceSize() may be 1 for an emptied inBuf, leading
to maybeMakeSpaceAvailable() returning zero due to `read_size < 2` being
true. That zero result prevents HttpStateData from making progress by
reading new virgin response bytes. Mistreatment of "emptied inBuf" cases
was exposed by 2023 commit 50c5af88 changes. More work is needed to
address other, more prominent "emptied inBuf" cases.

Author: eduard-bagdasaryan
Author Date: 2025-04-23 12:59:03 UTC

Do not treat tiny virgin response buffer space specially (#2056)

A long series of commits sprinkled Squid code with special treatment of
virgin response buffer space equal to one or two bytes:

* 2005 commit 2afaba07 adds a 2-byte hack to HttpStateData. The hack
  avoids stalling transactions by never asking
  StoreEntry::delayAwareRead() to read 1 byte.

* 2006 commit 253cacc adds the same 2-byte hack to Ftp::Client.

* 2007 commit 478cfe99 adds a 2-byte hack to
  Icap::ModXact::virginConsume() so that ICAP transactions do not stall
  when HttpStateData hack prevents new virgin bytes from being read into
  a BodyPipe buffer that has just one or two space bytes left. The ICAP
  hack was probably one-byte more aggressive than necessary...

* 2011 commit 0ad2b63b adds +1 byte hacks to ServerStateData and
  ClientHttpRequest to work around the same StoreEntry::delayAwareRead()
  problem leading to excessive buffering with slow clients.

* A followup 2012 commit 4dc2b072 adjusts StoreEntry::delayAwareRead()
  to allow 1-byte reads, making _all_ of the above hacks unnecessary.
  However, that commit only removed +1 byte hacks added in 2011, missing
  HttpStateData and ICAP hacks added in 2005, 2006 and 2007, probably
  because that 2012 work focused on fixing the bug introduced by 2011
  commit (that was triggered by slow client-Squid communication).

We now remove the remaining known hacks that worked around
StoreEntry::delayAwareRead() inability to read 1 byte (i.e. the first
three hacks on the above list).

These changes also happen to fix transaction stalls with read_ahead_gap
set to 1 and probably some other rare use cases: Transactions could
stall because inBuf.spaceSize() may be 1 for an emptied inBuf, leading
to maybeMakeSpaceAvailable() returning zero due to `read_size < 2` being
true. That zero result prevents HttpStateData from making progress by
reading new virgin response bytes. Mistreatment of "emptied inBuf" cases
was exposed by 2023 commit 50c5af88 changes. More work is needed to
address other, more prominent "emptied inBuf" cases.

Author: squidadm
Author Date: 2025-04-23 07:37:02 UTC

Source Format Enforcement (#2041)

Author: squidadm
Author Date: 2025-04-17 17:10:39 UTC

Source Format Enforcement (#2043)

Reference point for automated CONTRIBUTORS updates

Author: Amos Jeffries
Author Date: 2024-09-30 11:19:58 UTC

5.10 final

Author: Amos Jeffries
Author Date: 2022-08-09 23:34:54 UTC

Bug 3193 pt2: NTLM decoder truncating strings (#1114)

The initial bug fix overlooked large 'offset' causing integer
wrap to extract a too-short length string.

Improve debugs and checks sequence to clarify cases and ensure
that all are handled correctly.

Author: Armin Wolfermann
Author Date: 2020-02-04 20:15:00 UTC

fix security patch

Author: Francesco Chemolli
Author Date: 2018-10-21 13:06:33 UTC

Mingw build fixes

There are type errors and data access errors in windows-specific bits of
code.

Author: Garri Djavadyan
Author Date: 2016-12-17 13:56:49 UTC

Bug 4169: HIT marked as MISS when If-None-Match does not match

Author: Garri Djavadyan
Author Date: 2016-12-17 12:59:41 UTC

Bug 4169: HIT marked as MISS when If-None-Match does not match

Author: Garri Djavadyan
Author Date: 2016-12-17 10:04:42 UTC

Bug 4169: HIT marked as MISS when If-None-Match does not match

Author: Amos Jeffries
Author Date: 2016-12-17 06:37:44 UTC

Do not override user defined -std option

Author: Source Maintenance
Author Date: 2015-04-23 13:02:10 UTC

Docs: Update CONTRIBUTORS

Author: Amos Jeffries
Author Date: 2012-08-08 08:16:36 UTC

merged from trunk

Author: serassio <>
Author Date: 2008-03-02 23:54:50 UTC

Windows port: Add reconfigure action to cachemgr.

Author: robertc <>
Author Date: 2002-10-24 20:53:10 UTC

cbdata debug history