GitLab

Detail:
  Drop COMPAT_INET4 support
Admin:
  aof/ShareFS built with cc 5.89 in the All32 environment
  aof/ShareFSSA built with cc 5.89 in the Disc environment

Version 3.61. Tagged as 'ShareFSBlob-3_61'

aof/ShareFS built with cc 5.80 in the All32 environment
aof/ShareFSSA built with cc 5.80 in the Disc environment

Version 3.60. Tagged as 'ShareFSBlob-3_60'

As per bug 448, the "Free" menu item on each filer should have an ellipsis because it opens a persistent dialogue box.

Detail:
  Add a binary of the standalone version of the module with supporting code to manage registration with ResourceFS, so that a disc based module can be created.
  Add the corresponding install rule to the BBE makefile.
Admin
  Built using Disc environment & cc 5.78.

Admin:
  Mirror the change in srcbuild-0_54 manually.
  Tagged as ShareFS-3_59-1

Detail:
  * Use ShareFSFiler$DefaultDir rather than ...Path
  * Avert possible buffer overflows in append_dotdefault_path()
  * Remove unused source file and function containing SWP (not ARMv8 compatible)
Admin:
  Submission from Rob Sprowson.
  Binary built using All32 environment & cc 5.71.

Version 3.59. Tagged as 'ShareFS-3_59'

Detail:
  The errno lookup function was also including the last OS error, but not
checking if there actually was an OS error, leading to an address of
&00000004 being passed to MessageTrans to substitute.
  Since the C library (and TCP/IP library) errors aren't in the same number
space, just remove this unrelated text from the function and say "Network
error" instead.
  Also modified errno lookup to get the message token right for errors > 99.
Admin:
  Submission from Rob Sprowson.
  Binary built using All32 environment & cc 5.71.

Admin:
  This is still a binary import. Fix provided by J Ballance.

Version 3.57. Tagged as 'ShareFS-3_57'

Previously held in AUNMsgs, moved here and installed during the resources phase.
Tagged as ShareFS-3_55-OwnResources.

Detail:
  Various command line and filename buffers enlarged to accommodate the
  larger sizes permitted since RISC OS 4.
Admin:
  This is still a binary import. Initial fault detected by J Ballance using
  fsbash, but in practice a large number of buffers within ShareFS were
  affected.

Version 3.55. Tagged as 'ShareFS-3_55'

Rebuilt with cc 5.71 in an All32 environment.
Tagged as ShareFS-3_54-2.

Retagged as 'ShareFS-3_54'

Admin:
  Typo correction submitted by Jeffrey Lee

Version 3.54. Tagged as 'ShareFS-3_54'

Detail:
  This is a binary import for licensing reasons. Source changes received from
  Jeffrey Lee under NDA. His notes:
  * fslib/h/evil - Removed ESCAPE_BYTE & IOC_IMASKA, changes MONOTIME to be a
    call to OS_ReadMonotonicTime
  * fslib/c/esc, fslib/h/esc - Change testesc() to call OS_ReadEscapeState
    instead of using ESCAPE_BYTE #define
  * ShareFS/c/remotefs, ShareFS/c/window - tweaked usage of MONOTIME to avoid
    calling OS_ReadMonotonicTime more times than necessary (just in case there
    was some method to Acorn's madness, and they were peeking zero page due to
    performance concerns). Functions now tend to cache the time in a
    'monotime' local variable.
  * ShareFS/c/window - There was a loop in domsg() which would just sit and
    burn CPU time waiting for the monotonic time to update if the new client
    key (which is a hash of MONOTIME) was identical to the old one. This has
    now been changed to use Portable_Idle, so will be a bit nicer on machines
    which support wait-for-interrupt. I'm not sure how often that loop
    actually does anything, but after spotting it I just couldn't leave it
    there as it was!
Admin:
  Jeffrey has tested this on both his Pi (with high processor vectors) and
  Iyonix (without), in both client & server roles.

Version 3.53. Tagged as 'ShareFS-3_53'

Detail:
  Previous build of this binary blob used unaligned LDR instructions
  which behave differently on ARMv7.
Admin:
  No functional change. Untested.

Tagged as 'ShareFS-3_52-2'

This is the result of executing the BBE build phase on a complete Tungsten
ROM build. It may be linked into any other compatible ROM builds in the same
way as if the source was available, but obviously it cannot be studied or
improved upon in its current form, and it is not portable to any incompatible
architecture variants or ABIs.
It is hoped that in future this will be replaced by a full source release.