ObjFW  History of src/runtime/amiga-library.m of 96c24a6b6c9cd340

This will be moved to the "amiga-library" branch entirely in order to
not block the 1.0 release, as this is still far from finished. check-in: [f0b0aacf36] user: js, branch: trunk, size: 0

There is nothing left that needs errno. file: [6e1b91ee55] check-in: [8c90e88939] user: js, branch: trunk, size: 19209 [annotate] [blame] [check-ins using] [diff]

This prevents leaking resources like open files. file: [aa5ebeffcc] check-in: [169832c303] user: js, branch: trunk, size: 19267 [annotate] [blame] [check-ins using] [diff]

It's private and not needed. file: [1bd0f3b745] check-in: [661c36631a] user: js, branch: trunk, size: 18593 [annotate] [blame] [check-ins using] [diff]

This is theoretical, as this is happening in Forbid() and hence nothing
should be able to change it, even after FreeMem(). However, some
debugging tools might break Forbid(), and this commit fixes those. file: [ed1a8dcdb1] check-in: [4d470a65d4] user: js, branch: trunk, size: 18695 [annotate] [blame] [check-ins using] [diff]

If it is an inline, changing how tagged pointers work internally would
break the ABI and require recompiling everything, making future changes
harder. file: [3b48513197] check-in: [2fee820b5d] user: js, branch: trunk, size: 18390 [annotate] [blame] [check-ins using] [diff]

"make clean" not cleaning everything is what resulted in the missing
import to go unnoticed. file: [752043dfa7] check-in: [5d4e54363d] user: js, branch: amiga-library, size: 17799 [annotate] [blame] [check-ins using] [diff]

This avoids needing to #undef std{in,out,err} and then providing a new
symbol. file: [750189a19b] check-in: [514944f1ea] user: js, branch: amiga-library, size: 17773 [annotate] [blame] [check-ins using] [diff]

These are based on runtime/amiga-{library,glue,end}.m. file: [e1cb0d467e] check-in: [f6ab72da6e] user: js, branch: amiga-library, size: 17797 [annotate] [blame] [check-ins using] [diff]

tlskey.m used OFMapTable only on AmigaOS. This became a problem when
autorelease pools were moved into the runtime, as autorelease pools use
TLS. The build then broke, as there suddenly was a dependency from the
runtime on ObjFW.

This now uses the hashtable from the runtime and also no longer uses
OFList, thus fixing the build. As we always use the runtime on AmigaOS
anyway, this is fine. file: [3100a40497] check-in: [49aee5736e] user: js, branch: trunk, size: 17797 [annotate] [blame] [check-ins using] [diff]

amiga-gcc now links in __init_eh(). However, since we init EH ourselves,
there's no need for it, so we replace it with an empty one. file: [a59cd2d1dd] check-in: [7342a6e98b] user: js, branch: trunk, size: 16550 [annotate] [blame] [check-ins using] [diff]

This again works with both, SjLj and DWARF exceptions. The latest public
SDK still uses SjLj, so it still needs to be supported for now. file: [55ee69eb97] check-in: [153b825cdb] user: js, branch: trunk, size: 16523 [annotate] [blame] [check-ins using] [diff]

This is more compatible with the various conventions used on various
systems. file: [980704d720] check-in: [79fe29dbf4] user: js, branch: trunk, size: 16477 [annotate] [blame] [check-ins using] [diff]

sfdc does not support SysV functions and is not part of the MorphOS SDK. file: [7e60f6ed92] check-in: [8f9c2b55ce] user: js, branch: trunk, size: 16484 [annotate] [blame] [check-ins using] [diff]

Otherwise, it would be necessary to make unwinding work between 68k and
PPC code, and would get even worse with OS4 support.

This also includes using separate .sfd files. file: [1262807e23] check-in: [1e70711672] user: js, branch: trunk, size: 13965 [annotate] [blame] [check-ins using] [diff]

ObjFW_RT used to be a separate project that followed the BSD style, as
it was written in pure C, while ObjFW's style is based on the BSD style
with changes to make it a better fit for Objective-C. This commit
changes ObjFW_RT to use the same style as ObjFW. file: [9f8ec83747] check-in: [ef6d69931e] user: js, branch: trunk, size: 13963 [annotate] [blame] [check-ins using] [diff]

The only reason the runtime should ever terminate the process is because
a fatal error occurred, and for that it should use abort(). file: [a831aab769] check-in: [a37a78d51b] user: js, branch: trunk, size: 14024 [annotate] [blame] [check-ins using] [diff]

This has been fixed upstream in amiga-gcc. file: [cbe13c4c89] check-in: [b31bd48f90] user: js, branch: trunk, size: 14088 [annotate] [blame] [check-ins using] [diff]

The reason for this is that, in theory, this also supports versions
older than 3. file: [52c1c18633] check-in: [5831626f68] user: js, branch: trunk, size: 14148 [annotate] [blame] [check-ins using] [diff]

This currently uses __saveds on all exported functions for simplicity.
As an optimization, __saveds can be removed from some functions later
after careful verification that no code path starting from that function
accesses any data. file: [7a454ada46] check-in: [482698c5be] user: js, branch: trunk, size: 8021 [annotate] [blame] [check-ins using] [diff]