ObjFW  History of src/OFZIPArchive.m of d998e3c492e9b279

Some names are too generic and might cause conflicts (e.g. zoo). file: [8040666849] check-in: [13f0321eff] user: js, branch: trunk, size: 28746 [annotate] [blame] [check-ins using] [diff]

This allows for a much cleaner solution to avoid the internal read
buffer of e.g. a TLS connection never being processed while waiting for
a delimiter. file: [76b549ca53] check-in: [a61e0594b4] user: js, branch: trunk, size: 28718 [annotate] [blame] [check-ins using] [diff]

Unfortunately, I could not find a file to test this with, as this is an
extremely rare corner case that basically doesn't seem to exist with
real world files: It requires an extremely small part size for an
archive with thousands of files. file: [03f366c882] check-in: [a847357585] user: js, branch: trunk, size: 28748 [annotate] [blame] [check-ins using] [diff]

This cannot yet extract a file from the archive if the file spans
multiple parts. However, for a multipart archive with multiple files,
all files that only span a single part can be extracted. file: [460801d892] check-in: [630ea6bc5a] user: js, branch: trunk, size: 26479 [annotate] [blame] [check-ins using] [diff]

It's generally uncommon to prefix those.

Additionally, it might be worthwhile to try to get those actually
registered, which requires usage. file: [7434ba6daa] check-in: [9b3cae6cba] user: js, branch: trunk, size: 24619 [annotate] [blame] [check-ins using] [diff]

In order to not create a retain cycle, this changes the reference to the
last returned stream to an unsafe unretained one that the stream itself
needs to reset to nil in its dealloc.

Since when closing a stream for writing the archive needs to be updated,
the code for that was moved out of OFZIPArchive and into the stream
writer, since it now has a reference to the archive anyway. file: [9d548c4f33] check-in: [6527c97d03] user: js, branch: trunk, size: 24622 [annotate] [blame] [check-ins using] [diff]

This allows constructing a URI to access a file within an archive. file: [d0b4f286d3] check-in: [19bbbbd95d] user: js, branch: trunk, size: 24769 [annotate] [blame] [check-ins using] [diff]

This is for consistency with OFTarArchiveEntry and OFLHAArchiveEntry. file: [e9a337a52f] check-in: [94c326ab78] user: js, branch: trunk, size: 24010 [annotate] [blame] [check-ins using] [diff]

The old write API made it too easy to lose bytes when a stream is set to
non-blocking mode. The new API always throws when not all bytes were
written, which forces handling the number of bytes being written being
smaller than the number of bytes requested to be written. file: [a5dd4c3dce] check-in: [d1d36ae522] user: js, branch: trunk, size: 23736 [annotate] [blame] [check-ins using] [diff]

This is only migrating some places, others will be migrated as they are
touched. file: [c62bbdf049] check-in: [a0fd103a0b] user: js, branch: trunk, size: 23867 [annotate] [blame] [check-ins using] [diff]

This means refusing to close twice, calling -[close] from -[dealloc] and
not calling -[cancelAsyncRequests].

Calling -[cancelAsyncRequests] in -[close] is too dangerous, as -[close]
gets called by -[dealloc]: If the queue is the last reference to the
object, at the point where -[cancelAsyncRequests] removes it from the
queue, the object will start to deallocate and call into
-[cancelAsyncRequests] again, which is still in the middle of removing
it and now finds itself with an inconsistent state. file: [f2e2222c23] check-in: [3b43d51006] user: js, branch: trunk, size: 23778 [annotate] [blame] [check-ins using] [diff]

With this property, it's possible to specify whether the memory should
be protected from swapping or not. This makes it easier for e.g. the
crypto hash classes: They can now just always use OFSecureData without
wasting unswappable memory if it's not needed. file: [f3370d3321] check-in: [e629dc83a9] user: js, branch: trunk, size: 23531 [annotate] [blame] [check-ins using] [diff]

This also adds a parameter for the actual and expected checksum. file: [9e27ae6fff] check-in: [8f21e8acb5] user: js, branch: trunk, size: 23570 [annotate] [blame] [check-ins using] [diff]

Instead, explicitly declare them, as OF_ASSUME_NONNULL_{BEGIN,END} does
not apply to implicit return types. This means that after this commit,
all init methods have a nonnull return type, as they should have. file: [f82daa6eae] check-in: [2f4e0df8be] user: js, branch: trunk, size: 22669 [annotate] [blame] [check-ins using] [diff]

Interestingly, Clang did not warn about this, but GCC does. file: [e300d0925e] check-in: [1f4407ddf1] user: js, branch: trunk, size: 22475 [annotate] [blame] [check-ins using] [diff]

This was accidentally changed when ZIP64 write support was added. file: [85b7d55558] check-in: [6a474ba752] user: js, branch: trunk, size: 22462 [annotate] [blame] [check-ins using] [diff]

Treating the ZIP64 extra field as part of the entry itself makes things
a lot easier when (re)writing the central directory. file: [c1fa8a1f8d] check-in: [8ebf6ed443] user: js, branch: trunk, size: 21490 [annotate] [blame] [check-ins using] [diff]

This is now only necessary for reading and appending. file: [a7134be3c5] check-in: [c404c33cf1] user: js, branch: trunk, size: 14579 [annotate] [blame] [check-ins using] [diff]

This is in preparation for OFTarArchive write support. file: [77923608cb] check-in: [3f34a360bf] user: js, branch: trunk, size: 14326 [annotate] [blame] [check-ins using] [diff]

This removes "b" for binary and always uses binary, as there is no good
reason to not use binary. file: [8c3891b987] check-in: [23e57c5040] user: js, branch: trunk, size: 13822 [annotate] [blame] [check-ins using] [diff]

There is not much point in having a bi-directional stream for Deflate
compression / decompression. file: [1d8f8c26f4] check-in: [f56a50ee50] user: js, branch: trunk, size: 13823 [annotate] [blame] [check-ins using] [diff]

Instead, let reads return 0 and let writes append after the end. file: [4f8a23b8a4] check-in: [f9cd4f9cab] user: js, branch: trunk, size: 13823 [annotate] [blame] [check-ins using] [diff]

This matches Apple's style. file: [af4fea4ffc] check-in: [6b77a5dd8b] user: js, branch: trunk, size: 13832 [annotate] [blame] [check-ins using] [diff]

Casts are now written like types in variable declarations. file: [df3e45938b] check-in: [4af49a13c3] user: js, branch: trunk, size: 13892 [annotate] [blame] [check-ins using] [diff]

Forgot to add 2017, even though I already did quite some changes in
2017. file: [16b7ebc50b] check-in: [44f45c2e35] user: js, branch: trunk, size: 13876 [annotate] [blame] [check-ins using] [diff]

The minimum version in the central directory is the minimum version
required to parse the central directory entry, while the minimum version
in the local header is the minimum version required to extract the file. file: [1422bca77d] check-in: [f013a6748a] user: js, branch: trunk, size: 13870 [annotate] [blame] [check-ins using] [diff]

Right now, this is only useful for OFTCPSocket, as this is the only
class so far not establishing the stream in the init method. However,
this adds the general infrastructure to allow reuse to all subclasses of
OFStream. file: [c12f6e9a97] check-in: [1de551cb5f] user: js, branch: trunk, size: 13916 [annotate] [blame] [check-ins using] [diff]

The reason for renaming to OFInflateStream was to have one stream for
decompression and one for compression in order to reduce memory usage if
only one of the two is needed, as the ivar layout will be smaller then.
However, it is more consistent with other stream classes to have one
stream that can handle both. The increased memory footprint of having
ivars for compression and decompression can be solved by having a
pointer to those instead. This will not incur any performance penalty,
as the pointer will be dereferenced instead of the ivars, meaning the
overhead is only getting the pointer from the ivars once. file: [f2af10f3e4] check-in: [134a1121c7] user: js, branch: trunk, size: 13899 [annotate] [blame] [check-ins using] [diff]

OFZIPArchive only needs an OFSeekableStream. file: [5b4c2f9e03] check-in: [080eb2f31b] user: js, branch: trunk, size: 13899 [annotate] [blame] [check-ins using] [diff]

(U)INT_FAST*_MAX is broken on Android, which makes (u)int_fast*_t
dangerous to use while it does not provide any real speed benefit. file: [b5e5e4f9e4] check-in: [d3158d091f] user: js, branch: trunk, size: 14025 [annotate] [blame] [check-ins using] [diff]

While at it, also update the mail address. file: [3d3f1ac742] check-in: [cec0f072f8] user: js, branch: 0.8, size: 14067 [annotate] [blame] [check-ins using] [diff]

While at it, also update the mail address. file: [c97bd73161] check-in: [2a27cf3000] user: js, branch: trunk, size: 14040 [annotate] [blame] [check-ins using] [diff]

Now that we can require GCC >= 4.6, we no longer need to fall back to
using a category on old compilers. file: [217e433e7c] check-in: [2aca549d60] user: js, branch: trunk, size: 14036 [annotate] [blame] [check-ins using] [diff]

This increases the required GCC version from 4.0 to 4.6 (exception:
Apple GCC, which already supports this with >= 4.0 starting with OS X
10.5). Since even GCC 4.6 is really old by now, there is no point in
still supporting something even older and making the code ugly because
of that. While some hardware and OS support was dropped from GCC 4.6
compared to GCC 4.0, there is nothing in there that would be an
interesting target with the exception of BeOS maybe - but a port to BeOS
can also be achieved using the Haiku support. The other dropped OSes are
mostly old versions of OSes while newer ones are still being supported
(and those newer versions of those OSes still support the same
hardware). file: [e3768699d3] check-in: [48980f2297] user: js, branch: trunk, size: 14055 [annotate] [blame] [check-ins using] [diff]

It makes more sense to have one class for inflating and one for
deflating instead of having one that does both, as having one that does
both requires more memory per instance when only inflating or deflating
is used. file: [da28dbf09c] check-in: [2e9336b1ef] user: js, branch: trunk, size: 13909 [annotate] [blame] [check-ins using] [diff]

Every OFSeekableStream can be used now. file: [6109b2512a] check-in: [aa1bb213e0] user: js, branch: trunk, size: 13896 [annotate] [blame] [check-ins using] [diff]

Some implementations set some bits only in either the Local File Header
or the Central Directory Entry, but not in both. file: [ec81a35432] check-in: [178b39b840] user: js, branch: trunk, size: 13405 [annotate] [blame] [check-ins using] [diff]

The old behaviour where the exception would access errno directly on
creation of the exception was very fragile. The two main problems with
it were that sometimes it would pick up an errno even though none had
been set and in other cases that when the exception was created errno
had already been overridden.

This also greatly increases errno handling on Win32, especially in
conjunction with sockets. It can still be improved further, though. file: [f5b578d7a5] check-in: [62e2de30b9] user: js, branch: trunk, size: 13472 [annotate] [blame] [check-ins using] [diff]

The reason for this is that some systems, for example Android, use
32 bits for off_t while having size related fields like st_size as
long long. file: [21935bfcfb] check-in: [46b89b1c8e] user: js, branch: trunk, size: 13442 [annotate] [blame] [check-ins using] [diff]

This means that OFObject.h imports macros.h now, making it unnecessary
to manually import macros.h in almost every file. And while at it, also
import autorelease.h in OFObject.h, so that this doesn't need to be
manually imported in almost every file as well. file: [e1f5b7ea93] check-in: [13ee56edf3] user: js, branch: trunk, size: 13366 [annotate] [blame] [check-ins using] [diff]

This belongs to OFZIPArchiveEntry. Consequently, it has been renamed to
of_zip_archive_entry_find_extra_field() and made public, so that it can
be used by applications that got the extra field from the
OFZIPArchiveEntry. file: [3ed67d1bc6] check-in: [adf754d5fc] user: js, branch: trunk, size: 13256 [annotate] [blame] [check-ins using] [diff]

This is in preparation for adding UDP sockets, as UDP sockets and TCP
sockets have no common superclass, as one is stream-oriented while the
other is packet-oriented.

Read and write exceptions are for any object now, as they are useful for
a lot more than just for streams, while the others (bind, listen, etc.)
are for any socket now (the type is id in this case, though, as there is
no common superclass). file: [85d6f8f249] check-in: [8d2a5052fd] user: js, branch: trunk, size: 13930 [annotate] [blame] [check-ins using] [diff]

While sorting -[entries] reduces hard disk seeks, it allows a denial of
service by creating an archive with a huge central directory without
actual files. As usually the order in the central directory matches the
order of the actual files, this minor speed increase is not worth the
attack vector. file: [20c9faa50f] check-in: [be628bbb84] user: js, branch: trunk, size: 13808 [annotate] [blame] [check-ins using] [diff]

The higher byte identifies the OS and is not helpful in the comparison,
as it's better to just try to decompress it instead of checking if we
are on the specified OS. file: [b1f97cb82f] check-in: [1ce5d53f93] user: js, branch: trunk, size: 9924 [annotate] [blame] [check-ins using] [diff]

Archive comments allow up to 65535 bytes after the end of central
directory record, thus the magic bytes for the end of central directory
record are not always at end - 22. file: [8bbe969b1a] check-in: [e2ad3ceacb] user: js, branch: trunk, size: 9897 [annotate] [blame] [check-ins using] [diff]

Also adds support for data descriptors. file: [0ac2251133] check-in: [53e46a8326] user: js, branch: trunk, size: 9717 [annotate] [blame] [check-ins using] [diff]

Still a lot to do, see list of FIXMEs in the source. file: [f436d4f99e] check-in: [1c78b3a4db] user: js, branch: trunk, size: 10974 [annotate] [blame] [check-ins using]