(cherry picked from commit 50dd7fea)

When running repack while doing repo writes, `packfile_load__cb()` can see some temporary files in the directory that are bigger than the usual, and makes `memcmp` overflow on the `p->pack_name` string. ASAN detected this. This just uses `strncmp`, that should not have any performance impact and is safe for comparing strings of different sizes.

```
ERROR: AddressSanitizer: heap-buffer-overflow on address 0x61200001a3f3 at pc 0x7f4a9e1976ec bp 0x7ffc1f80e100 sp 0x7ffc1f80d8b0
READ of size 89 at 0x61200001a3f3 thread T0
SCARINESS: 26 (multi-byte-read-heap-buffer-overflow)
    #0 0x7f4a9e1976eb in __interceptor_memcmp.part.78 (/build/cfgr-admin#link-tree/libtools_build_sanitizers_asan-ubsan-py.so+0xcf6eb)
    #1 0x7f4a518c5431 in packfile_load__cb /build/libgit2/0.27.0/src/libgit2-0.27.0/src/odb_pack.c:213
    #2 0x7f4a518d9582 in git_path_direach /build/libgit2/0.27.0/src/libgit2-0.27.0/src/path.c:1134
    #3 0x7f4a518c58ad in pack_backend__refresh /build/libgit2/0.27.0/src/libgit2-0.27.0/src/odb_pack.c:347
    #4 0x7f4a518c1b12 in git_odb_refresh /build/libgit2/0.27.0/src/libgit2-0.27.0/src/odb.c:1511
    #5 0x7f4a518bff5f in git_odb__freshen /build/libgit2/0.27.0/src/libgit2-0.27.0/src/odb.c:752
    #6 0x7f4a518c17d4 in git_odb_stream_finalize_write /build/libgit2/0.27.0/src/libgit2-0.27.0/src/odb.c:1415
    #7 0x7f4a51b9d015 in Repository_write /build/pygit2/0.27.0/src/pygit2-0.27.0/src/repository.c:509
```

(cherry picked from commit d22cd1f4)

The current error handling for the multiline variable parser is a bit
fragile, as each error condition has its own code to clear memory.
Instead, unify error handling as far as possible to avoid this
repetitive code. While at it, make use of `GITERR_CHECK_ALLOC` to
correctly handle OOM situations and verify that the buffer we print into
does not run out of memory either.

(cherry picked from commit bc63e1ef)

(cherry picked from commit 38b85255)

(cherry picked from commit a03113e8)

Release v0.26.6

OSS-fuzz has reported a potential out-of-bounds read when processing a
"ng" smart packet:

==1==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x6310000249c0 at pc 0x000000493a92 bp 0x7ffddc882cd0 sp 0x7ffddc882480
	READ of size 65529 at 0x6310000249c0 thread T0
	SCARINESS: 26 (multi-byte-read-heap-buffer-overflow)
	#0 0x493a91 in __interceptor_strchr.part.35 /src/llvm/projects/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc:673
	#1 0x813960 in ng_pkt libgit2/src/transports/smart_pkt.c:320:14
	#2 0x810f79 in git_pkt_parse_line libgit2/src/transports/smart_pkt.c:478:9
	#3 0x82c3c9 in git_smart__store_refs libgit2/src/transports/smart_protocol.c:47:12
	#4 0x6373a2 in git_smart__connect libgit2/src/transports/smart.c:251:15
	#5 0x57688f in git_remote_connect libgit2/src/remote.c:708:15
	#6 0x52e59b in LLVMFuzzerTestOneInput /src/download_refs_fuzzer.cc:145:9
	#7 0x52ef3f in ExecuteFilesOnyByOne(int, char**) /src/libfuzzer/afl/afl_driver.cpp:301:5
	#8 0x52f4ee in main /src/libfuzzer/afl/afl_driver.cpp:339:12
	#9 0x7f6c910db82f in __libc_start_main /build/glibc-Cl5G7W/glibc-2.23/csu/libc-start.c:291
	#10 0x41d518 in _start

When parsing an "ng" packet, we keep track of both the current position
as well as the remaining length of the packet itself. But instead of
taking care not to exceed the length, we pass the current pointer's
position to `strchr`, which will search for a certain character until
hitting NUL. It is thus possible to create a crafted packet which
doesn't contain a NUL byte to trigger an out-of-bounds read.

Fix the issue by instead using `memchr`, passing the remaining length as
restriction. Furthermore, verify that we actually have enough bytes left
to produce a match at all.

OSS-Fuzz-Issue: 9406

Travis has upgraded the default Xcode images from 8.3 to 9.4 on 31st
July 2018, including an upgrade to macOS 10.13. Unfortunately, this
breaks our CI builds on our maintenance branches. As we do not want to
include mayor changes to fix the integration right now, we force use of
the old Xcode 8.3 images.

Release v0.26.5

When checking whether a delta base offset and length fit into the base
we have in memory already, we can trigger an overflow which breaks the
check. This would subsequently result in us reading memory from out of
bounds of the base.

The issue is easily fixed by checking for overflow when adding `off` and
`len`, thus guaranteeting that we are never indexing beyond `base_len`.
This corresponds to the git patch 8960844a7 (check patch_delta bounds
more carefully, 2006-04-07), which adds these overflow checks.

Reported-by: Riccardo Schirone <rschiron@redhat.com>

When computing the offset and length of the delta base, we repeatedly
increment the `delta` pointer without checking whether we have advanced
past its end already, which can thus result in an out-of-bounds read.
Fix this by repeatedly checking whether we have reached the end. Add a
test which would cause Valgrind to produce an error.

Reported-by: Riccardo Schirone <rschiron@redhat.com>
Test-provided-by: Riccardo Schirone <rschiron@redhat.com>

Our delta code was originally adapted from JGit, which itself adapted it
from git itself. Due to this heritage, we inherited a bug from git.git
in how we compute the delta offset, which was fixed upstream in
48fb7deb5 (Fix big left-shifts of unsigned char, 2009-06-17). As
explained by Linus:

    Shifting 'unsigned char' or 'unsigned short' left can result in sign
    extension errors, since the C integer promotion rules means that the
    unsigned char/short will get implicitly promoted to a signed 'int' due to
    the shift (or due to other operations).

    This normally doesn't matter, but if you shift things up sufficiently, it
    will now set the sign bit in 'int', and a subsequent cast to a bigger type
    (eg 'long' or 'unsigned long') will now sign-extend the value despite the
    original expression being unsigned.

    One example of this would be something like

            unsigned long size;
            unsigned char c;

            size += c << 24;

    where despite all the variables being unsigned, 'c << 24' ends up being a
    signed entity, and will get sign-extended when then doing the addition in
    an 'unsigned long' type.

    Since git uses 'unsigned char' pointers extensively, we actually have this
    bug in a couple of places.

In our delta code, we inherited such a bogus shift when computing the
offset at which the delta base is to be found. Due to the sign extension
we can end up with an offset where all the bits are set. This can allow
an arbitrary memory read, as the addition in `base_len < off + len` can
now overflow if `off` has all its bits set.

Fix the issue by casting the result of `*delta++ << 24UL` to an unsigned
integer again. Add a test with a crafted delta that would actually
succeed with an out-of-bounds read in case where the cast wouldn't
exist.

Reported-by: Riccardo Schirone <rschiron@redhat.com>
Test-provided-by: Riccardo Schirone <rschiron@redhat.com>

Release v0.26.4

We might modify caches due to us trying to load the configuration to figure out
what kinds of filesystem protections we should have.

We still compare case-insensitively to protect more thoroughly as we don't know
what specifics we'll see on the system and it's the behaviour from git.

When dealing with `core.proectNTFS` and `core.protectHFS` we do check
against `.gitmodules` but we still have a failing test as the non-filesystem
codepath does not check for it.

Any part of the library which asks the question can pass in the mode to have it
checked against `.gitmodules` being a symlink.

This is particularly relevant for adding entries to the index from the worktree
and for checking out files.

This is so we have it available for the path validity checking. In a later
commit we will start rejecting `.gitmodules` files as symlinks.

We may take in names from the middle of a string so we want the caller to let us
know how long the path component is that we should be checking.

We want to reject these as they cause compatibility issues and can lead to git
writing to files outside of the repository.

These will be used by the checkout code to detect them for the particular
filesystem they're on.

These can't go into the public API yet as we don't want to introduce API or ABI
changes in a security release.

Given a path component it knows what to pass to the filesystem-specific
functions so we're protected even from trees which try to use the 8.3 naming
rules to get around us matching on the filename exactly.

The logic and test strings come from the equivalent git change.

It checks against the 8.3 shortname variants, including the one which includes
the checksum as part of its name.

This lets us check for other kinds of reserved files.

Otherwise we would also admit `..\..\foo\bar` as a valid path and fail to
protect Windows users.

Ideally we would check for both separators without the need for the copied
string, but this'll get us over the RCE.

If the we decide that the "name" of the submodule (i.e. its path inside
`.git/modules/`) is trying to escape that directory or otherwise trick us, we
ignore the configuration for that submodule.

This leaves us with a half-configured submodule when looking it up by path, but
it's the same result as if the configuration really were missing.

The name check is potentially more strict than it needs to be, but it lets us
re-use the check we're doing for the checkout. The function that encapsulates
this logic is ready to be exported but we don't want to do that in a security
release so it remains internal for now.

We should pretend such submdules do not exist as it can lead to RCE.

Update the settings to use a specific read-only token for accessing our
test repositories in Bitbucket.

At present, we have three online tests against bitbucket: one which
specifies the credentials in the payload, one which specifies the
correct credentials in the URL and a final one that specifies the
incorrect credentials in the URL.  Bitbucket has begun responding to the
latter test with a 403, which causes us to fail.

Break these three tests into separate tests so that we can skip the
latter until this is resolved on Bitbucket's end or until we can change
the test to a different provider.