apt, branch 1.3.5

Release 1.3.5

2017-02-27T14:25:21Z

CMake: Install statvfs.h to include/sys, not just include/

2017-02-22T15:53:46Z

We are including sys/statvfs.h, not statvfs.h, so make sure our dummy in the correct spot. (cherry picked from commit c050f4fa30ccda637f3e05edddf750b768c8590d)

Don't use -1 fd and AT_SYMLINK_NOFOLLOW for faccessat()

2017-02-22T15:53:46Z

-1 is not an allowed value for the file descriptor, the only allowed non-file-descriptor value is AT_FDCWD. So use that instead. AT_SYMLINK_NOFOLLOW has a weird semantic: It checks whether we have the specified access on the symbolic link. It also is implemented only by glibc on Linux, so it's inherently non-portable. We should just drop it. Thanks: James Clarke for debugging these issues Reported-by: James Clarke (cherry picked from commit 25f54c960d7a4ceca7bd3e21f87baf48d6cbc2d3)

Do not package names representing .dsc/.deb/... files

2017-02-22T15:53:46Z

In the case of build-dep and other commands where a file can be passed we must make sure not to normalize the path name as that can have odd side effects, or well, cause the operation to do nothing. Test for build-dep-file is adjusted to perform the vcard check once as "vcard" and once as "VCard", thus testing that this solves the reported bug. We inline the std::transform() and optimize it a bit to not write anything in the common case (package names are defined to be lowercase, the whole transformation is just for names that should not exist...) to counter the performance hit of the added find() call (it's about 0.15% more instructions than with the existing transform, but we save about 0.67% in writes...). Closes: #854794 (cherry picked from commit 85ee4036c68d8ecd2c973d413a17aca81380900b)

Only merge acquire items with the same meta key

2017-02-22T15:53:46Z

Since the introduction of by-hash, two differently named files might have the same real URL. In our case, the files icons-64x64.tar.gz and icons-128x128.tar.gz of empty tarballs. APT would try to merge them and end with weird errors because it completed the first download and enters the second stage for decompressing and verifying. After that it would queue a new item to copy the original file to the location, but that copy item would be in the wrong stage, causing it to use the hashes for the decompressed item. Closes: #838441 (cherry picked from commit 7b78e8bef1fc9de22d826db1db9df25f97d3710c)

COPYING.GPL: Update to recent version (address, LGPL name)

2017-02-22T15:53:46Z

Just copied over from common-licenses. Seems we missed to do that earlier. Gbp-Dch: ignore (cherry picked from commit 84285d17bab32a0ceafe31a5b2be61cc4f520b42)

travis: Do not build documentation in root job

2017-02-22T15:53:46Z

This speeds up testing things as root, which is good, because we usually test as user. Gbp-Dch: ignore (cherry picked from commit 17dadc40aefbbdd002b864edd157e4795e54fba3)

travis: Run test suites for root and user in separate build jobs

2017-02-22T15:53:46Z

This hopefully cuts down on the test time. Optimally, we'd just have one build job and parallize, but that requires a tty or something, probably due to GNU parallel? Gbp-Dch: ignore (cherry picked from commit 9b7c71f145e51c2d655ef09fca434d02db08331d)

basehttp: Only read Content-Range on 416 and 206 responses

2017-02-22T15:53:46Z

This fixes issues with sourceforge where the redirector includes such a Content-Range in a 302 redirect. Since we do not really know what file is meant in a redirect, let's just ignore it for all responses other than 416 and 206. Maybe we should also get rid of the other errors, and just ignore the field in those cases as well? LP: #1657567 (cherry picked from commit 4759a702081297bde66982efed8b2b7fd39ca27c)

stop rred from leaking debug messages on recovered errors

2017-02-22T15:53:46Z

rred can fail for a plentory of reasons, but its failure is usually recoverable (Ign lines) so it shouldn't leak unrequested debug messages to an observing user. Closes: #850759 (cherry picked from commit 2984d7aec37e09b473c7b99f43d20622c25dc99d)