apt, branch 1.1.7Debians commandline package managerhttps://git.kalnischkies.de/apt/atom?h=1.1.72015-12-26T18:11:10ZRelease 1.1.72015-12-26T18:11:10ZJulian Andres Klodejak@debian.org2015-12-26T18:10:21Zurn:sha1:25203cbbed0b7ebc3eb75fa96ab3ed53be148168Get rid of memmove() in our read buffering2015-12-26T16:38:40ZJulian Andres Klodejak@debian.org2015-12-26T16:38:40Zurn:sha1:83e22e26f9f10472aed97f889967c86ee218d28d
This further improves our performance, and rred on uncompressed
files now spents 78% of its time in writing. Which means that
we should really look at buffering those.
rred: Allow passing files as arguments for compressor testing2015-12-26T14:29:00ZJulian Andres Klodejak@debian.org2015-12-26T14:25:33Zurn:sha1:46cddb8c1a393cc7d03fbd897f2504ae9f47179d
This introduces a -t mode in which the first argument is input,
the second is output and the remaining are diffs.
This allows us to test patching compressed files, which are
detected using their file extension.
Use a hardcoded buffer size of 4096 to fix performance2015-12-26T13:49:00ZJulian Andres Klodejak@debian.org2015-12-26T13:49:00Zurn:sha1:0b29c72bdfc1466d47567cc3191b9661f81d3d3f
The code uses memmove() to move parts of the buffer to the
front when the buffer is only partially read. By simply
reading one page at a time, the maximum size of bytes that
must be moved has a hard limit, and performance improves:
In one test case, consisting of a 430 MB Contents file,
and a 75K PDiff, applying the PDiff previously took about
48 seconds and now completes in 2 seconds.
Further speed up can be achieved by buffering writes, they
account for about 60% of the run-time now.
Mark all FileFdPrivate classes as hidden2015-12-24T16:14:15ZJulian Andres Klodejak@debian.org2015-12-24T16:14:15Zurn:sha1:65ac6aad5f707849a127202f808d087d1fcaddc9
Gbp-Dch: ignore
1.1.6 Christmas release2015-12-24T16:03:21ZJulian Andres Klodejak@debian.org2015-12-24T16:03:21Zurn:sha1:3ac050d1ccfa33dbed236a565511fd3fe84b8465fix new[] vs delete mismatch introduced by b3db9d812015-12-23T18:42:53ZDavid Kalnischkiesdavid@kalnischkies.de2015-12-23T18:42:53Zurn:sha1:117551478551958d62d9555c595f2990805137c0
And as we are at it lets fix the 'style' issue I introduced with the
filefd changes as well.
Reported-By: gcc -fsanitize's & cppcheck
Git-Dch: Ignore
ensure we got a lock in clean operation2015-12-23T16:48:16ZDavid Kalnischkiesdavid@kalnischkies.de2015-12-23T16:48:16Zurn:sha1:8d968cc2d1e11da81f3bab878c1a413a97ca27ed
We try to acquired the locks, but we didn't stop if we failed to get it…
Closes: 808561
use a dynamic buffer for ReadLine2015-12-23T16:16:13ZDavid Kalnischkiesdavid@kalnischkies.de2015-12-23T16:16:13Zurn:sha1:b3db9d81325769fb9d9c6518f9d917f90328bef1
We don't need the buffer that often - only for ReadLine - as it is only
occasionally used, so it is actually more efficient to allocate it if
needed instead of statically by default. It also allows the caller to
influence the buffer size instead of hardcoding it.
Git-Dch: Ignore
implement a buffer system for FileFd::ReadLine2015-12-23T15:42:12ZDavid Kalnischkiesdavid@kalnischkies.de2015-12-23T15:42:12Zurn:sha1:f63123c385738b059cd854b694a0f62558fb8cbe
The default implementation of ReadLine was very naive by just reading
each character one-by-one. That is kinda okay for libraries implementing
compression as they have internal buffers (but still not great), but
while working with files directly or via a pipe as there is no buffer
there so all those reads are in fact system calls.
This commit introduces an internal buffer in the FileFd implementation
which is only used by ReadLine. The more low-level Read and all other
actions remain unbuffered – they just changed to deal with potential
"left-overs" in the buffer correctly.
Closes: 808579