Summary of changes in tzdata2016j (2016-11-22 23:17:13 -0800):
Saratov, Russia switches from +03 to +04 on 2016-12-04 at 02:00.
This creates a new zone Europe/Saratov.
New zone Asia/Atyrau for Atyrau Region, Kazakhstan, is like
Asia/Aqtau except it switched from +04/+05 to +05/+06 in spring
1999, not fall 1994.
Asia/Gaza and Asia/Hebron now use "EEST", not "EET", to denote
summer time before 1948. The old use of "EET" was a typo.
(one which used a different key for the signature of the data file...)
Allow either key to work. Also update the name of the sets list
file to match modern reality (only affects instructions issued to user.)
I skipped committing these changes until it had been used a few times
to verify that it actually works properly... it seems to.
Summary of changes in tzdata2016i (2016-11-01 23:19:52 -0700):
Cyprus split into two time zones on 2016-10-30 (new zone is
Asia/Famagusta and is UTC+3 year round). Tonga reintroduces
summer time on 2016-11-06 (assumed for now to be aligned with Fiji).
This year's summer time switch (from +08 to +11) for Antarctica/Casey
occurred 2016-10-22.
Also (minor) adjustments to some historic data for Italy (most
recent applies to time of day of switch out of summer time in period
1967-1970 & 1972-1974, other changes relate to 1910's and 1940's.)
Summary of changes in tzdata2016h (2016-10-19 23:17:57 -0700):
Asia/Gaza and Asia/Hebron end of summer time for 2016 is
2016-10-29 at 01:00 rather than 2016-10-21 at 00:00 (which
is within hours after this update). Guess that future
end dates will be the last Sat of October.
Corrected some historic time (and one date) transition
times for Turkey (all 1990 or before.)
Switch Sri-Lanka to use numeric abbreviation for time zone
as the tzdata invented one is not correct, and the correct
one seems not to be widely used.
Summary of changes in tzdata2016g (2016-09-13 08:56:38 -0700):
Timezone switch in Turkey (summer time becomes standard time)
Transition time corrections for historic timestamps in
America/Los_Angeles
zones using USSR rules in early 20th century
Some (more) time zone abbreviations converted to numeric form
Asia/Rangoon becomes Asia/Yangon (with backward compat link)
Summary of changes in tzdata2016g (2016-09-13 08:56:38 -0700):
Timezone switch in Turkey (summer time becomes standard time)
Transition time corrections for historic timestamps in
America/Los_Angeles
zones using USSR rules in early 20th century
Some (more) time zone abbreviations converted to numeric form
Asia/Rangoon becomes Asia/Yangon (with backward compat link)
Summary of changes in tzdata2016f (2016-07-05 16:26:51 +0200):
* The Egyptian government changed its mind on short notice, and
Africa/Cairo will not introduce DST starting 2016-07-07 after all.
* Asia/Novosibirsk switches from +06 to +07 on 2016-07-24 at 02:00.
* Asia/Novokuznetsk and Asia/Novosibirsk now use numeric time zone
abbreviations instead of invented ones.
* Europe/Minsk's 1992-03-29 spring-forward transition was at 02:00 not 00:00.
Summary of changes in tzdata2016e (2016-06-14 08:46:16 -0700):
* Africa/Cairo observes DST in 2016 from July 7 to the end of October.
(change obsoleted by 2016f)
* Locations while uninhabited now use '-00', not 'zzz', as a
placeholder time zone abbreviation.
* Asia/Baku's 1992-09-27 transition from +04 (DST) to +04 (non-DST) was
at 03:00, not 23:00 the previous day.
* zic now outputs a dummy transition at time 2**31 - 1 in zones
whose POSIX-style TZ strings contain a '<'.
* Changes affecting documentation and commentary.
Following changes have been made since the last import (taken from https://www.sqlite.org/news.html):
2016-05-18 - Release 3.13.0
SQLite version 3.13.0 is a regularly schedule maintenance release containg
performance enhancements and fixes for obscure bugs.
2016-04-18 - Release 3.12.2
Yikes! The 3.12.0 and 3.12.1 releases contain a backwards compatibility bug!
Tables that declare a column with type "INTEGER" PRIMARY KEY (where the
datatype name INTEGER is quoted) generate an incompatible database file. The
mistake came about because the developers have never thought to put a typename
in quotes before, and so there was no documentation of that capability nor any
tests. (There are tests now, though, of course.) Instances of quoting the
datatype name are probably infrequent in the wild, so we do not expect the
impact of this bug to be too severe. Upgrading is still strongly recommended.
Fixes for three other minor issues were included in this patch release. The
other issues would have normally been deferred until the next scheduled
release, but since a patch release is being issued anyhow, they might as well
be included.
2016-04-08 - Release 3.12.1
SQLite version 3.12.1 is an emergency patch release to address a crash bug that
snuck into version 3.12.0. Upgrading from version 3.12.0 is highly recommended.
Another minor problem involving datatypes on view columns, and a query planner
deficiency are fixed at the same time. These two issues did not justify a new
release on their own, but since a release is being issued to deal with the
crash bug, we included these other fixes for good measure.
2016-03-29 - Release 3.12.0
SQLite version 3.12.0 is a regularly scheduled maintenance release. A notable
change in this release is an increase in the default page size for newly
created database files. There are also various performance improvements. See
the change log for details.
2016-03-03 - Release 3.11.1
SQLite version 3.11.1 is a patch release that fixes problems in the new FTS5
extension and increases a default setting in the spellfix1 extension, and
implements enhancements to some of the Windows makefiles. The SQLite core is
unchanged from 3.11.0. Upgrading is optional.
2016-02-15 - Release 3.11.0
SQLite version 3.11.0 is a regularly scheduled maintenance release.
2016-01-20 - Release 3.10.2
Yikes! An optimization attempt gone bad resulted in a bug in the LIKE operator
which is fixed by this patch release. Three other minor but low-risk fixes are
also included in the patch.
2016-01-14 - Release 3.10.1
SQLite version 3.10.1 is a bug-fix release primarily targeting the fix for the
query planner bug cb3aa0641d9a4 discovered by Mapscape. Also included is a
minor API enhancement requested by the Firefox developers at Mozilla. The
differences from version 3.10.0 are minimal.
2016-01-06 - Release 3.10.0
SQLite version 3.10.0 is a regularly scheduled maintenance release.
2015-11-02 - Release 3.9.2
SQLite version 3.9.2 is a patch release fixing two obscure bugs. (Details: (1),
(2)). Upgrade only if you are having problems.
2015-10-16 - Release 3.9.1
SQlite version 3.9.1 is a small patch to version 3.9.0 that includes a few
simple build script and #ifdef tweaks to make the code easier to compile on a
wider variety of platform. There are no functional changes, except for a single
minor bug-fix in the json1 extension to stop it from recognizing form-feed
(ASCII 0x0c) as a whitespace character, in conformance with RFC7159.
2015-10-14 - Release 3.9.0
SQLite version 3.9.0 is a regularly schedule maintenance release. Key changes
include:
Begin using semantic versioning.
JSON SQL functions
The FTS5 full-text search engine
Support for indexes on expressions
Support for table-valued functions
See the change log for a long and more complete list of changes.
2015-07-29 - Release 3.8.11.1
SQLite version 3.8.11.1 is a patch release that fixes two arcane issues that
were reported shortly after 3.8.11 was released. Upgrade from 3.8.11 only in
the unlikely event that one of these obscure issues affect your code.
2015-07-27 - Release 3.8.11
SQLite version 3.8.11 is a regularly scheduled maintenance release. See the
change log for details.
2015-05-20 - Release 3.8.10.2
Yikes! Index corruption after a sequence of valid SQL statements!
It has been many years since anything like this bug has snuck into an official
SQLite release. But for the pasts seven months (version 3.8.7 through version
3.8.10.1) if you do an INSERT into a carefully crafted schema in which there
are two nested triggers that convert an index key value from TEXT to INTEGER
and then back to TEXT again, the INTEGER value might get inserted as the index
key instead of the correct TEXT, resulting in index corruption. This patch
release adds a single line of code to fix the problem.
If you do actually encounter this problem, running REINDEX on the damaged
indexes will clear it.
2015-05-09 - Release 3.8.10.1
The 3.8.10 release did not add the new SQLITE_ENABLE_DBSTAT_VTAB compile-time
option to the sqlite3_compileoption_used() interface. This patch release fixes
that omission. And while we are at it, the associated dbstat virtual table was
enhanced slightly and a harmless compiler warning was fixed.
There is no reason to upgrade from version 3.8.10 unless you are using the new
SQLITE_ENABLE_DBSTAT_VTAB compile-time option.
2015-05-07 - Release 3.8.10
SQLite version 3.8.10 is a regularly scheduled maintenance release. This
release features performance improvements, fixes to several arcane bugs found
by the AFL fuzzer, the new "sqldiff.exe" command-line utility, improvements to
the documentation, and other enhancements. See the release notes for additional
information.
2015-04-08 - Release 3.8.9
SQLite version 3.8.9 is a regularly scheduled maintenance release. New features
in this release include the PRAGMA index_xinfo command, the sqlite3_status64()
interface, and the ".dbinfo" command of the command-line shell. See the release
notes for additional information.
2015-02-25 - Release 3.8.8.3
The 3.8.8.3 patch release fixes an obscure problem in the SQLite code generator
that can cause incorrect results when the qualifying expression of a partial
index is used inside the ON clause of a LEFT JOIN. This problem has been in the
code since support for partial indexes was first added in version 3.8.0.
However, it is difficult to imagine a valid reason to every put the qualifying
constraint inside the ON clause of a LEFT JOIN, and so this issue has never
come up before.
Any applications that is vulnerable to this bug would have encountered problems
already. Hence, upgrading from the previous release is optional.
2015-01-30 - Release 3.8.8.2
The 3.8.8.2 patch release fixes a single minor problem: It ensures that the
sqlite3_wal_checkpoint(TRUNCATE) operation will always truncate the write-ahead
log even if log had already been reset and contained no new content. It is
unclear if this is a bug fix or a new feature.
Something like this would normally go into the next regularly scheduled
release, but a prominent SQLite user needed the change in a hurry so we were
happy to rush it out via this patch.
There is no reason to upgrade unless you actually need the enhanced behavior of
sqlite3_wal_checkpoint(TRUNCATE).
2015-01-20 - Release 3.8.8.1
Within hours of releasing version 3.8.8, a bug was reported against the
10-month-old 3.8.4 release. As that bug exists in all subsequent releases, the
decision was made to issue a small patch to the 3.8.8 before it came into
widespread use.
See ticket f97c4637102a3ae72b7911 for a description of the bug.
The changes between versions 3.8.8 and 3.8.8.1 are minimal.
2015-01-16 - Release 3.8.8
SQLite version 3.8.8 is a regularly schedule maintenance release of SQLite.
There are no dramatic new features or performance enhancements in this release,
merely incremental improvements. Most of the performance gain in this release
comes from refactoring the B-Tree rebalancing logic to avoid unnecessary
memcpy() operations. New features include the PRAGMA data_version statement and
the ability to accept a VALUES clause with no arbitrary limit on the number of
rows. Several obscure bugs have been fixed, including some multithreading races
and a work-around for a compiler bug on some Macs.
See the change log for a longer list of enhancements and bug fixes.
2014-12-09 - Release 3.8.7.4
SQLite version 3.8.7.4 an unscheduled bug-fix release. Changes from the
previous release and from version 3.8.7 are minimal.
This release fixes adds in a mutex that is required by the changes of the
3.8.7.3 patch but was accidentally omitted. The mutex was not required by any
of the internal SQLite tests, but Firefox crashes without it. Test cases have
been added to ensure that mutex is never again missed.
2014-12-06 - Release 3.8.7.3
SQLite version 3.8.7.3 an unscheduled bug-fix release. Changes from the
previous release and from version 3.8.7 are minimal.
This release fixes two obscure bugs that can result in incorrect query results
and/or application crashes, but not (as far as we can tell) security
vulnerabilities. Both bugs have been latent in the code across multiple prior
releases and have never before been encountered, so they are unlikely to cause
problems. Nevertheless it seems prudent to publish fixes for them both. See the
change log for details.
2014-11-19 - Release 3.8.7.2
SQLite version 3.8.7.2 is a patch and bug-fix release. Changes from the
previous release are minimal.
The primary reason for this release is to enhance the ROLLBACK command so that
it allows running queries on the same database connection to continue running
as long as the ROLLBACK does not change the schema. In all previous versions of
SQLite, a ROLLBACK would cause pending queries to stop immediately and return
SQLITE_ABORT or SQLITE_ABORT_ROLLBACK. Pending queries still abort if the
ROLLBACK changes the database schema, but as of this patch release, the queries
are allowed to continue running if the schema is unmodified.
In addition to the ROLLBACK enhancement, this patch release also includes fixes
for three obscure bugs. See the change log for details.
2014-10-30 - Release 3.8.7.1
SQLite version 3.8.7.1 is a bug-fix release.
The primary reason for this bug-fix release is to address a problem with
updating the value of fields at the end of a table that were added using ALTER
TABLE ADD COLUMN. This problem 1 first appeared in the 3.8.7 release.
Another minor annoyance in the 3.8.7 release was the fact that the Android
build tried to use the strchrnul() function from the standard C library but
that function is not available on Android. Android builds had to add
-DHAVE_STRCHRNUL=0 to work around the problem. This patch fixes that so that
Android builds should now work without any changes.
The operation of PRAGMA journal_mode=TRUNCATE has been enhanced so that it
invokes fsync() after truncating the journal file when PRAGMA synchronous=FULL.
This helps to preserve transaction durability in the case of a power loss
occurring shortly after commit.
Finally, a couple of long-standing and obscure problems associated with run
UPDATE and DELETE on VIEWs were fixed.
The changes from 3.8.7 are minimal.
2014-10-17 - Release 3.8.7
SQLite version 3.8.7 is a regularly scheduled maintenance release. Upgrading
from all prior versions is recommended.
Most of the changes from the previous release have been micro-optimizations
designed to help SQLite run a little faster. Each individual optimization has
an unmeasurably small performance impact. But the improvements add up. Measured
on a well-defined workload (which the SQLite developers use as a proxy for a
typical application workload) using cachegrind on Linux and compiled with gcc
4.8.1 and -Os on x64 linux, the current release does over 20% more work for the
same number of CPU cycles compared to the previous release. Cachegrind is not a
real CPU, and the workload used for measurement is only a proxy. So your
performance may vary. We expect to see about half the measured and reported
improvement in real-world applications. 10% is less than 20% but it is still
pretty good, we think.
This release includes a new set of C-language interfaces that have unsigned
64-bit instead of signed 32-bit length parameters. The new APIs do not provide
any new capabilities. But they do make it easier to write applications that are
more resistant to integer overflow vulnerabilities.
This release also includes a new sorter that is able to use multiple threads to
help with large sort operations. (Sort operations are sometimes required to
implement ORDER BY and/or GROUP BY clauses and are almost always required for
CREATE INDEX.) The multi-threads sorter is turned off by default and must be
enabled using the PRAGMA threads SQL command. Note that the multi-threaded
sorter provides faster real-time performance for large sorts, but it also uses
more CPU cycles and more energy.
2014-08-15 - Release 3.8.6
SQLite version 3.8.6 is a regularly scheduled maintenance release. Upgrading
from all previous versions is recommended.
This release contains the usual assortment of obscure bug fixes. One bug,
however, deserves special attention. A problem appeared in the CREATE INDEX
command beginning with version 3.8.2 (2013-12-06) that allowed, under some
circumstances, a UNIQUE index to be created on a column that was not unique.
Once the index was created, no new non-unique entries could be inserted, but
preexisting non-unique entries would remain. See ticket 9a6daf340df99ba93c for
further information. In addition to fixing this bug, the PRAGMA integrity_check
command has been enhanced to detect non-uniqueness in UNIQUE indices, so that
if this bug did introduce any problems in databases, those problems can be
easily detected.
Other noteworthy changes include the addition of support for hexadecimal
integers (ex: 0x1234), and performance enhancements to the IN operator which,
according to mailing list reports, help some queries run up to five times
faster.
Version 3.8.6 uses 25% fewer CPU cycles than version 3.8.0 from approximately
one year ago, according to valgrind and the test/speedtest1.c test program. On
the other hand, the compiled binary for version 3.8.6 is about 5% larger than
3.8.0. The size increase is due in part to the addition of new features such as
WITHOUT ROWID tables and common table expressions.
2014-06-04 - Release 3.8.5
SQLite version 3.8.5 is a regularly scheduled maintenance release. Upgrading
from the previous version is recommended.
Version 3.8.5 fixes more than a dozen obscure bugs. None of these bugs should
be a problem for existing applications. Nor do any of the bugs represent a
security vulnerability. Nevertheless, upgrading is recommended to prevent
future problems.
In addition to bug fixes, the 3.8.5 release adds improvements to the query
planner, especially regarding sorting using indices and handling OR terms in
the WHERE clause for WITHOUT ROWID tables. The ".system" and ".once"
dot-commands were added to the command-line interface. And there were
enhancements to the FTS4 and RTREE virtual tables. See the change log for
details.
2014-04-03 - Release 3.8.4.3
The optimizations added in version 3.8.4 caused some queries that involve
subqueries in the FROM clause, DISTINCT, and ORDER BY clauses, to give an
incorrect result. See ticket 98825a79ce145 for details. This release adds a
one-character change to a single line of code to fix the problem.
2014-03-26 - Release 3.8.4.2
The code changes that resulted in the performance improvements in version 3.8.4
missed a single buffer overflow test, which could result in a read past the end
of a buffer while searching a database that is corrupted in a particular way.
Version 3.8.4.2 fixes that problem using a one-line patch.
We are not aware of any problems in version 3.8.4 when working with well-formed
database files. The problem fixed by this release only comes up when reading
corrupt database files.
2014-03-11 - Release 3.8.4.1
SQLite version 3.8.4.1 is a patch against version 3.8.4 that fixes two minor
issues:
Work around a C-preprocessor macro conflict that causes compilation problems
for some configurations of Visual Studio.
Adjust the cost computation for the skip-scan optimization for improved
performance.
Both of these issues came to light within minutes of tagging the previous
release. Neither issue is serious but they can be annoying. Hence, the decision
was made to do a quick patch release to address both issues.
2014-03-10 - Release 3.8.4
SQLite version 3.8.4 is a maintenance release featuring performance
enhancements and fixes for a number of obscure bugs. There are no significant
new features in SQLite version 3.8.4. However, the number of CPU cycles
(measured by valgrind) needed to do many common operations has be reduced by
about 12% relative to the previous release, and by about 25% relative to
version 3.7.16 from approximately one year ago.
Version 3.8.4 of SQLite fixes several corner-case bugs that were found since
the previous release. These bugs were unlikely to appear in practice, and none
represent a security vulnerability. Nevertheless, developers are encouraged to
upgrade from all prior releases.
Summary of changes in tzdata2016d (2016-04-17 22:50:29 -0700):
* America/Caracas switches from -0430 to -04 on 2016-05-01 at 02:30.
* Asia/Magadan switches from +10 to +11 on 2016-04-24 at 02:00.
* New zone Asia/Tomsk, split off from Asia/Novosibirsk.
* New zone Europe/Kirov, split off from Europe/Volgograd.
* Russia and nearby locations had daylight-saving transitions on
1992-03-29 at 02:00 and 1992-09-27 at 03:00, instead of on
1992-03-28 at 23:00 and 1992-09-26 at 23:00.
* Many corrections to historical time in Kazakhstan from 1991
through 2005.
* Mention Internet RFCs 7808 (TZDIST) and 7809 (CalDAV time zone references).
Summary of changes in tzdata2016c (2016-03-23 00:51:27 -0700):
* Azerbaijan no longer observes DST.
* Chile reverts from permanent to seasonal DST.
* Europe/Kaliningrad and Europe/Vilnius changed from +03/+04 to
+02/+03 on 1989-03-26, not 1991-03-31.
* Several updates and URLs for historical and proposed Russian changes.
Summary of changes in tzdata2016b (2016-03-12 17:30:14 -0800):
* New zones Europe/Astrakhan and Europe/Ulyanovsk for Astrakhan and
Ulyanovsk Oblasts, Russia, both of which will switch from +03 to +04 on
2016-03-27 at 02:00 local time.
* As a trial of a new system that needs less information to be made up,
the new zones use numeric time zone abbreviations like "+04"
instead of invented abbreviations like "ASTT".
* Haiti will not observe DST in 2016.
* Palestine's spring-forward transition on 2016-03-26 is at 01:00, not 00:00.
* Europe/Chisinau observed DST during 1990, and switched from +04 to
+03 at 1990-05-06 02:00, instead of switching from +03 to +02.
* 1991 abbreviations in Europe/Samara should be SAMT/SAMST, not
KUYT/KUYST.
* tzselect's diagnostics and checking, and checktab.awk's checking,
have been improved.
* tzcode now builds under MinGW.
* tzselect now tests Julian-date TZ settings more accurately.
* Comments in zone tables have been improved.
* tzselect again limits its menu comments so that menus fit on a
24x80 alphanumeric display.
* A new web page tz-how-to.html.
* In the Theory file, the description of possible time zone abbreviations in
tzdata has been cleaned up, as the old description was unclear and
inconsistent.
Summary of changes in tzdata2016a (2016-01-26 23:28:02 -0800):
* America/Cayman will not observe daylight saving this year after all.
* Asia/Chita switches from +0800 to +0900 on 2016-03-27 at 02:00.
* Asia/Tehran now has DST predictions for the year 2038 and later,
to be March 21 00:00 to September 21 00:00.
* America/Metlakatla switched from PST all year to AKST/AKDT on
2015-11-01 at 02:00.
* America/Santa_Isabel has been removed, and replaced with a
backward compatibility link to America/Tijuana.
* Asia/Karachi's two transition times in 2002 were off by a minute.
* Changes affecting documentation and commentary.
Summary of changes in tzdata2015g (2015-10-01 00:39:51 -0700):
* Turkey's 2015 fall-back transition is scheduled for Nov.
* Norfolk moves from +1130 to +1100 on 2015-10-04 at 02:00 local time.
* Fiji's 2016 fall-back transition is scheduled for January 17, not 24.
* Fort Nelson, British Columbia will not fall back on 2015-11-01.
* Norfolk observed DST from 1974-10-27 02:00 to 1975-03-02 02:00.
* Changes affecting documentation.
Summary of changes in tzdata2015f (2015-08-10 18:06:56 -0700):
* North Korea switches to +0830 on 2015-08-15.
* Uruguay no longer observes DST.
* Moldova starts and ends DST at 00:00 UTC, not at 01:00 UTC.
* The two characters '%z' in a zone format now stand for the UTC
offset, e.g., '-07' for seven hours behind UTC and '+0530' for
five hours and thirty minutes ahead.
* Comments for America/Halifax and America/Glace_Bay have been improved.
* Data entries have been simplified for Atlantic/Canary, Europe/Simferopol,
Europe/Sofia, and Europe/Tallinn.
* Changes affecting documentation.
Summary of changes in tzdata2015e (2015-06-13 10:56:02 -0700):
* Morocco will suspend DST from 2015-06-14 03:00 through 2015-07-19 02:00,
not 06-13 and 07-18 as we had guessed.
* Assume Cayman Islands will observe DST starting next year, using US rules.
* The file 'iso3166.tab' now uses UTF-8, so that its entries can better
spell the names of Åland Islands, Côte d'Ivoire, and Réunion.
Summary of changes in tzdata2015d (2015-04-24 08:09:46 -0700):
* Egypt will not observe DST in 2015 and will consider canceling it
permanently. For now, assume no DST indefinitely.
* America/Whitehorse switched from UTC-9 to UTC-8 on 1967-05-28, not
1966-07-01.
* The abbreviations for Hawaii-Aleutian standard and daylight times
have been changed from HAST/HADT to HST/HDT, as per US Government
Printing Office style.
Summary of changes in tzdata2015c (2015-04-11 08:55:55 -0700):
* Egypt's spring-forward transition is at 24:00 on April's last Thursday,
not 00:00 on April's last Friday. This affects 2015, 2026, 2037,
2043, etc.
* Changes to pre-1991 Chile-related time stamps in America/Santiago,
Antarctica/Palmer, and Pacific/Easter.
* The America/Montreal zone has been turned into a link, as it differed
from an existing zone only for pre-1970 time stamps.
Bump version and soname for 5.2.1.
Date: 2015-02-26 13:01:09 +0200
Update NEWS for 5.2.1.
Date: 2015-02-22 19:38:48 +0200
xz: Use pipe2() if available.
Date: 2015-02-21 23:40:26 +0200
liblzma: Fix a compression-ratio regression in LZMA1/2 in fast mode.
The bug was added in the commit
f48fce093b07aeda95c18850f5e086d9f2383380 and thus
affected 5.1.4beta and 5.2.0. Luckily the bug cannot
cause data corruption or other nasty things.
Date: 2015-02-21 23:00:19 +0200
xz: Fix the fcntl() usage when creating a pipe for the self-pipe trick.
Now it reads the old flags instead of blindly setting O_NONBLOCK.
The old code may have worked correctly, but this is better.
Date: 2015-02-10 15:29:34 +0200
Update THANKS.
Date: 2015-02-10 15:28:30 +0200
tuklib_cpucores: Use cpuset_getaffinity() on FreeBSD if available.
In FreeBSD, cpuset_getaffinity() is the preferred way to get
the number of available cores.
Thanks to Rui Paulo for the patch. I edited it slightly, but
hopefully I didn't break anything.
Date: 2015-02-09 22:08:37 +0200
xzdiff: Make the mktemp usage compatible with FreeBSD's mktemp.
Thanks to Rui Paulo for the fix.
Date: 2015-02-03 21:45:53 +0200
Add a few casts to tuklib_integer.h to silence possible warnings.
I heard that Visual Studio 2013 gave warnings without the casts.
Thanks to Gabi Davar.
Date: 2015-01-26 21:24:39 +0200
liblzma: Set LZMA_MEMCMPLEN_EXTRA depending on the compare method.
Date: 2015-01-26 20:40:16 +0200
Update THANKS.
Date: 2015-01-26 20:39:28 +0200
liblzma: Silence harmless Valgrind errors.
Thanks to Torsten Rupp for reporting this. I had
forgotten to run Valgrind before the 5.2.0 release.
Date: 2015-01-09 21:50:19 +0200
xz: Fix comments.
Date: 2015-01-09 21:35:06 +0200
Update THANKS.
Date: 2015-01-09 21:34:06 +0200
xz: Don't fail if stdout doesn't support O_NONBLOCK.
This is similar to the case with stdin.
Thanks to Brad Smith for the bug report and testing
on OpenBSD.
Date: 2015-01-07 19:18:20 +0200
xz: Fix a memory leak in DOS-specific code.
Date: 2015-01-07 19:08:06 +0200
xz: Don't fail if stdin doesn't support O_NONBLOCK.
It's a problem at least on OpenBSD which doesn't support
O_NONBLOCK on e.g. /dev/null. I'm not surprised if it's
a problem on other OSes too since this behavior is allowed
in POSIX-1.2008.
The code relying on this behavior was committed in June 2013
and included in 5.1.3alpha released on 2013-10-26. Clearly
the development releases only get limited testing.
Date: 2015-01-06 20:30:15 +0200
Tests: Don't hide unexpected error messages in test_files.sh.
Hiding them makes no sense since normally there's no error
when testing the "good" files. With "bad" files errors are
expected and then it makes sense to keep the messages hidden.
Date: 2014-12-30 11:17:16 +0200
Update Solaris notes in INSTALL.
Mention the possible "make check" failure on Solaris in the
Solaris-specific section of INSTALL. It was already in
section 4.5 but it is better mention it in the OS-specific
section too.
Date: 2014-12-26 12:00:05 +0200
Build: POSIX shell isn't required if scripts are disabled.
Date: 2014-12-21 20:48:37 +0200
DOS: Update Makefile.
Date: 2014-12-21 19:50:38 +0200
Windows: Fix bin_i486 to bin_i686 in build.bash.
Date: 2014-12-21 18:58:44 +0200
Docs: Use lzma_cputhreads() in 04_compress_easy_mt.c.
Date: 2014-12-21 18:56:44 +0200
Docs: Update docs/examples/00_README.txt.
Date: 2014-12-21 18:11:17 +0200
Bump version and soname for 5.2.0.
I know that soname != app version, but I skip AGE=1
in -version-info to make the soname match the liblzma
version anyway. It doesn't hurt anything as long as
it doesn't conflict with library versioning rules.
Date: 2014-12-21 18:05:03 +0200
Avoid variable-length arrays in the debug programs.
Date: 2014-12-21 18:01:45 +0200
Build: Include 04_compress_easy_mt.c in the tarball.
Date: 2014-12-21 18:00:38 +0200
Fix build when --disable-threads is used.
Date: 2014-12-21 15:56:15 +0100
po/fr: improve wording for help for --lzma1/--lzma2.
Date: 2014-12-21 15:55:48 +0100
po/fr: missing line in translation of --extreme.
Date: 2014-12-21 14:32:33 +0200
Update NEWS for 5.2.0.
Date: 2014-12-21 14:32:22 +0200
Update NEWS for 5.0.8.
Date: 2014-12-21 14:07:54 +0200
xz: Fix a comment.
Date: 2014-12-20 20:43:14 +0200
Update INSTALL about the dependencies of the scripts.
Date: 2014-12-20 20:42:33 +0200
Windows: Update build instructions.
Date: 2014-12-20 20:41:48 +0200
Windows: Update the build script and README-Windows.txt.
The 32-bit build is now for i686 or newer because the
prebuilt MinGW-w64 toolchains include i686 code in the
executables even if one uses -march=i486.
The build script builds 32-bit SSE2 enabled version too.
Run-time detection of SSE2 support would be nice (on any OS)
but it's not implemented in XZ Utils yet.
Date: 2014-12-19 15:51:50 +0200
Windows: Define TUKLIB_SYMBOL_PREFIX in config.h.
It is to keep all symbols in the lzma_ namespace.
Date: 2014-12-16 21:00:09 +0200
xz: Update the man page about --threads.
Date: 2014-12-16 20:57:43 +0200
xz: Update the man page about --block-size.
Date: 2014-12-10 22:26:57 +0100
po/fr: several more translation updates: reword and handle --ignore-check.
Date: 2014-12-10 22:23:01 +0100
po/fr: yet another place where my email address had to be updated.
Date: 2014-12-10 22:22:20 +0100
po/fr: fix several typos that have been around since the beginning.
Date: 2014-12-03 20:02:31 +0100
po/fr: last batch of new translations for now.
Four new error messages.
Date: 2014-12-03 20:01:32 +0100
po/fr: translations for --threads, --block-size and --block-list.
Date: 2014-12-03 20:00:53 +0100
po/fr: remove fuzzy marker for error messages that will be kept in English.
The following is a copy of a comment inside fr.po:
Note from translator on "file status flags".
The following entry is kept un-translated on purpose. It is difficult to
translate and should only happen in exceptional circumstances which means
that translating would:
- lose some of the meaning
- make it more difficult to look up in search engines; it might happen one
in
a million times, if we dilute the error message in 20 languages, it will be
almost impossible to find an explanation and support for the error.
Date: 2014-12-03 19:58:25 +0100
po/fr: several minor updates and better wording.
Meaning doesn't change at all: it's only for better wording and/or
formatting of a few strings.
Date: 2014-12-03 19:56:12 +0100
po/fr: update my email address and copyright years.
Date: 2014-11-26 10:08:26 +0100
fr.po: commit file after only "update-po" so actual is readable.
Date: 2014-12-02 20:04:07 +0200
liblzma: Document how lzma_mt.block_size affects memory usage.
Date: 2014-11-28 20:07:18 +0200
Update INSTALL about a "make check" failure in test_scripts.sh.
Date: 2014-11-26 20:12:27 +0200
Remove LZMA_UNSTABLE macro.
Date: 2014-11-26 20:10:33 +0200
liblzma: Update lzma_stream_encoder_mt() API docs.
Date: 2014-11-25 12:32:05 +0200
liblzma: Verify the filter chain in threaded encoder initialization.
This way an invalid filter chain is detected at the Stream
encoder initialization instead of delaying it to the first
call to lzma_code() which triggers the initialization of
the actual filter encoder(s).
Date: 2014-11-17 19:11:49 +0200
Build: Update m4/ax_pthread.m4 from Autoconf Archive.
Date: 2014-11-17 18:52:21 +0200
Build: Replace obsolete AC_HELP_STRING with AS_HELP_STRING.
Date: 2014-11-17 18:43:19 +0200
Build: Fix Autoconf warnings about escaped backquotes.
Thanks to Daniel Richard G. for pointing out that it's
good to sometimes run autoreconf -fi with -Wall.
Date: 2014-11-10 18:54:40 +0200
xzdiff: Use mkdir if mktemp isn't available.
Date: 2014-11-10 18:45:01 +0200
xzdiff: Create a temporary directory to hold a temporary file.
This avoids the possibility of "File name too long" when
creating a temp file when the input file name is very long.
This also means that other users on the system can no longer
see the input file names in /tmp (or whatever $TMPDIR is)
since the temporary directory will have a generic name. This
usually doesn't matter since on many systems one can see
the arguments given to all processes anyway.
The number X chars to mktemp where increased from 6 to 10.
Note that with some shells temp files or dirs won't be used at all.
Date: 2014-11-10 15:38:47 +0200
liblzma: Fix lzma_mt.preset in lzma_stream_encoder_mt_memusage().
It read the filter chain from a wrong variable. This is a similar
bug that was fixed in 9494fb6d0ff41c585326f00aa8f7fe58f8106a5e.
Date: 2014-11-10 14:49:55 +0200
Update THANKS.
Date: 2014-10-29 21:28:25 +0200
Update .gitignore files.
Date: 2014-10-29 21:15:35 +0200
Build: Prepare to support Automake's subdir-objects.
Due to a bug in Automake, subdir-objects won't be enabled
for now.
http://debbugs.gnu.org/cgi/bugreport.cgi?bug=17354
Thanks to Daniel Richard G. for the original patches.
Date: 2014-10-24 20:09:29 +0300
Translations: Update the Italian translation.
Thanks to Milo Casagrande.
Date: 2014-10-18 18:51:45 +0300
Translations: Update the Polish translation.
Thanks to Jakub Bogusz.
Date: 2014-10-14 17:30:30 +0200
l10n: de.po: Change translator email address.
Although the old address is still working, the new one should
be preferred. So this commit changes all three places in de.po
accordingly.
Signed-off-by: Andre Noll <maan@tuebingen.mpg.de>
Date: 2014-10-14 17:30:29 +0200
l10n: de.po: Update German translation
Signed-off-by: Andre Noll <maan@systemlinux.org>
Date: 2014-10-14 17:30:28 +0200
l10n: de.po: Fix typo: Schießen -> Schließen.
That's a funny one since "schießen" means to shoot :)
Signed-off-by: Andre Noll <maan@systemlinux.org>
Date: 2014-10-09 19:42:26 +0300
Update THANKS.
Date: 2014-10-09 19:41:51 +0300
Add support for AmigaOS/AROS to tuklib_physmem().
Thanks to Fredrik Wikstrom.
Date: 2014-10-09 18:42:14 +0300
xzgrep: Avoid passing both -q and -l to grep.
The behavior of grep -ql varies:
- GNU grep behaves like grep -q.
- OpenBSD grep behaves like grep -l.
POSIX doesn't make it 100 % clear what behavior is expected.
Anyway, using both -q and -l at the same time makes no sense
so both options simply should never be used at the same time.
Thanks to Christian Weisgerber.
Date: 2014-09-25 09:22:45 +0700
l10n: vi.po: Update Vietnamese translation
Signed-off-by: Trần Ngọc Quân <vnwildman@gmail.com>
Date: 2014-09-25 18:38:48 +0300
Build: Detect supported compiler warning flags better.
Clang and nowadays also GCC accept any -Wfoobar option
but then may give a warning that an unknown warning option
was specified. To avoid adding unsupported warning options,
the options are now tested with -Werror.
Thanks to Charles Diza.
Date: 2014-09-20 21:01:21 +0300
Update NEWS for 5.0.7.
Date: 2014-09-20 19:42:56 +0300
liblzma: Fix a portability problem in Makefile.am.
POSIX supports $< only in inference rules (suffix rules).
Using it elsewhere is a GNU make extension and doesn't
work e.g. with OpenBSD make.
Thanks to Christian Weisgerber for the patch.
Date: 2014-09-14 21:54:09 +0300
Bump the version number to 5.1.4beta.
Date: 2014-09-14 21:50:13 +0300
Update NEWS for 5.0.6 and 5.1.4beta.
Date: 2014-09-14 21:02:41 +0300
Update TODO.
Date: 2014-08-05 22:32:36 +0300
xz: Add --ignore-check.
Date: 2014-08-05 22:15:07 +0300
liblzma: Add support for LZMA_IGNORE_CHECK.
Date: 2014-08-05 22:03:30 +0300
liblzma: Add support for lzma_block.ignore_check.
Note that this slightly changes how lzma_block_header_decode()
has been documented. Earlier it said that the .version is set
to the lowest required value, but now it says that the .version
field is kept unchanged if possible. In practice this doesn't
affect any old code, because before this commit the only
possible .version was 0.
Date: 2014-08-04 19:25:58 +0300
liblzma: Use lzma_memcmplen() in the BT3 match finder.
I had missed this when writing the commit
5db75054e900fa06ef5ade5f2c21dffdd5d16141.
Thanks to Jun I Jin.
Date: 2014-08-04 00:25:44 +0300
Update THANKS.
Date: 2014-08-03 21:32:25 +0300
liblzma: SHA-256: Optimize the Maj macro slightly.
The Maj macro is used where multiple things are added
together, so making Maj a sum of two expressions allows
some extra freedom for the compiler to schedule the
instructions.
I learned this trick from
<http://www.hackersdelight.org/corres.txt>.
Date: 2014-08-03 21:08:12 +0300
liblzma: SHA-256: Optimize the way rotations are done.
This looks weird because the rotations become sequential,
but it helps quite a bit on both 32-bit and 64-bit x86:
- It requires fewer instructions on two-operand
instruction sets like x86.
- It requires one register less which matters especially
on 32-bit x86.
I hope this doesn't hurt other archs.
I didn't invent this idea myself, but I don't remember where
I saw it first.
Date: 2014-08-03 20:38:13 +0300
liblzma: SHA-256: Remove the GCC #pragma that became unneeded.
The unrolling in the previous commit should avoid the
situation where a compiler may think that an uninitialized
variable might be accessed.
Date: 2014-08-03 20:33:38 +0300
liblzma: SHA-256: Unroll a little more.
This way a branch isn't needed for each operation
to choose between blk0 and blk2, and still the code
doesn't grow as much as it would with full unrolling.
Date: 2014-08-03 19:56:43 +0300
liblzma: SHA-256: Do the byteswapping without a temporary buffer.
Date: 2014-07-25 22:38:28 +0300
liblzma: Use lzma_memcmplen() in normal mode of LZMA.
Two locations were not changed yet because the simplest change
assumes that the initial "len" may be greater than "limit".
Date: 2014-07-25 22:30:38 +0300
liblzma: Simplify LZMA fast mode code by using memcmp().
Date: 2014-07-25 22:29:49 +0300
liblzma: Use lzma_memcmplen() in fast mode of LZMA.
Date: 2014-07-25 21:16:23 +0300
Update THANKS.
Date: 2014-07-25 21:15:07 +0300
liblzma: Use lzma_memcmplen() in the match finders.
This doesn't change the match finder output.
Date: 2014-07-25 20:57:20 +0300
liblzma: Add lzma_memcmplen() for fast memory comparison.
This commit just adds the function. Its uses will be in
separate commits.
This hasn't been tested much yet and it's perhaps a bit early
to commit it but if there are bugs they should get found quite
quickly.
Thanks to Jun I Jin from Intel for help and for pointing out
that string comparison needs to be optimized in liblzma.
Date: 2014-07-12 21:10:09 +0300
Update THANKS.
Date: 2014-07-12 20:06:08 +0300
Translations: Add Vietnamese translation.
Thanks to Trần Ngọc Quân.
Date: 2014-06-29 20:54:14 +0300
xz: Update the help message of a few options.
Updated: --threads, --block-size, and --block-list
Added: --flush-timeout
Date: 2014-06-18 22:07:06 +0300
xz: Use lzma_cputhreads() instead of own copy of tuklib_cpucores().
Date: 2014-06-18 22:04:24 +0300
liblzma: Add lzma_cputhreads().
Date: 2014-06-18 19:11:52 +0300
xz: Check for filter chain compatibility for --flush-timeout.
This avoids LZMA_PROG_ERROR from lzma_code() with filter chains
that don't support LZMA_SYNC_FLUSH.
Date: 2014-06-13 19:21:54 +0300
xzgrep: List xzgrep_expected_output in tests/Makefile.am.
Date: 2014-06-13 18:58:22 +0300
xzgrep: Improve the test script.
Now it should be close to the functionality of the original
version by Pavel Raiskup.
Date: 2014-06-11 21:03:25 +0300
xzgrep: Add a test for the previous fix.
This is a simplified version of Pavel Raiskup's
original patch.
Date: 2014-06-11 20:43:28 +0300
xzgrep: exit 0 when at least one file matches.
Mimic the original grep behavior and return exit_success when
at least one xz compressed file matches given pattern.
Original bugreport:
https://bugzilla.redhat.com/show_bug.cgi?id=1108085
Thanks to Pavel Raiskup for the patch.
Date: 2014-06-09 21:21:24 +0300
xz: Force single-threaded mode when --flush-timeout is used.
Date: 2014-05-25 22:05:39 +0300
Update THANKS.
Date: 2014-05-25 21:45:56 +0300
liblzma: Use lzma_alloc_zero() in LZ encoder initialization.
This avoids a memzero() call for a newly-allocated memory,
which can be expensive when encoding small streams with
an over-sized dictionary.
To avoid using lzma_alloc_zero() for memory that doesn't
need to be zeroed, lzma_mf.son is now allocated separately,
which requires handling it separately in normalize() too.
Thanks to Vincenzo Innocente for reporting the problem.
Date: 2014-05-25 19:25:57 +0300
liblzma: Add the internal function lzma_alloc_zero().
Date: 2014-05-08 18:03:09 +0300
xz: Fix uint64_t vs. size_t which broke 32-bit build.
Thanks to Christian Hesse.
Date: 2014-05-04 11:09:11 +0300
Docs: Update comments to refer to lzma/lzma12.h in example programs.
Date: 2014-05-04 11:07:17 +0300
liblzma: Rename the private API header lzma/lzma.h to lzma/lzma12.h.
It can be confusing that two header files have the same name.
The public API file is still lzma.h.
Date: 2014-04-25 17:53:42 +0300
Build: Fix the combination of --disable-xzdec --enable-lzmadec.
In this case "make install" could fail if the man page directory
didn't already exist at the destination. If it did exist, a
dangling symlink was created there. Now the link is omitted
instead. This isn't the best fix but it's better than the old
behavior.
Date: 2014-04-25 17:44:26 +0300
Build: Add --disable-doc to configure.
Date: 2014-04-24 18:06:24 +0300
Update INSTALL.
Add a note about failing "make check". The source of
the problem should be fixed in libtool (if it really is
a libtool bug and not mine) but I'm unable to spend time
on that for now. Thanks to Nelson H. F. Beebe for reporting
the issue.
Add a note about a possible need to run "ldconfig" after
"make install".
Date: 2014-04-09 17:26:10 +0300
xz: Rename a variable to avoid a namespace collision on Solaris.
I don't know the details but I have an impression that there's
no problem in practice if using GCC since people have built xz
with GCC (without patching xz), but renaming the variable cannot
hurt either.
Thanks to Mark Ashley.
Date: 2014-01-29 20:19:41 +0200
Docs: Add example program for threaded encoding.
I didn't add -DLZMA_UNSTABLE to Makefile so one has to
specify it manually as long as LZMA_UNSTABLE is needed.
Date: 2014-01-29 20:13:51 +0200
liblzma: Fix lzma_mt.preset not working with lzma_stream_encoder_mt().
It read the filter chain from a wrong variable.
Date: 2014-01-20 11:20:40 +0200
liblzma: Fix typo in a comment.
Date: 2014-01-12 19:38:43 +0200
Windows: Add config.h for building liblzma with MSVC 2013.
This is for building liblzma. Building xz tool too requires
a little more work. Maybe it will be supported, but for most
MSVC users it's enough to be able to build liblzma.
C99 support in MSVC 2013 is almost usable which is a big
improvement over earlier versions. It's "almost" because
there's a dumb bug that breaks mixed declarations after
an "if" statements unless the "if" statement uses braces:
https://connect.microsoft.com/VisualStudio/feedback/details/808650/visual-studio-2013-c99-compiler-bughttps://connect.microsoft.com/VisualStudio/feedback/details/808472/c99-support-of-mixed-declarations-and-statements-fails-with-certain-types-and-constructs
Hopefully it will get fixed. Then liblzma should be
compilable with MSVC 2013 without patching.
Date: 2014-01-12 17:41:14 +0200
xz: Fix a comment.
Date: 2014-01-12 17:04:33 +0200
Windows: Add MSVC defines for inline and restrict keywords.
Date: 2014-01-12 16:44:52 +0200
liblzma: Avoid C99 compound literal arrays.
MSVC 2013 doesn't like them. Maybe they aren't so good
for readability either since many aren't used to them.
Date: 2014-01-12 12:50:30 +0200
liblzma: Remove a useless C99ism from sha256.c.
Unsurprisingly it makes no difference in compiled output.
Date: 2014-01-12 12:17:08 +0200
xz: Fix use of wrong variable.
Since the only call to suffix_set() uses optarg
as the argument, fixing this bug doesn't change
the behavior of the program.
Date: 2014-01-12 12:11:36 +0200
Fix typos in comments.
Date: 2013-11-26 18:20:16 +0200
Update THANKS.
Date: 2013-11-26 18:20:09 +0200
liblzma: Document the need for block->check for lzma_block_header_decode().
Thanks to Tomer Chachamu.
Date: 2013-11-12 16:38:57 +0200
xz: Update the man page about --block-size and --block-list.
Date: 2013-11-12 16:30:53 +0200
Update THANKS.
Date: 2013-11-12 16:29:48 +0200
xz: Make --block-list and --block-size work together in single-threaded.
Previously, --block-list and --block-size only worked together
in threaded mode. Boundaries are specified by --block-list, but
--block-size specifies the maximum size for a Block. Now this
works in single-threaded mode too.
Thanks to James M Leddy for the original patch.
Date: 2013-10-26 13:26:14 +0300
Bump the version number to 5.1.3alpha.
Date: 2013-10-26 13:25:02 +0300
Update NEWS for 5.1.3alpha.
Date: 2013-10-26 12:47:04 +0300
Update TODO.
Date: 2013-10-25 22:41:28 +0300
xz: Document behavior of --block-list with threads.
This needs to be updated before 5.2.0.
Date: 2013-10-22 20:03:12 +0300
xz: Document --flush-timeout=TIMEOUT on the man page.
Date: 2013-10-22 19:51:55 +0300
xz: Take advantage of LZMA_FULL_BARRIER with --block-list.
Now if --block-list is used in threaded mode, the encoder
won't need to flush at each Block boundary specified via
--block-list. This improves performance a lot, making
threading helpful with --block-list.
The flush timer was reset after LZMA_FULL_FLUSH but since
LZMA_FULL_BARRIER doesn't flush, resetting the timer is
no longer done.
Date: 2013-10-02 20:05:23 +0300
liblzma: Support LZMA_FULL_FLUSH and _BARRIER in threaded encoder.
Now --block-list=SIZES works with in the threaded mode too,
although the performance is still bad due to the use of
LZMA_FULL_FLUSH instead of the new LZMA_FULL_BARRIER.
Date: 2013-10-02 12:55:11 +0300
liblzma: Add LZMA_FULL_BARRIER support to single-threaded encoder.
In the single-threaded encoder LZMA_FULL_BARRIER is simply
an alias for LZMA_FULL_FLUSH.
Date: 2013-09-17 11:57:51 +0300
liblzma: Add block_buffer_encoder.h into Makefile.inc.
This should have been in b465da5988dd59ad98fda10c2e4ea13d0b9c73bc.
Date: 2013-09-17 11:55:38 +0300
xz: Add a missing test for TUKLIB_DOSLIKE.
Date: 2013-09-17 11:52:28 +0300
Add native threading support on Windows.
Now liblzma only uses "mythread" functions and types
which are defined in mythread.h matching the desired
threading method.
Before Windows Vista, there is no direct equivalent to
pthread condition variables. Since this package doesn't
use pthread_cond_broadcast(), pre-Vista threading can
still be kept quite simple. The pre-Vista code doesn't
use anything that wasn't already available in Windows 95,
so the binaries should run even on Windows 95 if someone
happens to care.
Date: 2013-09-11 14:40:35 +0300
Build: Remove a comment about Automake 1.10 from configure.ac.
The previous commit supports silent rules and that requires
Automake 1.11.
Date: 2013-09-09 20:37:03 +0300
Build: Create liblzma.pc in a src/liblzma/Makefile.am.
Previously it was done in configure, but doing that goes
against the Autoconf manual. Autoconf requires that it is
possible to override e.g. prefix after running configure
and that doesn't work correctly if liblzma.pc is created
by configure.
A potential downside of this change is that now e.g.
libdir in liblzma.pc is a standalone string instead of
being defined via ${prefix}, so if one overrides prefix
when running pkg-config the libdir won't get the new value.
I don't know if this matters in practice.
Thanks to Vincent Torri.
Date: 2013-08-04 15:24:09 +0300
Fix the previous commit which broke the build.
Apparently I didn't even compile-test the previous commit.
Thanks to Christian Hesse.
Date: 2013-08-03 13:52:58 +0300
Windows: Add Windows support to tuklib_cpucores().
It is used for Cygwin too. I'm not sure if that is
a good or bad idea.
Thanks to Vincent Torri.
Date: 2013-08-02 15:59:46 +0200
macosx: separate liblzma package
Date: 2013-08-02 15:58:44 +0200
macosx: set minimum to leopard
Date: 2011-08-07 13:13:30 +0200
move configurables into variables
Date: 2013-07-15 14:08:41 +0300
Update THANKS.
Date: 2013-07-15 14:08:02 +0300
Build: Fix the detection of missing CRC32.
Thanks to Vincent Torri.
Date: 2013-07-04 14:18:46 +0300
xz: Add preliminary support for --flush-timeout=TIMEOUT.
When --flush-timeout=TIMEOUT is used, xz will use
LZMA_SYNC_FLUSH if read() would block and at least
TIMEOUT milliseconds has elapsed since the previous flush.
This can be useful in realtime-like use cases where the
data is simultanously decompressed by another process
(possibly on a different computer). If new uncompressed
input data is produced slowly, without this option xz could
buffer the data for a long time until it would become
decompressible from the output.
If TIMEOUT is 0, the feature is disabled. This is the default.
This commit affects the compression side. Using xz for
the decompression side for the above purpose doesn't work
yet so well because there is quite a bit of input and
output buffering when decompressing.
The --long-help or man page were not updated yet.
The details of this feature may change.
Date: 2013-07-04 13:41:03 +0300
xz: Don't set src_eof=true after an I/O error because it's useless.
Date: 2013-07-04 13:25:11 +0300
xz: Fix the test when to read more input.
Testing for end of file was no longer correct after full flushing
became possible with --block-size=SIZE and --block-list=SIZES.
There was no bug in practice though because xz just made a few
unneeded zero-byte reads.
Date: 2013-07-04 12:51:57 +0300
xz: Move some of the timing code into mytime.[hc].
This switches units from microseconds to milliseconds.
New clock_gettime(CLOCK_MONOTONIC) will be used if available.
There is still a fallback to gettimeofday().
Date: 2013-07-01 14:35:03 +0300
Update THANKS.
Date: 2013-07-01 14:34:11 +0300
xz: Silence a warning seen with _FORTIFY_SOURCE=2.
Thanks to Christian Hesse.
* Edit version numbers.
* Make the script deal better with being re-run multiple times. It
now keeps a status file for each non-trivial step, and doesn't repeat
the step if the sttaus file exists.
* Instead of spawning a subshell to resolve merge conflicts, print a message
and exit, with a suggestion to rte-run the script after conflicts
are resolved.
* Check whether there are any conflicts before trying to do anything
about conflicts.
