v3c

lp:v3c

Created by Philip Ashmore and last modified
Get this branch:
bzr branch lp:v3c

Related bugs

Related blueprints

Branch information

Owner:
Philip Ashmore
Project:
v3c
Status:
Mature

Import details

Import Status: Failed

This branch is an import of the HEAD branch of the Git repository at git://git.code.sf.net/p/v3c/code.

The import has been suspended because it failed 5 or more times in succession.

Last successful import was .

Import started on neumayer and finished taking 40 seconds — see the log
Import started on pear and finished taking 20 seconds — see the log
Import started on russkaya and finished taking 20 seconds — see the log
Import started on pear and finished taking 15 seconds — see the log

Recent revisions

104. By Philip Ashmore

Version 2.5.7-01

* Added TreeStreambuf2
  This variant lets you specify a prefix and suffix for entries.

* Added trace-tree_streambuf-test
  Combining tracing and trees.

* Added m_ps->flush() to PrettyPrintStreambuf::close()
  Now it plays nice when "trace" is supplied as the stream.

* Added v3c-raspi to v3c/v3c-local.sh

103. By Philip Ashmore

Version 2.5.6-02

* Changed the default Ubuntu release to "Raring"
  I did a clean Ubuntu install just to be sure.
  I've been doing "apt-get update" and never checked if there was a
  newer release - sorry Quantal and Raring!

  As usual you can change the release being targetted by setting
  v3c_ubuntu_release on the make command line or in the environment.

102. By Philip Ashmore

Version 2.5.6-01

* Added ModelViewController implementation
  Ok, so there's no Controller but there are models and views.
  A view can "watch" multiple models.
  Models and views can be deleted in any order - there's a test.
  The only event type already defined is eEvDestroyed - model destroyed.
  You have to pass unhandled model events to the base View class.

* Fixed v3c-local.sh, which fixes make v3c-local

* Fixed long-standing PrettyPrintStreambuf bug
  The sync() implementation made some wrong assumptions.

  The PrettyPrintStreambuf and TreeStreambuf classes can handle cerr
  correctly now as well.

* Added isOpen() to PrettyPrintStreambuf.

* Added dir_part() extension_part() name_part() mkstemps() and
  resolve_path() to v3c/filesystem.hpp

* Added cwd() in v3c/filesystem.hpp
  It's the same as pwd().

* endsWith(), replaceAll() and join() added to v3c/string_utils.hpp

101. By Philip Ashmore

Version 2.5.5-01

* Added the "v3c-local" target to the top-level makefile
  See README for details.

* Added V3C_VER and v3c_version() to v3c.h
  V3C_VER contains
    10000 * $package_major_version + 100 * $package_minor_version +
    $package_micro_version
  , which is 20505 for this version.

  v3c_version() allows programs that dlopen libv3c to check its version.

* Added mkstemp to v3c/filesystem.hpp
  This takes care of where to create the file if you don't tell it
  where, and c++-wraps the ::mkstemp call.

  This is a small feature addition, hence the micro version bump.

* Added "SHELL := $(SHELL) -e" to v3c.mak, if "-e" not already there
  This forces all make commands to fail if there's any error.

* Modified the "v3c" program and the v3c::i_process_streambuf class to
  handle a SHELL environment variable that includes parameters.

* Added "v3c nth" to print the nth argument, not counting the index arg

* smart_ptr_cow_traits -> smart_cow_ptr_traits in smart_ptr.hpp

100. By Philip Ashmore

Version 2.5.4-03

* Fixed a smart pointer bug
  Flex likes to do "yylval = 0", and if yylval's type is a smart
  pointer, you got a leak.

99. By Philip Ashmore

Version 2.5.4-02

* Fixed git checkouts
  Either I hadn't tested it in a while or some new problem occurred,
  but I changes how git checkouts are done for "make git tag=X" and
  "make git branch=X" so they no longer affect any git repo objects,
  although git does mention that the newly created repo is in a
  'detached HEAD' state.

98. By Philip Ashmore

Version 2.5.4-01

* Lintian clean (N: 2 tags overridden (2 warnings))
  See debian/treedb.lintian-overrides.in for details.

* Removed -release \${PACKAGE_LT_RELEASE} from PACKAGE_LT_LDFLAGS
  This is done in v3c.m4.in and v3c-make-helper.

  It results in libtool creating shared libraries without the package
  version in their name, like other packages.

* Added a man page for v3c-make-helper

* Changed "sign_debian" to "v3c_sign_debian"
  In v3c/v3c.mak and v3c/v3c-make-helper.

* Split debian/changelog.in for Debian/Ubuntu
  There are now two files - debian/changelog-debian.in and
  debian/changelog-ubuntu.in .
  Which one gets chosen depends on whether you "make debian" or
  "make ubuntu".

  This means that after you
  1. copy v3c's debian/changelog-ubuntu.in file to your package
  2. add it to your (git) version control system
  3. git-mv debian/changelog.in debian/changelog-debian.in
  4. Update configure.ac.in (for both)
  5. Update Makefile.am's EXTRA_DIST to include both
  "make release ubuntu" and you're good to go!

* Added V3C_UBUNTU_RELEASE to v3c.m4.in and v3c-make-helper
  This allows you to

     make v3c_ubuntu_release=quantal release ubuntu

  to target a specific Ubuntu release - the default for now is "precise".

  You can also export it in the shell environment when building multiple
  Ubuntu packages.

  It does this by defining V3C_UBUNTU_RELEASE with AC_SUBST, so any *.in
  file using @V3C_UBUNTU_RELEASE@ gets the value, like
  debian/changelog-ubuntu.in.

  This is a significant enough change to warrant a large version bump.

97. By Philip Ashmore

Version 2.5.3-01

* Added Debian/Ubuntu build requirements into configure.ac.in

* Debian/Ubuntu builds can now be run with make -jN
  So "make -j7 git branch=2.5.2-02 release ubuntu" is a lot faster,
  if you have the cores.
  The version in git needs to be parallel-friendly too for this to work.

* Added source fortification
  This adds
  CPPFLAGS: -D_FORTIFY_SOURCE=2
  CXXFLAGS: -fstack-protector --param=ssp-buffer-size=4 -Wformat -Werror=format-security
  CPPFLAGS: -fstack-protector --param=ssp-buffer-size=4 -Wformat -Werror=format-security

* Added v3c_strcpy() to v3c/utils.h to workaround _FORTIFY_SOURCE
  This is needed for trailing unsized ("[]") character arrays and
  _FORTIFY_SOURCE - the "regular" strcpy detects that the destination
  length is zero and aborts the program, even if the space has been
  reserved in the destination.

  Im my opinion the compiler should issue a warning/error and not leave
  this until runtime.

  This is a significant enough change to warrant a large version bump.

* Fixed PACKAGE_LT_LDFLAGS definition in v3c.m4.in and v3c-make-helper
  The "make" program handles $ABC as ${A}BC so this would silently fail.

* Removed "-@PACKAGE_API_VERSION@" for lib_LTLIBRARIES
  As a result of fixing PACKAGE_LT_LDFLAGS, libtool now gets the
  -release flag which it uses to add "-@PACKAGE_API_VERSION@" before the
  ".so" with shared libraries which it should have done all along.

  Because of this, shared libraries were getting a double-dose of
  version strings - silly.

  So now, the libraries don't have the -@PACKAGE_API_VERSION@ in
  lib_LTLIBRARIES, which is what other packages do.

96. By Philip Ashmore

Version 2.5.2-01

* Enable -jN builds

95. By Philip Ashmore

Version 2.5.1-01

The "bucky-ball" release.

* TraceWriter nesting works
  More trace tests added to catch regressions.

* Added pointer methods to RefPtr in referenced.hpp

* misc cleanups
  Missed the "tchar.hpp" -> "tchar.h" include change in compiler.hpp.
  Added include files in a few places needed for gcc/g++ 4.7.

Branch metadata

Branch format:
Branch format 7
Repository format:
Bazaar repository format 2a (needs bzr 1.16 or later)
This branch contains Public information 
Everyone can see this information.

Subscribers