#148: Missing files in the distribution of Olena 0.11
------------------------+---------------------------------------------------
Reporter: levill_r | Owner: Olena Team
Type: defect | Status: new
Priority: trivial | Milestone: Olena 0.11.1
Component: Milena | Version: 0.11
Keywords: EXTRA_DIST |
------------------------+---------------------------------------------------
Reported by Théo, relaunching `bootstrap.sh` from a distributed tarball of
Olena 0.11:
{{{
este% ./bootstrap.sh
sh: Can't open ./cleanup.sh
sh: Can't open ./gen_filelists.sh
}}}
--
Ticket URL: <https://trac.lrde.org/olena/ticket/148>
Olena <http://olena.lrde.epita.fr>
Olena, a generic and efficient C++ image processing library.
#217: Nasty bug in mln::data::impl::in memcpy_ with g++ 4.2 on Debian GNU/Linux
on IA-32
----------------------+-----------------------------------------------------
Reporter: levill_r | Owner: Olena Team
Type: defect | Status: new
Priority: critical | Milestone: Olena 1.1
Component: Milena | Version: 1.0
Keywords: |
----------------------+-----------------------------------------------------
Running the following test exhibits this issue:
{{{
make -C milena/apps/graph-morpho CXX="ccache g++-4.2" checkTESTS=samples-
image2d TESTS_ENVIRONMENT="valgrind"
}}}
This seems to be a bug (or a weird implementation-defined behavior) in
Debian's `g++` 4.2 on IA-32, due to some strict-aliasing rule not
respected in Milena's code. I added a `#warning` in `memcpy_.hh` in r4608.
But the thing is, `g++` 4.2 is the main compiler in Debian GNU/Linux 5.0,
which is used on most GCC-based Olena buildbots, meaning we won't get a
successful build on these hosts until we
circumvent this bug. That's bad.
--
Ticket URL: <https://trac.lrde.org/olena/ticket/217>
Olena <http://olena.lrde.epita.fr>
Olena, a generic and efficient C++ image processing library.
#201: Use « olena » instead of « oln » as an identifier
----------------------+-----------------------------------------------------
Reporter: levill_r | Owner: Olena Team
Type: defect | Status: new
Priority: major | Milestone:
Component: Olena | Version: 1.0
Keywords: |
----------------------+-----------------------------------------------------
We should no longer use « `oln` » ; « `olena` » is much more clearer.
* The Subversion repository is called `oln`, because `olena` was already
used (arg!); we should not do this mistake again: the future Git
repository shall be named `olena`.
* Likewise, Olena's build farm URL
(https://buildfarm.lrde.org/buildfarm/oln/) should be renamed to `olena`.
If you find some other places where the project is called `oln` instead of
`olena`, please report them here. TIA.
--
Ticket URL: <https://trac.lrde.org/olena/ticket/201>
Olena <http://olena.lrde.epita.fr>
Olena, a generic and efficient C++ image processing library.
#126: Speed up tests
-------------------------+--------------------------------------------------
Reporter: levill_r | Owner: Olena Team
Type: enhancement | Status: new
Priority: major | Milestone: Olena 1.0ß
Component: Milena | Version: 1.0
Keywords: |
-------------------------+--------------------------------------------------
Some tests take a very long time to execute. Main causes are:
* input image(s) are too big [[BR]] => they should be reduced (sometimes,
testing on `lena.pgm` is just too costly);
* the default, non-optimized compilation settings (`-O0 -ggdb`) produces
very inefficient code for certain algorithms [[BR]] => we should locally
increase the level of optimization (e.g., `-O1` or `-O2`) for these tests,
and '''document''' this in `Makefile.am` (explain why we made changes in
`CXXFLAGS` for a given test).
We might want to make a list of lengthy tests here before solving this
ticket.
--
Ticket URL: <https://trac.lrde.org/olena/ticket/126>
Olena <http://olena.lrde.epita.fr>
Olena, a generic and efficient C++ image library.
#214: Use Automake 1.11's AM_DEFAULT_SOURCE_EXT
-------------------------+--------------------------------------------------
Reporter: levill_r | Owner: Olena Team
Type: enhancement | Status: new
Priority: minor | Milestone:
Component: Milena | Version: 1.0
Keywords: Make |
-------------------------+--------------------------------------------------
Quoting the [http://www.mail-archive.com/automake@gnu.org/msg15022.html
announce of Automake 1.11]:
{{{
- The default source file extension (.c) can be overridden with
AM_DEFAULT_SOURCE_EXT now.
}}}
Which allows us to replace
{{{
bin_PROGRAMS = foo bar baz
foo_SOURCES = foo.cc
bar_SOURCES = bar.cc
baz_SOURCES = baz.cc
}}}
with
{{{
bin_PROGRAMS = foo bar baz
AM_DEFAULT_SOURCE_EXT = .cc
}}}
(Of course, « `AM_DEFAULT_SOURCE_EXT = .cc` » should be factored in some
`common.mk` file.)
This just rocks if you consider the tons of `*_SOURCES` lines in the
zillions of `tests/*/Makefile.am` files.
On the downward side, this change requires Automake 1.11, which is not
even in Debian unstable as of July 24, 2009. We might want to wait a
little bit before requiring this version.
--
Ticket URL: <https://trac.lrde.org/olena/ticket/214>
Olena <http://olena.lrde.epita.fr>
Olena, a generic and efficient C++ image processing library.
#207: Use Automake 1.11
-------------------------+--------------------------------------------------
Reporter: levill_r | Owner: levill_r
Type: enhancement | Status: new
Priority: major | Milestone: Olena 1.1
Component: Olena | Version: 1.0
Keywords: |
-------------------------+--------------------------------------------------
Automake 1.11 provides a better framework for tests: each test can
generate a log file, and the test suite is a mere rule generating a
summary log, depending on all these logs. The `check` rule is no longer
based on a shell `for` loop, and `make` can truly honors its `-j`
(parallel instances) option when testing. Of course, one need to be
careful with side effects, notably files created by tests, as one test may
overwrite the output(s) of another tests. Akim recommends that test `foo`
writes its outputs to `foo.dir`.
Moreover, this permits many improvements:
* One can generate pretty-printed versions of the tests using, e.g.,
`rst2html` (if the log contains an RST output) on the logs.
* One can have a `recheck` rule reruning only the failed tests, since the
first line of the log records the outcome of the test (pass or fail).
* Also, provided dependencies are well set, this can probably help one
write rules to rerun only the needed tests when a change occurred in the
source tree (but I don't know yet how to do this).
--
Ticket URL: <https://trac.lrde.org/olena/ticket/207>
Olena <http://olena.lrde.epita.fr>
Olena, a generic and efficient C++ image processing library.
#129: Reintegrate oln.m4 (full or a part of it) into Olena 1.0
-----------------------------------+----------------------------------------
Reporter: levill_r | Owner: levill_r
Type: enhancement | Status: new
Priority: minor | Milestone: Olena 1.0ß
Component: Olena | Version: 1.0
Keywords: configure autoconf M4 |
-----------------------------------+----------------------------------------
In Olena 0.11, `oln.m4` (which might be renamed to `olena.m4`) provided
useful services to both the Olena package (search for libraries, set up
flags for tests, etc.) and autoconfiscated packages using Olena, by
providing a `--with-oln` (to be renamed to `--with-olena`) flag.
We should reintegrate a part of these features, possibly in two M4 macro
packages:
* one for Olena,
* one for Olena-based projects.
--
Ticket URL: <https://trac.lrde.org/olena/ticket/129>
Olena <http://olena.lrde.epita.fr>
Olena, a generic and efficient C++ image library.