[sword-devel] verseparsing-utf8

Dmitrijs Ledkovs dmitrij.ledkov at gmail.com
Sun May 24 14:15:34 MST 2009

2009/5/24 Jonathan Marsden <jmarsden at fastmail.fm>:
> Dmitrijs Ledkovs wrote:
>> Verseparsing-utf8 is failing.....
> Confirmed.  I saw that too a while back...
>> ../src/mgr/filemgr.cpp: In member function ‘signed char
>> sword::FileMgr::trunc(sword::FileDesc*)’:
>> ../src/mgr/filemgr.cpp:276: warning: ignoring return value of ‘ssize_t
>> write(int, const void*, size_t)’, declared with attribute
>> warn_unused_result
> These are annoying to read, but relatively harmless (unless you do
> -Werror of course :) .  They do mean that the code concerned is not
> really doing sufficient error checking, which should IMO be fixed; I
> think I provided some ideas/a patch for at least some of these write()
> calls earlier.

So if it fails to write (e.g. out of memory / HDD) then sword will
still think everything succeeded fine? Seems quite dangerous to me
I've had my system running with 0MB free space left on HDD for a
while. It wasn't fun when programs assumed that certain things

>> LocaleMgr::getLocale failed to find de
> These mean the SWORD library can't find the SWORD de locale data on your
> system.  I'm not sure it will be able to do so, unless you actually have
> SWORD installed.  In your example you did:
>> $ ./configure --with-icu --enable-tests
> so you'd need the SWORD locale data under /usr/local/share/sword/ (as I
> understand it) for these tests to work as designed.

well I installed our debian package and yes _now_ the test works.

> If these tests are supposed to work "out of the tarball" with no SWORD
> library and locale data yet installed, then they probably need fixing so
> that they are able to "find" the SWORD locale data inside the
> tarball/build area??

Sounds reasonable to me. That's how every other well-known test suites
work they test what they have just build not was is installed and
maybe is functional. And the previous cppunittest was working with the
build data, not installed data.

Again I'm not a C++ expert so I can only report the bug. On the bright
side i now now how to run both cppunittest and the shell script based
one at runtime. So dear sword developers if you want testsuite to be
run on 12+ architectures by debian buildd machines please write one
which works ;-) either by updating cppunittest so that it doesn't fail
or move to a shell script one.

With best regards

Dmitrijs Ledkovs (for short Dima),
Ледков Дмитрий Юрьевич

More information about the sword-devel mailing list