Re: [Bug 465460] anyone has take a look?

[thomas harding laposte net (Thomas Harding)]

On Mon, Jun 02, 2008 at 08:33:31PM +0200, Hans Breuer wrote:
> I get mails about every dia bug change. But that does not mean I can make 
> immediately sense of all of them ;) This particular bug report appeared to 
> me as some improvement and some regression. But I don't really feel 
> qualified to review it. So input from someone who understands more of 
> DocBook than I do would really be appreciated.

?
I just dropped the entity hack about graphic extention, while
format attribute is good enough for Yelp, and the correct way
is to setup graphic.default.extension in stylesheet to build html.
 
> > Last patch filled addresses several problems about documentation:
> >   
> >   * scrollkeeper really works, so xml docs are now accessibles directly
> >     via yelp.
> Could you comment on your replacements of <para/> to <simpara/>. Does it 
> have to do with producing less pages only containing the headline and a link?

The goal is to let admonitions (note, tip, warning...) and related
text on the same line, as para is block and simpara not. Note I changed
only the first para in admonitions to simpara.

> >   * hardcopies are done for all languages (there are still problems
> >     with pl, at least on Debian, on ps and pdf outputs, as lll TeX
> >     macro is defined twice somewhere in the tools used).
> Is this a regression?

No, as there were simply no hardbooks built for pl and eu.

> >   * All documentation processes are done automatically, hardcopies
> >     are compiled while --with-hardbooks option is given to ./configure
> will check if it work out of the box for me when I have access to my Linux 
> box again.

Yep, note that if you don't have dblatex, pdf and postscripts outputs
will be discarded silently.

> > However, I detected a new problem:
> >   
> >   * when compiled without gnome support, Dia is not able to open
> >     and list documentation directory. I filled a short strace 
> >     output about that.
> THe exact error message given would be interesting. Looking at the code it 
> either includes the path or it dose not. The latter would be an issue with 
> the content of the directory (i.e. missing index.html. The former should 
> include more detailed information _why_ opening the directory failed.

There is no error message, even in a console.

read(3, "\1\2\233\27\0\0\0\0`\0\300\1\4\0\0\0(\0\0\0\4\0\0\0p(\34"..., 32) = 32
open("/usr/local/share/dia/help", O_RDONLY|O_NONBLOCK|O_LARGEFILE|O_DIRECTORY) = 6
fstat64(6, {st_mode=S_IFDIR|S_ISGID|0755, st_size=4096, ...}) = 0
fcntl64(6, F_SETFD, FD_CLOEXEC)         = 0
getdents64(6, /* 7 entries */, 4096)    = 168
getdents64(6, /* 0 entries */, 4096)    = 0
close(6)                                = 0
rt_sigaction(SIGINT, {SIG_IGN}, {0xb7079120, [], 0}, 8) = 0
rt_sigaction(SIGQUIT, {SIG_IGN}, {SIG_DFL}, 8) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
clone(child_stack=0, flags=CLONE_PARENT_SETTID|SIGCHLD, parent_tidptr=0xbfda5d44) = 5272
waitpid(5272, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0) = 5272
rt_sigaction(SIGINT, {0xb7079120, [], 0}, NULL, 8) = 0
rt_sigaction(SIGQUIT, {SIG_DFL}, NULL, 8) = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
--- SIGCHLD (Child exited) @ 0 (0) ---
gettimeofday({1211926549, 660413}, NULL) = 0

7 entries means the 4 languages plus "C" symlink, . and ..


> BTW: it would be better to have this information in a different bug-report 
> *if* the files are installed in the correct place.

Will fill.

> > Again, I would emphasize that documentation is a matter of quality on
> > programs.
> Yeah, but what helps the best documentation if it does not get read ;)

Mmmph :P

-- 
Thomas Harding