Rosetta should display something for consecutive spaces in the middle of a line of a msgid

Hi Malcolm, could you confirm if this is a dupe of bug 2627. Although 2627 mentions only bug comments, the fix for it would likely apply to everywhere in launchpad, including Rosetta.

It is the certainly the same issue as in bug 2627.
I cannot say if fixing it in Malone would also fix it in Rosetta, or if the fixes for both problems are the same.
Rosetta is already handling leading and trailing whitespace in a different way to Malone.

Carlos, Danilo,

Is this something that you guys are going to take care in rosetta only or could be fixed fixing bug 2627?

Malcolm... I think I just realize what you were complaining about... I didn't noticed that the browser was collapsing all whitespaces as being just one. I thought it was just a matter of displaying an special symbol for them because was hard to see them.

Diogo, about your question, I really don't know. If there is an easy way to solve this in our infrastructure, it's fine for me to change our code to use bug #2627 solution, if it's not possible, I would like mpt's opinion about this, whether we should use ' ' or just the middle dot symbol we use at the beginning or end of a line.

  wouldn't be enough, because you can't assume *everyone* is copying and pasting. :-) I think the space symbol (at least, once the fix for bug 46 lands) would be better.

However, I don't think either approach would actually be useful. For example, if a developer has a string containing a row of table headings, the appropriate number of spaces before a heading (so that it lines up with its column) will depend on the length of the previous heading, which will depend on the translation. So for Rosetta to suggest "four spaces should go here" would be misleading. Translating such a string is going to be hard regardless of translation system, so I suggest this is actually a localizability bug in the software.

On 20/09/06, Matthew Paul Thomas <email address hidden> wrote:

> However, I don't think either approach would actually be useful. For
> example, if a developer has a string containing a row of table headings,
> the appropriate number of spaces before a heading (so that it lines up
> with its column) will depend on the length of the previous heading,
> which will depend on the translation. So for Rosetta to suggest "four
> spaces should go here" would be misleading. Translating such a string is
> going to be hard regardless of translation system, so I suggest this is
> actually a localizability bug in the software.

You're right, a string like that makes translation hard.
procps contains this string in the "free" program:
" total used free shared buffers cached\n"
procps isn't translatable, so this doesn't matter.

However, the most common strings with consecutive spaces are "--help"
strings, where the text in the first column is the option name, and
the text in the second column is a description. The first column
should not be translated, so rosetta would be displaying the correct
number of spaces needed to line up the second column.

It seems that translators are getting this wrong:

$ LANGUAGE=es_ES make --help
Modo de empleo: make [opciones] [objetivo] ...
Opciones:
  -b, -m No se tendrá en cuenta por compatibilidad.
  -B, --always-make Hace incondicionalmente todos los objetivos.
  -C DIRECTORIO, --directory=DIRECTORIO
                              Se cambia al DIRECTORIO antes de hacer nada.
  -d Se imprimirán grandes cantidades de información de depurado.
  --debug[=BANDERAS] Se imprimirán varios tipos de información de depurado.
  -e, --environment-overrides
                Las variables ambientales se imponen a las de los makefiles.
  -f ARCHIVO, --file=ARCHIVO, --makefile=ARCHIVO
                                   Lee al ARCHIVO como un makefile.
  -h, --help Muestra este mensaje y finaliza.
  -i, --ignore-errors No se toman en cuenta los errores
provenientes de las instrucciones.
  -I DIRECTORIO, --include-dir=DIRECTORIO
                       Busca dentro del DIRECTORIO los makefiles incluidos.
  -j [N], --jobs[=N] Se permiten N trabajos a la vez; si no se
especifica un
argumento son infinitos.
  -k, --keep-going Sigue avanzando aún cuando no se puedan
crear algunos objetivos.
  -l [N], --load-average[=N], --max-load[=N]
      No inicia con trabajos múltiples a menos que la carga estÃ(c)
por debajo de N.
  -L, --check-symlink-times Utiliza el último mtime entre los
enlaces simbólicos y los objetivos.
  -n, --just-print, --dry-run, --recon
                              No ejecuta ninguna instrucción; sólo
las muestra.
  -o ARCHIVO, --old-file=ARCHIVO, --assume-old=ARCHIVO
                           Supone que ARCHIVO es muy viejo y no lo reconstruye.
  -p, --print-data-base Se imprime la base de datos interna de `make'.
  -q, --question No se ejecutan las instrucciones;
el estado de salida
indicará si están actualizados.
  -r, --no-builtin-rules Se deshabilitan las reglas...

Example string
https://launchpad.net/distros/ubuntu/dapper/+source/gcc-3.4/+pots/gcc-3.4/en_GB/1157/+translate