Re: More verbosity

[Bruno Haible]

Hello,

c.buhtz@posteo.jp wrote:
> I'm using xgettext (GNU gettext-tools) 0.21 in Debian 12. I have three 
> issues influencing each other in the context of verbosity of xgettext.
> 
> 
> 1.
> My primary problem is that xgettext doesn't tell me (on stdout) which 
> source files it reads.

There is no particular message that gets printed for each file
that gets read. But...

> For example:
> 
>      xgettext --verbose --output=message.pot common/*.py qt/*.py
> 
> The message.pot is generated but I don't see which py-files where 
> involved.

a) When the xgettext command succeeds, it will have read all the argument
   files. You can list them through
     $ echo common/*.py qt/*.py
   or
     $ ls -1 common/*.py qt/*.py

b) To observe what the xgettext program really does, you can use strace:
     $ strace --output=message.pot common/*.py qt/*.py 2>&1 | grep '^open'

c) In xgettext's output, lines that start with '#:' reference the source
   files. See the GNU gettext manual, chapter "PO Files"
   <https://www.gnu.org/software/gettext/manual/html_node/PO-Files.html>.

> 2.
> The "--verbose" has no effect. No output is generated. Why then does 
> this option exist?

It has an effect for the extraction from Ruby source files.

> 3.
> The manpage tell me that "--verbose" does "increase verbosity level". 
> What is the limit of that level? Can I do "-vvvvvvvv"? Where is the 
> limit.

The variable that counts the number of 'v' is an 'int'. Therefore,
providing this option more than 2147483647 times will lead to undefined
behaviour.

Bruno