jhardlin (Friday, 04. June 2010)
> In git log, file history is scattered and difficult to find.

Ok, that's true. But I use 'git log' only if I'm really interested in 
some special information (e.g.: "Which idiot wrote this nonsense? Oh 
god, it was me again...").

> Every po file is signed. Why not xml?

What do mean with "signed"? The header?

> History tag exists in the xml language. Why not use it?

If I remember correctly we used to (mis)use <revhistory> elements in 
the multi-language XML files, but in a rather strange and useless way.

Adding <revhistory>s again and using them according to the examples in 
"DocBook: The Definitive Guide" may be better than using our selfmade 
"section history" comments (we should suppress the output to HTML, 
IMHO). But isn't one <revhistory> for every XML file an overkill?

> Sure, history is not necessary to work on a xml file.

More interesting IMHO: does it *help* you if you work on an XML file?

> [...] I am not young, 

Oh, so we are two! :-)

> and I find pleasant to have all the history of the file summarized
> at the beginning without needing to browse all the log over.

Try this shell command:

for d in appendix concepts dialogs filters glossary introduction \ 
menus toolbox tutorial using
        find src/$d -type f -name '*.xml' | head -n 20 | tail -n 3
done |
xargs sed -e '/[Ss]ection [Hh]istory/,/-->/!d'

This displays some more or less randomly picked section history 
comments, and IMHO most of them are totally useless. 

I still think we should remove them or find an alternative way to 
provide some meta information.


Attachment: signature.asc
Description: This is a digitally signed message part.

Gimp-docs mailing list

Reply via email to