Using a real parser generator to parse numbers (and dates)

[Lionel Elie Mamane]

On Wed, Feb 29, 2012 at 11:49:38AM +0100, Lionel Elie Mamane wrote:

> In the context of fdo#41166 (a bug in how our numbers/date parser
> handles ambiguities), I thought "wouldn't it be better if we used a
> proper parser generator and wrote a proper grammar instead of a
> hand-coded parser"? It would be easier to debug, easier to maintain,
> easier to read, ... and thus probably less buggy in the long run.

> Well, I've a good "proof of concept" start on making it a
> Boost::Spirit grammar.

Grmbl... Forgot the attachments. Here they are.

-- 
Lionel
-------------- next part --------------
A non-text attachment was scrubbed...
Name: lo_classes.hxx
Type: text/x-c++hdr
Size: 8044 bytes
Desc: not available
URL: <http://lists.freedesktop.org/archives/libreoffice/attachments/20120229/dcd881f5/attachment-0002.hxx>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: lo_classify.hxx
Type: text/x-c++hdr
Size: 7233 bytes
Desc: not available
URL: <http://lists.freedesktop.org/archives/libreoffice/attachments/20120229/dcd881f5/attachment-0003.hxx>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: use-spirit.patch
Type: text/x-diff
Size: 70957 bytes
Desc: not available
URL: <http://lists.freedesktop.org/archives/libreoffice/attachments/20120229/dcd881f5/attachment-0001.patch>