Re: commit: Most of the code for the GtkMathView plugin.

From: Luca Padovani <lpadovan_at_cs.unibo.it>
Date: Tue Jan 25 2005 - 11:47:33 CET

Hi Martin, Jean,

On Tue, 2005-01-25 at 11:37 +0100, Jean Bréfort wrote:
> > ./configure --prefix=******* --disable-boxml --disable-gmetadom
> > --disable-libxml2-reader --disable-custom-reader --enable-pipe
>
> Are all these flags mandatory?

Short answer: no, they are not.

Explanation: gtkmathview comes with 4 different frontends depending on
which XML method one wants to use for reading the source MathML
documents. The methods are:
- use libxml2, building the XML tree (method used by AbiWord)
- use libxml2, just reading the XML stream
- use gmetadom, building the DOM tree
- use a custom frontend (must be provided by the host application)

Because of the heavy use of C++ templates in gtkamthview source code,
the compilation of all these different frontends can take quite some
time. During development and bug tracking this gets really annoying. So,
all those --disable-* flags basically disable the frontends that are not
used by AbiWord.

The --enable-pipe flag enables a flag in GCC which makes it run slightly
faster (you can save up to 2/3 minutes when compiling gtkmathview from
scratch), but it's not mandatory.

The --prefix=* flag is necessary if the default installation directory
(/usr/local) is not what you want (or it is incompatible with the
distribution's policies, like under Debian). AFAIK, it is rather common
to have the --prefix flag specified in the package building scripts.

In order to detect if the right frontend has been compiled (where
"right" means the one AbiWord uses), one can do

        pkg-config mathview-libxml2 --cflags --libs

In fact this is what happens in ac-helpers/abi-mathview.m4.

As for including gtkmathview's source tree into AbiWord's, I hope this
is not necessary nor useful (I did my best to make gtkmathview pluggable
and configurable as much as possible :-) )

Let me know if anything's not clear

Cheers
--luca
Received on Tue Jan 25 11:48:17 2005

This archive was generated by hypermail 2.1.8 : Tue Jan 25 2005 - 11:48:19 CET