On 3/28/07, Sebastien Tandel <[EMAIL PROTECTED]> wrote:
>
> >>> I dont think it is really realistic to have all autogenerated files
> >>> always compile without any warnings.
> >>> Maybe we should instead split Makefile.common up into three parts :
> >>>
> >>> First part : normal dissectors
> >>>
> >>> Second part : ANS2WRS generated dissectors which take extra compile
> >>> time flags and definitions to suppress artefacts from the compiler.
> >>>
> >>> Third part : PIDL generated dissectors that once again take extra
> >>> compile parameters and definitions.
> >>>
> >>>
> >> Sounds like a reasonable solution. It would also have the advantage that
> >> we get a list of generated dissectors, which we currently don't have.
> >>
> >> But: Is there a "makefile magic", so you can have different CFLAGS
> >> settings for the different lists? Remember that we are using implicit
> >> rules here.
> >>
> >> Having three lists shouldn't be difficult to add, but I don't know how
> >> to get the "compile logic" for it - however, I'm really not an expert on
> >> this.
> >>
> >> If there's a solution for this makefile problem, I think it's the way to
> >> go ...
> >>
> >
> > When it comes to the ANS2WRS dissectors are concerned I think that
> > ANS2WRS is so mature now that maybe what we should do for those
> > dissectors would instead be to
> > remove them completely from epan/dissectors and make asn part of the
> > build process.
> > All these dissectors have their own makefile down in asn/* where one
> > can set additional or different compiler flags compared to those used
> > by the normal epan/dissectors.
> > These asn dissectors could even be linked to a separate dll :
> > libwireshark-asn.dll
> >
> Wouldn't be the register_all_protocols() function problematic?
We could create a plugin directory with all its paraphernalia and have
the asn1 dissectors be loaded like a plugin.
But still the best thing would be to have the asn1 generated
dissectors inepan/dissectors and have them built from the Makefile.
For the warnings we could have asn2wrs.py prepend the #pragma for
unused static function right bellow the signature to the generated
code (that's an ugly fix but I do not find a feasible way to have
asn2wrs not creating the unused functions).
The rest is monkey business: just to add the generated targets to the
Makefile[s] like this:
packet-h248.c: ../../asn1/h248/packet-h248-template.c \
../../asn1/h248/packet-h248-template.h \
../../asn1/h248/h248.cnf \
../../asn1/h248/h248v3.asn
[TAB](cd ../../asn1/h248 && make copy_files )
packet-h248.h: packet-h248.c
Luis
> Regards,
> Sebastien Tandel
> _______________________________________________
> Wireshark-dev mailing list
> Wireshark-dev@wireshark.org
> http://www.wireshark.org/mailman/listinfo/wireshark-dev
>
--
This information is top security. When you have read it, destroy yourself.
-- Marshall McLuhan
_______________________________________________
Wireshark-dev mailing list
Wireshark-dev@wireshark.org
http://www.wireshark.org/mailman/listinfo/wireshark-dev
