[M3devel] m3tests/cmp_fie vs. compare_file

Jay K jay.krell at cornell.edu
Thu Jul 1 12:26:15 CEST 2010


Understood. They do a lot of the same but merging requires both a decent top-down understanding of the
semantics to plan ahead slightly and then a close bottom-up line by line read to merge and preserve meaning.
(I seem to be speaking like an MBA this evening. Don't worry. :) )

I don't have the energy for it at the moment.


I am extending this area though, the stuff for recording and later comparing assembly output.
Background: Fixing more configure -enable-checking seems to regress code /slightly/, in bit insert/extract.
So I haven't commited said fix.
So I'm going to try the bitfield refs again (that Tony had done).
That might both fix and improve the code.


Once they are merged the result can probably keep the name compare_file.


Thanks for the overview. I'll try to look into this eventually and I'll try not to look into it too soon,
before other things that are probably more useful (and more interesting and fun).


Btw, there's stuff called "jux" in the code here.
Is it rude to request sticking to English? Thanks.


 - Jay


----------------------------------------
> Date: Thu, 1 Jul 2010 11:38:53 +0200
> From: wagner at elegosoft.com
> To: m3devel at elegosoft.com
> Subject: Re: [M3devel] m3tests/cmp_fie vs. compare_file
>
> Quoting Jay K :
>
> >
> > I find it confusing that m3-sys/m3tests/src/m3makefile contains both
> > cmp_file and compare_file (unless one was a thin wrapper over the
> > other, not the case)
> > Olaf you added cmp_file in
> > http://modula3.elegosoft.com/cgi-bin/cvsweb.cgi/cm3/m3-sys/m3tests/src/m3makefile.diff?r1=1.3;r2=1.4.
> >
> > Presumably they can be merged into one function?
>
> compare_file() writes output and doesn't return anything meaningful
> (it's a procedure), while cmp_file is a boolean state function that
> doesn't write anything to stdout.
>
> The first could be based on the second.
>
> Of course you can find a common signature, but then you'll have to
> change all occurences and add a use case parameter, which isn't nice,
> too.
>
> Perhaps you can suggest better names?
>
> Please do not change the semantics (i.e. the use of @cmp, for example)
> though, as that might break the regression tests.
>
> Olaf
> --
> Olaf Wagner -- elego Software Solutions GmbH
> Gustav-Meyer-Allee 25 / Gebäude 12, 13355 Berlin, Germany
> phone: +49 30 23 45 86 96 mobile: +49 177 2345 869 fax: +49 30 23 45 86 95
> http://www.elegosoft.com | Geschäftsführer: Olaf Wagner | Sitz: Berlin
> Handelregister: Amtsgericht Charlottenburg HRB 77719 | USt-IdNr: DE163214194
>
 		 	   		  


More information about the M3devel mailing list