Re: [Gnu-arch-users] robin's per-directory regexps

[Robin Farine]

>>>>> Tom Lord writes:

    > * the inventory tag of ".arch-inventory" files

    >   Robin's code always assigns a .arch-inventory file
    >   a tag of the form:

    >   ?relative/path/.arch-inventory

    >   This is a definate bug.

There is one patch to back out to remove this (patch-10 in my
archive). I first used a version which required an add or a tagline,
but then I thought that a name tag would just work. Apparently not.

    >   Correcting this bug, however, will break revisions 
    >   that already use the .arch-inventory patch.  

Some of mine will break but I was prepared to such changes so no
problem on my side to remove this.

    > * .arch-inventory as control file

    >   Surely we don't want inventory to consult a .arch-inventory
    >   file for per-directory regexps but then _not_ treat that file
    >   as source.

Yes, exactly.

    >   This change will change the semantics of `inventory' in that
    >   tree.

Sure.

    >   Now, I tend to think it is sufficiently unlikely that anyone
    >   is already using the name .arch-inventory as a non-source 
    >   file.   So I think I'm willing to go ahead and break upward
    >   compatability in that way.

And even if this clash happens, it would not endanger the archive.
Just force the user to either remove his .arch-inventory stuff or
change his script.



    > * semantics review

[excellent description of how the patch affects the inventory process]

    >   The idea of simply ignoring global regexps if dir-local ones are
    >   present is conceptually simpler, but I think less practical.  If
    >   eventually we want that, we can add an `unrecognized2' regexp
    >   keyword to .arch-inventory syntax, tested after the source regexp,
    >   so that:

    >   unrecognized2 .*

Yes, although the idea remains to have the global regexps handling
almost every file correctly. The dir-local regexps should only be used
to handle eccentricities of non-arch upstream trees. So I did not see
a real need for this use case.

    >   Finally, the patch contains no provision for a .arch-inventory file
    >   which applies recursively to subdirs.   I'm not sure one is actually
    >   needed.

Me neither, for the same reason as above.

    > * Changes to the "ancient defaults"

    >   The patch changes the "ancient default regexps" to add
    >   .arch-inventory to the excludes regexps.

    >   This is, strictly speaking, a bug:

Yeah, probably due to robotic behavior of my part.


    > * other issues

    >   There are some other minor issues about code formatting and coding 
    >   style -- more a reflection of my quirky fussiness over some issues
    >   than anything else.

You probably refer to the new data-structure and the CPP macro used to
compute an array index. I agree that regexp related structures and the
code that parse regexps should be refactorized. But I could not find a
satisfying solution which would not break current archives (mainly
changes in the format of =tagging-method).


Thanks Tom for reviewing so carefully this changes.


Robin