Compiler Warnings

[David L. Andre <DLA at MIT-AI>]

    RMS@MIT-AI 01/26/82 00:12:16 Re: Compiler Warnings
    I would like to see a hybrid scheme in which the warnings history list
    is the primary data structure but the Compiler Warnings buffer is also
    available.  It is easy to keep the buffer up to date by remembering a
    copy of the warnings history list used to update it last time,
    finding the differences between the two lists, and updating the buffer
    according to those differences.
This is almost exactly what I proposed.  The only difference, as I see
it, is that the Compiler Warnings buffer doesn't exist by default in my
scheme, and also I have a simpler idea on updating it (set up again when
changed).

    Another desirable feature would be that warnings about functions used
    but not defined could be stuck retroactively onto the warnings lists
    for the calling functions.  This would be the right place for them to
    appear to interact properly with Edit Compiler Warnings.
This is interesting.  In the case of misspelled functions, that would be
a help.  In the case of functions you haven't written yet, it would
serve no purpose.  I agree that the functions called but not defined
should be stored with the errors for the calling functions.  I also
think that PRINT-COMPILER-WARNINGS should continue to print a summary of
these at the end.