[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Safety of elisp-flymake-byte-compile (Was Re: [Emacs-diffs] scratch/

From: João Távora
Subject: Re: Safety of elisp-flymake-byte-compile (Was Re: [Emacs-diffs] scratch/allow-custom-load-paths)
Date: Fri, 14 Dec 2018 13:38:39 +0000
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/27.0.50 (windows-nt)

Stefan Monnier <address@hidden> writes:

>> Oh, I didn't know about load-history. So is this usable?
>>     (defun moderately-trusted-p (file)
>>       (catch 'done
>>         (let ((file (expand-file-name file))
>>               (lhist load-history)
>>               (lpath load-path)
>>               probe)
>>           (and
>>            (string-match "\\.elc?$" file)
>                                    ^
>                                   \\'

>>            (while (setq probe (pop lhist))
>>              (when (and (car probe)
>>                         (equal (file-name-sans-extension file)
>>                                (file-name-sans-extension (car probe))))
>>                (throw 'done t)))
>>            (while (setq probe (pop lpath))
>>              (when (equal (file-name-as-directory
>>                            (file-name-directory file))
>>                           (file-name-as-directory probe))
>>                (throw 'done t)))))))
>> Or do you want to write this in C for speed?  Where should this be
>> plugged into?  load in lread.c directly?
> I don't understand the question.  AFAIK this would only be used within
> bytecomp.el when processing a `defmacro` in order to decide whether it
> should be added to byte-compile-macro-environment or not.

Oh, right... of course, never mind.  Can you do that bit?  I suppose we
have to check some new var byte-compile-just-checking-p and
(moderately-trusted-p byte-compile-current-file), right?

Is it in `byte-compile-file-form-defalias'

>> We should still hand load-path and load-history, or a part of
>> load-history down to the flymake'in slave emacs, right?
> Ah, right, we'd then need to pass the `car` of load-history entries to
> the slave.

And not the load-path?


reply via email to

[Prev in Thread] Current Thread [Next in Thread]