Re: feature/package-vc has been merged

[Philip Kaludercic <philipk@posteo.net>]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> diff --git a/lisp/emacs-lisp/package.el b/lisp/emacs-lisp/package.el
>> index a7bcdd214c..bf6849af65 100644
>> --- a/lisp/emacs-lisp/package.el
>> +++ b/lisp/emacs-lisp/package.el
>> @@ -462,6 +462,15 @@ package-vc-p
>>    (inline-letevals (pkg-desc)
>>      (inline-quote (eq (package-desc-kind ,pkg-desc) 'vc))))
>>  
>> +(define-inline package-lisp-dir (pkg-desc)
>> +  "Return the directory with Lisp files for PKG-DESC."
>> +  (inline-letevals (pkg-desc)
>> +    (inline-quote
>> +     (let* ((extras (package-desc-extras ,pkg-desc))
>> +            (lisp-dir (alist-get :lisp-dir extras))
>> +            (dir (package-desc-dir ,pkg-desc)))
>> +       (file-name-directory (file-name-concat dir lisp-dir))))))
>
> Any reason why this needs to be inlined?  I'd expect the inlining to
> have a negligible effect since the body contains its fair share of
> further function calls, none of which are conditional and AFAICT all the
> calls to this pass a variable as argument, so the inlining will not
> expose further optimization opportunities.

You are probably right, it would be better to convert this into a
regular function.

>> @@ -827,7 +836,7 @@ package--reload-previously-loaded
>>  byte-compilation of the new package to fail."
>>    (with-demoted-errors "Error in package--load-files-for-activation: %s"
>>      (let* (result
>> -           (dir (package-desc-dir pkg-desc))
>> +           (dir (package-lisp-dir pkg-desc))
>>             ;; A previous implementation would skip `dir' itself.
>>             ;; However, in normal use reloading from the same directory
>>             ;; never happens anyway, while in certain cases external to
>
> Why do we need this change?

Because we are only interested in the sub-directory containing the lisp
files.

>> @@ -891,7 +900,7 @@ package-activate-1
>>            (package--reload-previously-loaded pkg-desc))
>>          (with-demoted-errors "Error loading autoloads: %s"
>>            (load (package--autoloads-file-name pkg-desc) nil t))
>> -        (add-to-list 'load-path (directory-file-name pkg-dir)))
>> +        (add-to-list 'load-path (package-lisp-dir pkg-desc)))
>>        ;; Add info node.
>>        (when (file-exists-p (expand-file-name "dir" pkg-dir))
>>          ;; FIXME: not the friendliest, but simple.
>
> This should really not be needed (actually, this `add-to-list` is not
> needed at all for any package (re)installed with Emacs≥26 (or maybe even
> 25, can't remember)).  The "normal" behavior is that it's the autoloads
> file which adds to `load-path` (which makes it possible for that
> autoloads file to add the `:lisp-dir` instead of the root dir, indeed).

I see what you mean.  But this would be change that is unrelated to
package-vc, so it could just be removed directly on master.

>> @@ -1080,9 +1089,10 @@ package-autoload-ensure-default-file
>>  (defvar autoload-timestamps)
>>  (defvar version-control)
>>  
>> -(defun package-generate-autoloads (name pkg-dir)
>> -  "Generate autoloads in PKG-DIR for package named NAME."
>> -  (let* ((auto-name (format "%s-autoloads.el" name))
>> +(defun package-generate-autoloads (pkg-desc pkg-dir)
>> +  "Generate autoloads for PKG-DESC in PKG-DIR."
>> +  (let* ((name (package-desc-name pkg-desc))
>> +         (auto-name (format "%s-autoloads.el" name))
>>           ;;(ignore-name (concat name "-pkg.el"))
>>           (output-file (expand-file-name auto-name pkg-dir))
>>           ;; We don't need 'em, and this makes the output reproducible.
>
> I thought an alternative was for `package-vc.el` to call this function
> with the `:lisp-dir` as `pkg-dir`, so we don't need to change this part
> of the code.

I might be missing something, but the previous signature was missing a
package description object that the change required.

If you think it is better, I could rename the function and add a
compatibility alias.

>> @@ -1118,7 +1140,7 @@ package--compile
>>    (let ((byte-compile-ignore-files (package--parse-elpaignore pkg-desc))
>>          (warning-minimum-level :error)
>>          (load-path load-path))
>> -    (byte-recompile-directory (package-desc-dir pkg-desc) 0 t)))
>> +    (byte-recompile-directory (package-lisp-dir pkg-desc) 0 t)))
>
> Why do we need this?  AFAIK this recurses into subdirectories so using
> `package-desc-dir` still compiles all the files just fine, no?

Same as above, if we know all the lisp code is located in one
sub-directory, there is no need to compile everything in the
directory -- which might contain test files or other scripts that were
not meant to be compiled.

>>  (defun package--native-compile-async (pkg-desc)
>>    "Native compile installed package PKG-DESC asynchronously.
>> @@ -1126,7 +1148,7 @@ package--native-compile-async
>>  `package-activate-1'."
>>    (when (native-comp-available-p)
>>      (let ((warning-minimum-level :error))
>> -      (native-compile-async (package-desc-dir pkg-desc) t))))
>> +      (native-compile-async (package-lisp-dir pkg-desc) t))))
>
> Same here.

Same here.

>> @@ -2419,7 +2441,7 @@ package--newest-p
>>  
>>  (declare-function comp-el-to-eln-filename "comp.c")
>>  (defvar package-vc-repository-store)
>> -(defun package--delete-directory (dir pkg-desc)
>> +(defun package--delete-directory (dir)
>>    "Delete PKG-DESC directory DIR recursively.
>>  Clean-up the corresponding .eln files if Emacs is native
>>  compiled."
>
> IIUC this part of the change is because `package-delete` does not delete
> package-vc packages, right?  If so, I support that 100%:

This change reverts a previous feature, back when package-vc didn't load
sub-directories, but cloned the repository into some XDG directory and
created a symbolic link from ~/.emacs.d/elpa to the right sub-directory.
To do that, knowledge of what package was been deleted was required:

-  (if (and (package-vc-p pkg-desc)
-           (require 'package-vc)   ;load `package-vc-repository-store'
-           (file-in-directory-p dir package-vc-repository-store))
-      (progn
-        (delete-directory
-         (expand-file-name
-          (car (file-name-split
-                (file-relative-name dir package-vc-repository-store)))
-          package-vc-repository-store)
-         t)
-        (delete-file (directory-file-name dir)))
-    (delete-directory dir t)))
+  (delete-directory dir t))

Now that this isn't being done anymore, I could revert the change and
simplify the code back to the original state (actually it has become
slightly more complicated since, because I noticed that
`package-vc-install-from-checkout' requires some special handling, but
that doesn't require a package description to be detected).

> Packages installed with `package-install` can be deleted without too
> much fuss because they usually don't contain any important info, but for
> `package-vc` this is completely different and we should either never
> delete them or at least not without a minimum of sanity checks
> (uncommited changes, stashes, additional branches) and very
> explicit prompt.

This is currently not done, but could be added.  I am imagining checking
of the package directory is the root directory of a repository (does
that work for all VCS?), and if so double-prompting the user.  But I
would be opposed to preventing users from deleting packages installed
using `package-vc', if only it would go against my workflow of fetching
the sources for a package, preparing and sending a patch, and then
reverting to the release package.

>> @@ -2549,7 +2560,7 @@ package-recompile
>>      ;; load them (in case they contain byte code/macros that are now
>>      ;; invalid).
>>      (dolist (elc (directory-files-recursively
>> -                  (package-desc-dir pkg-desc) "\\.elc\\'"))
>> +                  (package-lisp-dir pkg-desc) "\\.elc\\'"))
>>        (delete-file elc))
>>      (package--compile pkg-desc)))
>
> Same as above.

The explanation remains the same.

>
>         Stefan