[bug#49934,1/3] gnu: Add libx86emu

* gnu/packages/hardware.scm: (libx86emu): New variable.

--
2.32.0

On Sat, Aug 07, 2021 at 09:58:55PM +0000, phodina via Guix-patches via wrote:
> +  (home-page "https://github.com/wfeldt/libx86emu")

FYI this is the same package as <http://issues.guix.gnu.org/45744>.

I don’t know what became of it.

Hi

On Sunday, August 8th, 2021 at 12:10 PM, pelzflorian (Florian Pelz) <pelzflorian@pelzflorian.de> wrote:

> On Sat, Aug 07, 2021 at 09:58:55PM +0000, phodina via Guix-patches via wrote:
>
> > -   (home-page "https://github.com/wfeldt/libx86emu")
>
> FYI this is the same package as http://issues.guix.gnu.org/45744.
>
> I don’t know what became of it.

Thanks. I didn't know there's a patch already. The aim of this one is to get the hwinfo and hw-prober to get the hardware details.

I've not encountered this situation. So what do you think is the best approach?

Use the patch from Vincent Legoll and refactor the other two?

So that both issues are closed?

Kind regards
Petr

On Mon, Aug 09, 2021 at 12:00:00PM +0000, phodina wrote:
> Thanks. I didn't know there's a patch already. The aim of this one is to get the hwinfo and hw-prober to get the hardware details.

I think it is good to get your patch into shape and merged.  While I
haven’t been involved in package review (shame on me), I’ll give it a
try.

You should, after applying each of your patches, cd to the guix repo
and run `./etc/indent-code.el gnu/packages/hardware.scm` as described
in `info '(guix)Formatting Code'`.

`./pre-inst-env guix lint libx86emu` warns that your custom check
phase does not consider #:tests?  (i.e. it does not work with the
--without-tests transform option).

As for the package libx86emu proper:

> (define-public libx86emu
>   (package
>     (name "libx86emu")
>     (version "3.1")
>     (source (origin
>               (method git-fetch)
>               (uri (git-reference
>                     (url "https://github.com/wfeldt/libx86emu")
>                     (commit version)))
>               (file-name (git-file-name name version))
>               (sha256
>                (base32
>                 "104xqc6nj9rpi7knl3dfqvasf087hlz2n5yndb1iycw35a6j509b"))))
>     (build-system gnu-build-system)
>     (arguments
>      `(#:make-flags (list
>                      (string-append "CC=" ,(cc-for-target))
>                      (string-append "DESTDIR=" %output)
>                      (string-append "LIBDIR=/lib"))
>        #:phases (modify-phases %standard-phases
>                   (delete 'configure)
>                   (add-after 'unpack 'fix-version-and-usr
>                     (lambda* (#:key inputs #:allow-other-keys)
>                       (delete-file "git2log")

I am not sure, would it be better to delete this git2log in an origin
snippet?  I have not checked.



>                       (let* ((file (open-file "VERSION" "a")))
>                         (display ,version file)
>                         (close-port file))
>                       (substitute* "Makefile"
>                         (("/usr") "/"))))
>                   (replace 'check
>                     (lambda* _
>                       (invoke "make" "test"))))))
>     (native-inputs `(("git" ,git) ("perl" ,perl) ("nasm" ,nasm)))
>     (synopsis "Library for x86 emulation")
>     (description "Small library to emulate x86 instructions.  The focus here
> is not a complete emulation (go for qemu for this) but to cover enough for
> typical firmware blobs.

The start of the sentence is missing before @enumerate.  “You can”?

> @enumerate
> @item intercept any memory access or directly map real memory ranges
> @item intercept any i/o access, map real i/o ports, or block any real i/o
> @item intercept any interrupt
> @item provides hook to run after each instruction

Either “provides a hook” or “provides hooks”.


> @item recognizes a special x86 instruction that can trigger logging

s/recognizes/recognize/


> @item has integrated logging

Maybe s/has/use/


> @end enumerate")
>     (home-page "https://github.com/wfeldt/libx86emu")
>     (license license:bsd-1)))

The license is not bsd-1 but x11-style, I think.

Regards,
Florian

Hi Petr!

I have merged #49934 and #50283.

I'll combine our patches and send a new one. :)

Regards,
RG.

Hi,

I have sent a patch for hw-probe (https://issues.guix.gnu.org/49934#9), 
which requires some refinements.

[1] Some programs in the script are not patched, hence doesn't work in 
pure environment.
[2] Synopsis and description has to be revised.
[3] Finding a way to avoid propagation of hw-info.

@phodina, Would you be able to hack on the above.

Regards,
RG.

Thanks for the patch!

I’m afraid the 'patch build phase is much too repetitive.  You can
greatly simplify it by doing a substitution like this:

--8<---------------cut here---------------start------------->8---
(substitute* "hw-probe.pl"
  (("Cmd\\(\"([^\"]+)" _ command)
   (string-append "Cmd(\"" (which command))))
--8<---------------cut here---------------end--------------->8---

The regular expression has a sub-group that matches everything after the
opening double quote until it hits another double quote.  That group is
captured as “command”.  We then look up that string in the build
environment with “which” (not the executable of the same name), which
returns the absolute file name corresponding to the command.

Hi Ricardo!

> --8<---------------cut here---------------start------------->8---
> (substitute* "hw-probe.pl"
>    (("Cmd\\(\"([^\"]+)" _ command)
>     (string-append "Cmd(\"" (which command))))
> --8<---------------cut here---------------end--------------->8---

Thanks for the snippet and explanation.

I am wondering these,
[1] Does require import of specific modules?
[2] If `(which command)`'s output is empty, will it return #f?

Regards,
RG.

On Sat, Dec 25, 2021 at 04:17:32PM -0500, Raghav Gururajan via Guix-patches via wrote:
> Hi Ricardo!
> 
> > --8<---------------cut here---------------start------------->8---
> > (substitute* "hw-probe.pl"
> >    (("Cmd\\(\"([^\"]+)" _ command)
> >     (string-append "Cmd(\"" (which command))))
> > --8<---------------cut here---------------end--------------->8---
> 
> Thanks for the snippet and explanation.
> 
> I am wondering these,
> [1] Does require import of specific modules?
> [2] If `(which command)`'s output is empty, will it return #f?

For 2, looks like yes

(ins)efraim@3900XT ~$ which -a svn
which: no svn in (/home/efraim/Applications/.bin:/gnu/store/4iyar3w7k30b5j13v9cf5gmiz5livkyj-enlightenment-0.24.2/bin:/run/setuid-programs:/home/efraim/.config/guix/current/bin:/home/efraim/.guix-profile/bin:/home/efraim/.guix-profile/sbin:/run/current-system/profile/bin:/run/current-system/profile/sbin)
(ins)efraim@3900XT ~$ guile
GNU Guile 3.0.7
Copyright (C) 1995-2021 Free Software Foundation, Inc.

Guile comes with ABSOLUTELY NO WARRANTY; for details type `,show w'.
This program is free software, and you are welcome to redistribute it
under certain conditions; type `,show c' for details.

Enter `,help' for help.
(ins)scheme@(guile-user)> (use-modules (guix build utils))
(ins)scheme@(guile-user)> (which "svn")
$1 = #f
(ins)scheme@(guile-user)>

Hi,

Raghav Gururajan <rg@raghavgururajan.name> writes:

> Hi Ricardo!
>
>> --8<---------------cut here---------------start------------->8---
>> (substitute* "hw-probe.pl"
>>    (("Cmd\\(\"([^\"]+)" _ command)
>>     (string-append "Cmd(\"" (which command))))
>> --8<---------------cut here---------------end--------------->8---
>
> Thanks for the snippet and explanation.
>
> I am wondering these,
> [1] Does require import of specific modules?

No.  `which', the Scheme procedure, is part of (guix build utils).

> [2] If `(which command)`'s output is empty, will it return #f?

Yes, and I agree that's a problem.

Fortunately, the recently added search-input-file more usefully raises
an error when the file it is looking for doesn't exist, so I used that.

It was a bit painful to hunt all the commands but in the end it works
beautifully!  I pushed the result as
3611f99affc4d2edcf4141e3022dd0d2a0b04068.

Closing, thanks!

Maxim

Maxim,

> Yes, and I agree that's a problem.
> 
> Fortunately, the recently added search-input-file more usefully raises
> an error when the file it is looking for doesn't exist, so I used that.
> 
> It was a bit painful to hunt all the commands but in the end it works
> beautifully!  I pushed the result as
> 3611f99affc4d2edcf4141e3022dd0d2a0b04068.
> 
> Closing, thanks!

Thank a lot.

Regards,
RG.