From mboxrd@z Thu Jan  1 00:00:00 1970
From: Michael Tremer <michael.tremer@ipfire.org>
To: development@lists.ipfire.org
Subject: Re: [PATCH] Fix for bug 12539
Date: Thu, 10 Dec 2020 14:29:59 +0100
Message-ID: <B5C1DBA7-AFC2-48BE-A239-40B4612C4BD5@ipfire.org>
In-Reply-To:
 <trinity-b8c8795a-7d05-417b-b686-2984683b99c3-1607606038186@3c-app-gmx-bap41>
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="===============9016960703997313449=="
List-Id: <development.lists.ipfire.org>

--===============9016960703997313449==
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable

Hi,

> On 10 Dec 2020, at 14:13, Bernhard Bitsch <Bernhard.Bitsch(a)gmx.de> wrote:
>=20
> Hello,
>=20
>> Gesendet: Donnerstag, 10. Dezember 2020 um 13:48 Uhr
>> Von: "Michael Tremer" <michael.tremer(a)ipfire.org>
>> An: "Adolf Belka" <ahb.ipfire(a)gmail.com>
>> Cc: development(a)lists.ipfire.org
>> Betreff: Re: [PATCH] Fix for bug 12539
>>=20
>> Hello,
>>=20
>> Great work. This looks like how it should be.
>>=20
>>> On 7 Dec 2020, at 15:01, Adolf Belka <ahb.ipfire(a)gmail.com> wrote:
>>>=20
>>> The installer recognises cups and cups-filters both as cups and puts
>>> two instances of cups in the add-on services table.
>>> Based on input from Michael Tremer this patch replaces the command
>>> returning the second element between hyphens with one that takes
>>> what comes after "meta-" using Perl code rather than a shell command.
>>> The second find command was changed as per Michael's suggestion.
>>>=20
>>> Tested in my ipfire test bed system and only results in one cups
>>> entry.
>>> Signed-off-by: Adolf Belka <ahb.ipfire(a)gmail.com>
>>> ---
>>> html/cgi-bin/services.cgi | 9 +++++----
>>> 1 file changed, 5 insertions(+), 4 deletions(-)
>>>=20
>>> diff --git a/html/cgi-bin/services.cgi b/html/cgi-bin/services.cgi
>>> index 26ab4f314..36954ba70 100644
>>> --- a/html/cgi-bin/services.cgi
>>> +++ b/html/cgi-bin/services.cgi
>>> @@ -161,19 +161,20 @@ END
>>> 	my $lines=3D0; # Used to count the outputlines to make different bgcolor
>>>=20
>>> 	# Generate list of installed addon pak's
>>> -	my @pak =3D `find /opt/pakfire/db/installed/meta-* 2>/dev/null | cut -d=
"-" -f2`;
>>> +	opendir (DIR, "/opt/pakfire/db/installed") || die "Cannot opendir /opt/=
pakfire/db/installed/: $!";
>>> +	my @pak =3D sort readdir DIR;
>>> 	foreach (@pak){
>>> 		chomp($_);
>>> +		next unless (m/^meta-/);
>>> +		s/^meta-//;
>>=20
>> Although this is the least intuitive thing to do. I have no idea who desig=
ned Perl. I hope they are proud of all the chaos they have created :)
>>=20
>> -Michael
>=20
> Can't see where the solution is 'least intuitive'. As far as I can see from=
 the patch, now the directory of installed addons is read with Perl functions=
 ( not in a extra forked process ). The definition of filenames is used conse=
quently ( 'meta-<addon name>', possible services are stored in /etc/init.d/<a=
ddon name> ).

To be clear, I am complaining about the Perl language design instead of the c=
ode. In my view something like

  name =3D name.replace(=E2=80=9Cmeta-=E2=80=9C, =E2=80=9C=E2=80=9D)

is a thousand times easier to read.

Best,
-Michael

>=20
> - Bernhard
>=20
>=20
>>=20
>>>=20
>>> 		# Check which of the paks are services
>>> -		my @svc =3D `find /etc/init.d/$_ 2>/dev/null | cut -d"/" -f4`;
>>> -		foreach (@svc){
>>> +		if (-e "/etc/init.d/$_") {
>>> 			# blacklist some packages
>>> 			#
>>> 			# alsa has trouble with the volume saving and was not really stopped
>>> 			# mdadm should not stopped with webif because this could crash the sys=
tem
>>> 			#
>>> -			chomp($_);
>>> 			if ( $_ eq 'squid' ) {
>>> 				next;
>>> 			}
>>> --
>>> 2.29.2


--===============9016960703997313449==--