gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings

classic Classic list List threaded Threaded
31 messages Options
12
Reply | Threaded
Open this post in threaded view
|

gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings

Sergei Trofimovich-6
Hello World!

TL;DR:
  This evening I plan to mangle ~3000 ebuilds in the main tree
  by dropping trailing '.' in all 'DESCRIPTION=' fields (except "etc." case)

Long story:

As you may know newest portage release 2.2.11
got a minor (but chatty) QA warning:
    DESCRIPTION ends with a '.' character

    https://github.com/gentoo/portage/commit/06637c4215d55c57517739214c6e0fd6f8f53914
    https://bugs.gentoo.org/438976

About ~3000 ebuilds are affected, thus I've sketched a mangling script:

    https://github.com/trofi/gentoo-qa/blob/master/check_description.sh

It is nice to use to convert all your large overlays you sync to gx86, etc.

The script does not handle case of multiline description:
    DESCRIPTION="You have to
    clean that yourself."

I'll mangle/commit one package at a time. Hope not to interfere with your
workflow much.

If you have any objections/thoughts please do say so.

Thanks!

--

  Sergei

signature.asc (188 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings

Bertrand Jacquin-5
Hi,

On 2014-08-10 14:22, Sergei Trofimovich wrote:

> The script does not handle case of multiline description:
>     DESCRIPTION="You have to
>     clean that yourself."

You could handle this by reading metadata/md5-cache/*/* instead of
ebuild itself

But is multiline DESCRIPTION something recommended as it should contain
a short description ?

--
Beber

Reply | Threaded
Open this post in threaded view
|

Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings

William Hubbs
In reply to this post by Sergei Trofimovich-6
On Sun, Aug 10, 2014 at 03:22:11PM +0300, Sergei Trofimovich wrote:

> Hello World!
>
> TL;DR:
>   This evening I plan to mangle ~3000 ebuilds in the main tree
>   by dropping trailing '.' in all 'DESCRIPTION=' fields (except "etc." case)
>
> Long story:
>
> As you may know newest portage release 2.2.11
> got a minor (but chatty) QA warning:
>     DESCRIPTION ends with a '.' character
Why is this a QA warning in the first place?

I don't recall a policy mandating that descriptions can't end with '.'. I
asked our QA lead about it and was told that he didn't recall that we
have an official policy about it either. Also, the devmanual never
mentions any such requirement.

If someone can point me to something I'm missing, let me know.
Otherwise, I think the warning should be removed.

William

signature.asc (188 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings

Manuel Rüger
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

On 08/12/2014 03:48 AM, William Hubbs wrote:

> On Sun, Aug 10, 2014 at 03:22:11PM +0300, Sergei Trofimovich
> wrote:
>> Hello World!
>>
>> TL;DR: This evening I plan to mangle ~3000 ebuilds in the main
>> tree by dropping trailing '.' in all 'DESCRIPTION=' fields
>> (except "etc." case)
>>
>> Long story:
>>
>> As you may know newest portage release 2.2.11 got a minor (but
>> chatty) QA warning: DESCRIPTION ends with a '.' character
>
> Why is this a QA warning in the first place?
>
> I don't recall a policy mandating that descriptions can't end with
> '.'. I asked our QA lead about it and was told that he didn't
> recall that we have an official policy about it either. Also, the
> devmanual never mentions any such requirement.
>
> If someone can point me to something I'm missing, let me know.
> Otherwise, I think the warning should be removed.
>
> William
>

These links might be helpful:

http://git.overlays.gentoo.org/gitweb/?p=proj/portage.git;a=commit;h=06637c4215d55c57517739214c6e0fd6f8f53914

https://bugs.gentoo.org/show_bug.cgi?id=438976

http://comments.gmane.org/gmane.linux.gentoo.devel/80786


What's still missing is a patch for devmanual (if we still really want
to enforce this).


Manuel
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0
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=bmFg
-----END PGP SIGNATURE-----

Reply | Threaded
Open this post in threaded view
|

Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings

William Hubbs
On Tue, Aug 12, 2014 at 03:59:30AM +0200, Manuel Rüger wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA512
 
 *snip*

> These links might be helpful:
>
> http://git.overlays.gentoo.org/gitweb/?p=proj/portage.git;a=commit;h=06637c4215d55c57517739214c6e0fd6f8f53914
>
> https://bugs.gentoo.org/show_bug.cgi?id=438976
>
> http://comments.gmane.org/gmane.linux.gentoo.devel/80786
>
>
> What's still missing is a patch for devmanual (if we still really want
> to enforce this).
I read the thread, and there was no concensus about making this a
repoman check. Some people thought it was a good idea, but there was a
feeling that this sort of thing is trivial and shouldn't be worried
about.

William

signature.asc (188 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings

Tyler Pohl
how to i get off these mailing lists?


On Mon, Aug 11, 2014 at 7:42 PM, William Hubbs <[hidden email]> wrote:
On Tue, Aug 12, 2014 at 03:59:30AM +0200, Manuel Rüger wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA512

 *snip*

> These links might be helpful:
>
> http://git.overlays.gentoo.org/gitweb/?p=proj/portage.git;a=commit;h=06637c4215d55c57517739214c6e0fd6f8f53914
>
> https://bugs.gentoo.org/show_bug.cgi?id=438976
>
> http://comments.gmane.org/gmane.linux.gentoo.devel/80786
>
>
> What's still missing is a patch for devmanual (if we still really want
> to enforce this).

I read the thread, and there was no concensus about making this a
repoman check. Some people thought it was a good idea, but there was a
feeling that this sort of thing is trivial and shouldn't be worried
about.

William



--
Tyler Pohl
Reply | Threaded
Open this post in threaded view
|

Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings

Duncan-42
[Mailed direct and to list.]

Tyler Pohl posted on Mon, 11 Aug 2014 21:20:17 -0700 as excerpted:

> how to i get off these mailing lists?

Follow the instructions, as found in the headers of every mail on the
list including the one you replied to, or the ones on the site you
presumably signed up from?  Seriously:

Headers:

> List-Help: <mailto:[hidden email]>
> List-Unsubscribe: <mailto:[hidden email]>
> List-Subscribe: <mailto:[hidden email]>
> List-Id: Gentoo Linux mail <gentoo-dev.gentoo.org>


Page at http://www.gentoo.org/main/en/lists.xml, clearly linked via the
lists link on the gentoo homepage:

<quote>

To unsubscribe from a list, send an empty email to:

[hidden email]

Note: You must use the identical address that you subscribed with to
unsubscribe successfully. If your email address is now forwarded/
rewritten beyond your control, please contact the list owner via listname
+[hidden email] with a request for manual removal.

You will then recieve a unsubscription confirmation request (double opt-
in) from the list manager, that you must reply to if you wish to be
unsubscribed.

</quote>

--
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman


Reply | Threaded
Open this post in threaded view
|

Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings

Alex Xu
On 12/08/14 01:29 AM, Duncan wrote:
> Follow the instructions, as found in the headers of every mail on the
> list including the one you replied to, or the ones on the site you
> presumably signed up from?  Seriously:

s/presumably //, this list is closed-loop.


signature.asc (836 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings

hasufell
In reply to this post by William Hubbs
William Hubbs:

> On Tue, Aug 12, 2014 at 03:59:30AM +0200, Manuel Rüger wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA512
>  
>  *snip*
>
>> These links might be helpful:
>>
>> http://git.overlays.gentoo.org/gitweb/?p=proj/portage.git;a=commit;h=06637c4215d55c57517739214c6e0fd6f8f53914
>>
>> https://bugs.gentoo.org/show_bug.cgi?id=438976
>>
>> http://comments.gmane.org/gmane.linux.gentoo.devel/80786
>>
>>
>> What's still missing is a patch for devmanual (if we still really want
>> to enforce this).
>
> I read the thread, and there was no concensus about making this a
> repoman check. Some people thought it was a good idea, but there was a
> feeling that this sort of thing is trivial and shouldn't be worried
> about.
>

That thread is pretty odd.

First, a sentence does not need to have a predicate. I know that for 99%
sure in german and the english wikipedia article seems to suggest the
same. Correct me if I am wrong.

Second, there are valid descriptions that are full ordinary sentences
without referencing ${PN}:
"Access a working SSH implementation by means of a library".

In addition, repoman doesn't check for full sentences that reference
${PN}, such as:
"Portage is the package management and distribution system for Gentoo".

So we have another (useless) repoman warning with false positives.

Reply | Threaded
Open this post in threaded view
|

Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings

Rich Freeman
On Tue, Aug 12, 2014 at 8:47 AM, hasufell <[hidden email]> wrote:
>
> First, a sentence does not need to have a predicate. I know that for 99%
> sure in german and the english wikipedia article seems to suggest the
> same. Correct me if I am wrong.
>

In English your typical English class would teach that every sentence
must have a predicate.  From what Google tells me it technically isn't
entirely true, but every sentence generally does contain a verb.  So,
"library that implements SSL" is not a sentence under any
circumstances.

> Second, there are valid descriptions that are full ordinary sentences
> without referencing ${PN}:
> "Access a working SSH implementation by means of a library".
>
> In addition, repoman doesn't check for full sentences that reference
> ${PN}, such as:
> "Portage is the package management and distribution system for Gentoo".
>
> So we have another (useless) repoman warning with false positives.
>

Yeah, at best this seems a bit trivial.  Do we have a policy that
descriptions aren't allowed to be complete sentences?  Many of our
developers are not native English speakers in the first place, so
striving for grammatical perfection is a bit optimistic.  On top of
that, repoman certainly isn't a native English speaker, so expecting
it to achieve grammatical perfection is a really tall order.  And
please don't suggest making languagetool a dependency for portage...

I don't have a problem with QA recommending new tree policies, but if
they're going to do this the QA team ought to first ensure that the
team agrees (however they want to govern that), and then communicate
the policy before implementing it.  I'd also implement it in
documentation before doing so in repoman, otherwise we're going to
have a repoman full of 800 rules whose origin is a mystery.  I'm fine
with QA policies going into effect by default, but communicating them
allows objections to be raised and an appeal made to Council if
necessary before we get too far along.  This isn't just about due
process - it is hard for developers to even comply with a policy they
are unaware of.

Rich

Reply | Threaded
Open this post in threaded view
|

Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings

Ian Stakenvicius-2
In reply to this post by hasufell
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 12/08/14 08:47 AM, hasufell wrote:

> William Hubbs:
>> On Tue, Aug 12, 2014 at 03:59:30AM +0200, Manuel Rüger wrote:
>>> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512
>>
>> *snip*
>>
>>> These links might be helpful:
>>>
>>> http://git.overlays.gentoo.org/gitweb/?p=proj/portage.git;a=commit;h=06637c4215d55c57517739214c6e0fd6f8f53914
>>>
>>>
>>>
https://bugs.gentoo.org/show_bug.cgi?id=438976

>>>
>>> http://comments.gmane.org/gmane.linux.gentoo.devel/80786
>>>
>>>
>>> What's still missing is a patch for devmanual (if we still
>>> really want to enforce this).
>>
>> I read the thread, and there was no concensus about making this
>> a repoman check. Some people thought it was a good idea, but
>> there was a feeling that this sort of thing is trivial and
>> shouldn't be worried about.
>>
>
> That thread is pretty odd.
>
> First, a sentence does not need to have a predicate. I know that
> for 99% sure in german and the english wikipedia article seems to
> suggest the same. Correct me if I am wrong.
>
> Second, there are valid descriptions that are full ordinary
> sentences without referencing ${PN}: "Access a working SSH
> implementation by means of a library".
>
> In addition, repoman doesn't check for full sentences that
> reference ${PN}, such as: "Portage is the package management and
> distribution system for Gentoo".
>
> So we have another (useless) repoman warning with false positives.
>

TL;DR -- is there any technical reason as to why a DESCRIPTION ending
in '.' is bad?  Other than the fact that it adds 3000 unnecessary
bytes to the portage tree?  IE, does it have the possibility of
throwing off tools that strictly adhere to some random spec (although
it doesn't seem like PMS declares anything bad about this either)??

Perhaps we need to have a less-important repoman warning level
(something that can be quieted with a flag) for things like this?  In
terms of DESCRIPTION consistency I don't see it being a bad thing that
we have the warning, but i also don't see a point in changing the
entire tree to get rid of 3000 bytes, esp. since the ChangeLog entries
added to the tree will add at least 30,000 bytes :)





-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iF4EAREIAAYFAlPqHIYACgkQ2ugaI38ACPCvXQD7BQYtciffNZDCM03vMSlNAgQh
s4j3dw3tL9VDe/oiq7kA/25lVdaRqAc/mbdiI5surUOG9a0J+1sk/nrVft4ocnSs
=8273
-----END PGP SIGNATURE-----

Reply | Threaded
Open this post in threaded view
|

repoman --nonag (was Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings )

Ian Stakenvicius-2
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 12/08/14 09:54 AM, Ian Stakenvicius wrote:
>
> Perhaps we need to have a less-important repoman warning level
> (something that can be quieted with a flag) for things like this?
> In terms of DESCRIPTION consistency I don't see it being a bad
> thing that we have the warning, but i also don't see a point in
> changing the entire tree to get rid of 3000 bytes, esp. since the
> ChangeLog entries added to the tree will add at least 30,000 bytes
> :)
>

I'm wondering what everyone thinks of having a --nonag option to
repoman and shoving some of the more trivial/style-related repoman
'warnings' into a 'nag' level warning?  IIRC at least one of the QA
team members is so tired of the warnings that they want to make every
single one of them errors; the --nonag option would allow those
warnings to remain in repoman (ie to help guide new dev's or non-dev's
using repoman on their local repos) but since they don't relate to
actual technical breakage they can just be turned off during QA runs, etc.

Thoughts?

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iF4EAREIAAYFAlPqHwoACgkQ2ugaI38ACPAVvgEAqNY3pl+QartxGxiTnEPuycl3
4za+QK26KuNUGO0RJewA/0EIV6z92TG3hr+eLDViIJxfdB0GVTl6JV6ha/EQUZcY
=49jq
-----END PGP SIGNATURE-----

Reply | Threaded
Open this post in threaded view
|

Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings

hasufell
In reply to this post by Rich Freeman
Rich Freeman:
> so striving for grammatical perfection is a bit optimistic.

In that case, we should just rm the repoman warning and stop discussing
this matter.

Reply | Threaded
Open this post in threaded view
|

Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings

William Hubbs
In reply to this post by Rich Freeman
On Tue, Aug 12, 2014 at 09:26:07AM -0400, Rich Freeman wrote:
 
 *snip*

> Yeah, at best this seems a bit trivial.  Do we have a policy that
> descriptions aren't allowed to be complete sentences?  Many of our
> developers are not native English speakers in the first place, so
> striving for grammatical perfection is a bit optimistic.  On top of
> that, repoman certainly isn't a native English speaker, so expecting
> it to achieve grammatical perfection is a really tall order.  And
> please don't suggest making languagetool a dependency for portage...
 
No, we do not have, and there has been no request for, a qa policy that
requires description to not end with a '.'. Also, it is not documented
in the devmanual. So, it appears that this warning was put in place
without involving the QA team at all.

William

signature.asc (188 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: repoman --nonag (was Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings )

Rich Freeman
In reply to this post by Ian Stakenvicius-2
On Tue, Aug 12, 2014 at 10:04 AM, Ian Stakenvicius <[hidden email]> wrote:

>
> I'm wondering what everyone thinks of having a --nonag option to
> repoman and shoving some of the more trivial/style-related repoman
> 'warnings' into a 'nag' level warning?  IIRC at least one of the QA
> team members is so tired of the warnings that they want to make every
> single one of them errors; the --nonag option would allow those
> warnings to remain in repoman (ie to help guide new dev's or non-dev's
> using repoman on their local repos) but since they don't relate to
> actual technical breakage they can just be turned off during QA runs, etc.
>

What, specifically, are we considering trivial?

The whole point of repoman is to prevent devs from making mistakes.
Being able to turn off warnings is counterproductive.  Eliminating
warnings that don't need to be warnings is of course fine.

There is no value in having an escalating battle between warnings and
options to suppress them.

Rich

Reply | Threaded
Open this post in threaded view
|

Re: repoman --nonag (was Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings )

Ian Stakenvicius-2
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 12/08/14 12:36 PM, Rich Freeman wrote:

> On Tue, Aug 12, 2014 at 10:04 AM, Ian Stakenvicius <[hidden email]>
> wrote:
>>
>> I'm wondering what everyone thinks of having a --nonag option to
>> repoman and shoving some of the more trivial/style-related
>> repoman 'warnings' into a 'nag' level warning?  IIRC at least one
>> of the QA team members is so tired of the warnings that they want
>> to make every single one of them errors; the --nonag option would
>> allow those warnings to remain in repoman (ie to help guide new
>> dev's or non-dev's using repoman on their local repos) but since
>> they don't relate to actual technical breakage they can just be
>> turned off during QA runs, etc.
>>
>
> What, specifically, are we considering trivial?
>
> The whole point of repoman is to prevent devs from making
> mistakes. Being able to turn off warnings is counterproductive.
> Eliminating warnings that don't need to be warnings is of course
> fine.
>
> There is no value in having an escalating battle between warnings
> and options to suppress them.
>
> Rich
>

Well, there's warnings related to style, like
DESCRIPTION-ending-in-period, and then there's warnings relating to
technical or functional issues.  Of the second set, there are fatal
ones and then there are ones that aren't fatal but still important
(DEPENDENCY.badindev comes to mind).  I think the style or other
non-functional warnings (i can't actually think of any that aren't
style related, tbh) are great to have, and perhaps should even be
expanded if someone felt so inclined, but not at the expense of
additional noise all the time for groups like QA that are primarily
concerned about maintaining functionality.  So instead of, for
instance, dropping the DESCRIPTION-ending-in-period check, it could
instead be relegated to a "nag" that could be hidden with --nonag.

Essentially what it boils down to is that I don't see every non-fatal
warning as being equivalent in importance, and it might make sense to
push the ones that could be construed as recommendations rather than
warnings to a lighter level.

If there isn't any support for this idea, then of course let's skip it
and we can drop the check(s) instead if that's what's desired by the
community.  Then it's just a question of how far we might want to go
in terms of dropping checks.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iF4EAREIAAYFAlPqR28ACgkQ2ugaI38ACPCQfQEAgs9Zbpw9rkXjZpJUrM6s0/LZ
mGm1UeLe0iNN0zKn8JwBAJZ2NL1tEDA+8X15UHsT4mBTevK+I3cv9+l6R7j6AtGq
=ptmP
-----END PGP SIGNATURE-----

Reply | Threaded
Open this post in threaded view
|

Re: repoman --nonag (was Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings )

hasufell
Ian Stakenvicius:
> So instead of, for
> instance, dropping the DESCRIPTION-ending-in-period check, it could
> instead be relegated to a "nag" that could be hidden with --nonag.

It will still be broken, even if you hide it.

Reply | Threaded
Open this post in threaded view
|

Re: repoman --nonag (was Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings )

Ian Stakenvicius-2
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 12/08/14 01:08 PM, hasufell wrote:
> Ian Stakenvicius:
>> So instead of, for instance, dropping the
>> DESCRIPTION-ending-in-period check, it could instead be relegated
>> to a "nag" that could be hidden with --nonag.
>
> It will still be broken, even if you hide it.
>

Say it's fixed so it doesn't do false-positives anymore, etc. etc.

I don't consider a recommended style message to be 'broken' just
because it's not listed in the devmanual/PMS/etc as a requirement.
The implementation of it, on the other hand, yes that could be broken
and in this case should be fixed if we keep the check around.


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iF4EAREIAAYFAlPqSyYACgkQ2ugaI38ACPAMZAD/QMy3mmz9yL9kKLfcrNlf737X
9+iJjspqMrp/h8PV19oA/3fQExM/yGUBinM5CWFx6lvYz1pL2daeyxUgMRxtcxDB
=ki6s
-----END PGP SIGNATURE-----

Reply | Threaded
Open this post in threaded view
|

Re: repoman --nonag (was Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings )

Rich Freeman
On Tue, Aug 12, 2014 at 1:13 PM, Ian Stakenvicius <[hidden email]> wrote:
>
> I don't consider a recommended style message to be 'broken' just
> because it's not listed in the devmanual/PMS/etc as a requirement.
> The implementation of it, on the other hand, yes that could be broken
> and in this case should be fixed if we keep the check around.
>

If we are bothered enough by something to have repoman check it, we
can be bothered enough to add it to the devmanual.

I think we need to decide whether we care about periods at the ends of
DESCRIPTIONs.  If we do, then it should be a warning and devs should
fix their ebuilds at the next convenient opportunity.  If we don't,
then let's just drop the warning.

I'm fine with the separation of hard/soft errors, because some issues
could be situational and left to developer discretion.  However, we
wouldn't want to hide those, because if a dev introduces a new issue
we don't want them to not see the warning.

If somebody has a whitespace issue they should get a warning.  They
should be doing a scan before commit, and they should generally take
the time to fix the issue, even though it is just style.  What is the
point in having a style guideline if half of us are just going to
ignore warnings related to it.  That doesn't mean that our style
guidelines have to be over-the-top - the solution to that is to drop
requirements that aren't important, not to hide them.

If somebody wants to come up with a bunch of extra optional repoman
warnings for stuff like style, I think their time would be better
spent coming up with an ebuild pretty-printer or something which just
fixes things instead of whining about things that aren't policy.

Ultimately quality has to be something we invest in for each other's
sake.  If a rule isn't really benefiting anybody, then it doesn't
belong.  Within reason good style helps us all out - bash doesn't care
if the whole ebuild fits on one line with all the phases/variables/etc
in semi-random order, but we impose some sane style so that we can
work in the tree and not rip our eyes out.

--
Rich

Reply | Threaded
Open this post in threaded view
|

Re: repoman --nonag (was Re: gentoo-x86 tree cleanup for 'DESCRIPTION ends with a '.' character' warnings )

William Hubbs
On Tue, Aug 12, 2014 at 01:25:44PM -0400, Rich Freeman wrote:

> On Tue, Aug 12, 2014 at 1:13 PM, Ian Stakenvicius <[hidden email]> wrote:
> >
> > I don't consider a recommended style message to be 'broken' just
> > because it's not listed in the devmanual/PMS/etc as a requirement.
> > The implementation of it, on the other hand, yes that could be broken
> > and in this case should be fixed if we keep the check around.
> >
>
> If we are bothered enough by something to have repoman check it, we
> can be bothered enough to add it to the devmanual.
I also think "that something" should be added to the devmanual before it
is added to repoman so that developers aren't blind-sided by repoman
warnings like this.


> I think we need to decide whether we care about periods at the ends of
> DESCRIPTIONs.  If we do, then it should be a warning and devs should
> fix their ebuilds at the next convenient opportunity.  If we don't,
> then let's just drop the warning.
 
I think some will have periods and some won't depending on how the
description is written, so this warning is not one that should stay.

William


signature.asc (188 bytes) Download Attachment
12