Processed: Urgencies should all be lower case

[Debian Bug Tracking System]

Processing commands for control@...:

> clone 484359 -1
Bug#484359: [frontend/lintian] Urgencies are case-sensitive
Bug 484359 cloned as bug 484511.

> reassign -1 debian-policy
Bug#484511: [frontend/lintian] Urgencies are case-sensitive
Bug reassigned from package `release.debian.org' to `debian-policy'.

> retitle -1 Urgencies should all be lower case
Bug#484511: [frontend/lintian] Urgencies are case-sensitive
Changed Bug title to `Urgencies should all be lower case' from `[frontend/lintian] Urgencies are case-sensitive'.

> reassign 484359 lintian
Bug#484359: [frontend/lintian] Urgencies are case-sensitive
Bug reassigned from package `release.debian.org' to `lintian'.

> block 484359 by -1
Bug#484511: Urgencies should all be lower case
Bug#484359: [frontend/lintian] Urgencies are case-sensitive
Was not blocked by any bugs.
Blocking bugs of 484359 added: 484511

> thanks
Stopping processing here.

Please contact me if you need assistance.

Debian bug tracking system administrator
(administrator, Debian Bugs database)


--
To UNSUBSCRIBE, email to debian-policy-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Russ Allbery-2]

Sorry, please reply to this message rather than the other.  I got the
wrong bug number and sent this to Lintian instead of to Policy.

Russ Allbery <rra@...> writes:

> After discussion with Joerg Jaspert, I'm of the opinion that dak is the
> canonical place for valid urgencies to be set.
> Thus, sending this back to lintian, cloning to debian-policy, and
> blocking this one by the new debian-policy bug.
>
> -policy readers:
> urgencies are currently set as lower case only; dak will treat any
> unkown urgency (including upper cased ones) as 'low'. Britney also
> doesn't pay attention to non-lower-case urgencies.
>
> 5.6.17 says urgencies are case-*in*sensitive, and should be changed (IMO
> etc)

Given that Policy has said that Urgency is case-insensitive since 3.5.5
(June 2001) when it imported text from the original packaging manual that
said it was case-insensitive, I'm going to need some convincing that this
is really a bug in Policy as opposed to a bug in dak and britney.  Even
putting aside the question of who implemented what first, since I expect
both go back for years, we know that uppercase has been used in practice
and this doesn't seem like a useful place to enforce case.

Wouldn't the easiest solution be to just make dak (and britney) lowercase
the urgency string before working with it?

--
Russ Allbery (rra@...)               <http://www.eyrie.org/~eagle/>



--
To UNSUBSCRIBE, email to debian-bugs-dist-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Luk Claes]

Russ Allbery wrote:

> Sorry, please reply to this message rather than the other.  I got the
> wrong bug number and sent this to Lintian instead of to Policy.
>
> Russ Allbery <rra@...> writes:
>
>> After discussion with Joerg Jaspert, I'm of the opinion that dak is the
>> canonical place for valid urgencies to be set.
>> Thus, sending this back to lintian, cloning to debian-policy, and
>> blocking this one by the new debian-policy bug.
>>
>> -policy readers:
>> urgencies are currently set as lower case only; dak will treat any
>> unkown urgency (including upper cased ones) as 'low'. Britney also
>> doesn't pay attention to non-lower-case urgencies.
>>
>> 5.6.17 says urgencies are case-*in*sensitive, and should be changed (IMO
>> etc)
>
> Given that Policy has said that Urgency is case-insensitive since 3.5.5
> (June 2001) when it imported text from the original packaging manual that
> said it was case-insensitive, I'm going to need some convincing that this
> is really a bug in Policy as opposed to a bug in dak and britney.  Even
> putting aside the question of who implemented what first, since I expect
> both go back for years, we know that uppercase has been used in practice
> and this doesn't seem like a useful place to enforce case.
>
> Wouldn't the easiest solution be to just make dak (and britney) lowercase
> the urgency string before working with it?

I thought policy was documenting practice and not the other way around?

That should be enough to convince you...

Cheers

Luk



--
To UNSUBSCRIBE, email to debian-bugs-dist-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Russ Allbery-2]

Luk Claes <luk@...> writes:

> I thought policy was documenting practice and not the other way around?

Hm, no, that's not my perception of Policy.  New Policy changes should
reflect practice in Debian first unless the problem is significant enough
that Policy needs to pick a side before one wins in practice, but once
something is in Policy, practice needs to conform with Policy and not the
other way around.  Otherwise, what would be the point in having a Policy
document?

Of course, if someone believes Policy is wrong, they can always propose to
change it.  In this case, though, I think Policy's current position is
also the correct technical one to take.

--
Russ Allbery (rra@...)               <http://www.eyrie.org/~eagle/>



--
To UNSUBSCRIBE, email to debian-policy-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Luk Claes]

Russ Allbery wrote:

> Luk Claes <luk@...> writes:
>
>> I thought policy was documenting practice and not the other way around?
>
> Hm, no, that's not my perception of Policy.  New Policy changes should
> reflect practice in Debian first unless the problem is significant enough
> that Policy needs to pick a side before one wins in practice, but once
> something is in Policy, practice needs to conform with Policy and not the
> other way around.  Otherwise, what would be the point in having a Policy
> document?
>
> Of course, if someone believes Policy is wrong, they can always propose to
> change it.  In this case, though, I think Policy's current position is
> also the correct technical one to take.

Good luck with convincing the ones implementing it...

Why should it be case insensitive? It's not case insensitive now and I
do not see any technical reason to change that. It's not that package
names, sections or anything else is case insensitive, is it? Like
explained before, an urgency that is not existing (for instance one with
capitals) is treated as low urgency at least in britney.

Cheers

Luk



--
To UNSUBSCRIBE, email to debian-policy-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Russ Allbery-2]

Luk Claes <luk@...> writes:
> Russ Allbery wrote:

>> Of course, if someone believes Policy is wrong, they can always propose
>> to change it.  In this case, though, I think Policy's current position
>> is also the correct technical one to take.

> Good luck with convincing the ones implementing it...

You make it sound like it's an ASN.1 encoder or something.  If Joerg says
that he absolutely won't change dak, or the release team absolutely won't
change britney, then we certainly need to talk about that, but I think
it's just dak and britney that are affected, britney is mostly the only
place that it matters, and I'd be surprised if it's a hard code change.

> Why should it be case insensitive?

I think I already covered this, but to restate: because it's been
documented as case-insensitive for at least seven years and probably more
than a decade and because there's a long-standing minority practice of
using it in all caps for an urgency of high (possibly going back to the
packaging manual that documented urgencies in all caps while stating that
case was insignificant).

> It's not case insensitive now and I do not see any technical reason to
> change that. It's not that package names, sections or anything else is
> case insensitive, is it?

Those feel like different cases to me.  Package names and sections are
names whose meanings are external to the specification of the relevant
field and represent directories on disk or file names, which are also
case-sensitive.  Urgency is a keyword.  I think keywords are generally
better treated as case-insensitive.  An example that feels similar to me
is control field names.

--
Russ Allbery (rra@...)               <http://www.eyrie.org/~eagle/>



--
To UNSUBSCRIBE, email to debian-policy-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Joerg Jaspert]

On 11407 March 1977, Russ Allbery wrote:
> You make it sound like it's an ASN.1 encoder or something.  If Joerg says
> that he absolutely won't change dak,

I wont change it.

But I might look at patches changing it (or better, bzr trees ready to
merge), if someone really wants it changed.

>> Why should it be case insensitive?
> I think I already covered this, but to restate: because it's been
> documented as case-insensitive for at least seven years and probably more
> than a decade and because there's a long-standing minority practice of
> using it in all caps for an urgency of high (possibly going back to the
> packaging manual that documented urgencies in all caps while stating that
> case was insignificant).

The code in dak, in the current form, is there since 2002-02-13, when
jennifer (today process_unchecked) got added to the repository. Most
probably something similar existed in the code before this.
Its also nearly unchanged since then, with changes being cosmetical.

The code itself also resets all unknown (and upper-case == unknown)
values to low, so whoever thinks HIGH is doing good or helps in noticing
stuff nicely sets himself a low urgency...

>> It's not case insensitive now and I do not see any technical reason to
>> change that. It's not that package names, sections or anything else is
>> case insensitive, is it?
> Those feel like different cases to me.  Package names and sections are
> names whose meanings are external to the specification of the relevant
> field and represent directories on disk or file names, which are also
> case-sensitive.  Urgency is a keyword.  I think keywords are generally
> better treated as case-insensitive.  An example that feels similar to me
> is control field names.

I personally think people should just use lower case, there is no good
reason to have stuff mixed-case or all upper. It doesn't get you
anything.

--
bye, Joerg
#debian.de @ OFTC
(01:38) <michael> hui, hier wird sonntags gechattet :)
(01:39) <maxx> ja, aber nur zwischen 1:35 und 1:45, wenn der Sonntag der 1. im Monat ist :)
(01:39) <Sahneschnitter> wasn hier los? activity :)

[Manoj Srivastava-5]

On Wed, 04 Jun 2008 14:52:04 -0700, Russ Allbery <rra@...> said:

> Luk Claes <luk@...> writes:
>> I thought policy was documenting practice and not the other way
>> around?

> Hm, no, that's not my perception of Policy.  New Policy changes should
> reflect practice in Debian first unless the problem is significant
> enough that Policy needs to pick a side before one wins in practice,
> but once something is in Policy, practice needs to conform with Policy
> and not the other way around.  Otherwise, what would be the point in
> having a Policy document?

        I tend to agree.  Policy is not just documentation, it is trying
 to document standards to which components of the distributions are to
 be held to.

> Of course, if someone believes Policy is wrong, they can always
> propose to change it.  In this case, though, I think Policy's current
> position is also the correct technical one to take.

        AOL.

        manoj
--
Home on the Range was originally written in beef-flat.
Manoj Srivastava <srivasta@...> <http://www.golden-gryphon.com/>  
1024D/BF24424C print 4966 F272 D093 B493 410B  924B 21BA DABB BF24 424C



--
To UNSUBSCRIBE, email to debian-bugs-dist-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Russ Allbery-2]

Joerg Jaspert <joerg@...> writes:
> On 11407 March 1977, Russ Allbery wrote:

>> You make it sound like it's an ASN.1 encoder or something.  If Joerg
>> says that he absolutely won't change dak,
>
> I wont change it.
>
> But I might look at patches changing it (or better, bzr trees ready to
> merge), if someone really wants it changed.

Patch attached.  I can also create a bzr repository if that's helpful.

> The code in dak, in the current form, is there since 2002-02-13, when
> jennifer (today process_unchecked) got added to the repository. Most
> probably something similar existed in the code before this.  Its also
> nearly unchanged since then, with changes being cosmetical.

As you can see from the patch below, it looked like dak was actually
trying to follow Policy here but the code was buggy (the check was done in
the wrong order).

I also added a fix for logging that I'm not sure is necessary but which
can't hurt (if britney uses that log file, this would also fix britney,
but I don't know how that part of the process works).

=== modified file 'dak/process_accepted.py'
--- dak/process_accepted.py 2008-05-05 19:32:00 +0000
+++ dak/process_accepted.py 2008-06-05 17:19:18 +0000
@@ -81,6 +81,7 @@
 
     def log (self, source, version, urgency):
         "Log an event"
+        urgency = urgency.lower()
         self.log_file.write(" ".join([source, version, urgency])+'\n')
         self.log_file.flush()
         self.writes += 1

=== modified file 'dak/process_unchecked.py'
--- dak/process_unchecked.py 2008-05-05 19:32:00 +0000
+++ dak/process_unchecked.py 2008-06-05 17:15:43 +0000
@@ -906,10 +906,10 @@
     if changes["architecture"].has_key("source"):
         if not changes.has_key("urgency"):
             changes["urgency"] = Cnf["Urgency::Default"]
+        changes["urgency"] = changes["urgency"].lower()
         if changes["urgency"] not in Cnf.ValueList("Urgency::Valid"):
             reject("%s is not a valid urgency; it will be treated as %s by testing." % (changes["urgency"], Cnf["Urgency::Default"]), "Warning: ")
             changes["urgency"] = Cnf["Urgency::Default"]
-        changes["urgency"] = changes["urgency"].lower()
 
 ################################################################################
 
--
Russ Allbery (rra@...)               <http://www.eyrie.org/~eagle/>



--
To UNSUBSCRIBE, email to debian-policy-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Joey Hess]

Joerg Jaspert wrote:
> The code in dak, in the current form, is there since 2002-02-13, when
> jennifer (today process_unchecked) got added to the repository. Most
> probably something similar existed in the code before this.
> Its also nearly unchanged since then, with changes being cosmetical.

Nice theory, but I have made many uploads with urgency=HIGH between 2001
and 2005 and did not receive any mails about an unknown urgency for
those, but only for my last upload recently.

--
see shy jo

[Joerg Jaspert]

On 11408 March 1977, Joey Hess wrote:

>> The code in dak, in the current form, is there since 2002-02-13, when
>> jennifer (today process_unchecked) got added to the repository. Most
>> probably something similar existed in the code before this.
>> Its also nearly unchanged since then, with changes being cosmetical.
> Nice theory, but I have made many uploads with urgency=HIGH between 2001
> and 2005 and did not receive any mails about an unknown urgency for
> those, but only for my last upload recently.

No theory, the code and the revision history is pretty clear on it.

--
bye, Joerg
[...]
While Debian is certainly about beer, and in some cases may even be
about free beer, Debian is mainly about free speech.



--
To UNSUBSCRIBE, email to debian-bugs-dist-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Joerg Jaspert]

>> But I might look at patches changing it (or better, bzr trees ready to
>> merge), if someone really wants it changed.
> Patch attached.  I can also create a bzr repository if that's helpful.

For the future - yes please. Including a changelog entry.

> I also added a fix for logging that I'm not sure is necessary but which
> can't hurt (if britney uses that log file, this would also fix britney,
> but I don't know how that part of the process works).

Britney uses it, but its not needed at that point, as p_a will read the
stuff out of the .dak file written earlier by p_u.


--
bye, Joerg
<Aquariophile> geht nur in IE
<Aquariophile> unter win
<Aquariophile> autsch ich glaub das war ein eigentor

[Russ Allbery-2]

Joerg Jaspert <joerg@...> writes:

>> Patch attached.  I can also create a bzr repository if that's helpful.
>
> For the future - yes please. Including a changelog entry.

Will do in the future.

> Britney uses it, but its not needed at that point, as p_a will read the
> stuff out of the .dak file written earlier by p_u.

With this patch applied, I think that these bugs are now moot, but I
wanted to check before closing them.  Is there any further work required
in britney to treat urgencies as case-insensitive, or does the change in
the log file written by dak solve that problem as well?

--
Russ Allbery (rra@...)               <http://www.eyrie.org/~eagle/>



--
To UNSUBSCRIBE, email to debian-policy-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Adam D. Barratt-29]

On Sat, 2008-06-07 at 22:40 -0700, Russ Allbery wrote:
[...]
> With this patch applied, I think that these bugs are now moot, but I
> wanted to check before closing them.  Is there any further work required
> in britney to treat urgencies as case-insensitive, or does the change in
> the log file written by dak solve that problem as well?

As I understand it, the change in dak means that the first time britney
is aware of the packages urgency it will already have been lower-cased.
Whilst britney still handles urgencies internally in a case-sensitive
manner, it should never be receiving an initial urgency for a package
which isn't in lower case.

Assuming that I haven't missed anything above then I personally have no
objections to closing the bugs.

Adam



--
To UNSUBSCRIBE, email to debian-policy-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Neil McGovern-2]

On Sat, Jun 21, 2008 at 08:08:45PM +0100, Adam D. Barratt wrote:

> On Sat, 2008-06-07 at 22:40 -0700, Russ Allbery wrote:
> [...]
> > With this patch applied, I think that these bugs are now moot, but I
> > wanted to check before closing them.  Is there any further work required
> > in britney to treat urgencies as case-insensitive, or does the change in
> > the log file written by dak solve that problem as well?
>
> As I understand it, the change in dak means that the first time britney
> is aware of the packages urgency it will already have been lower-cased.
> Whilst britney still handles urgencies internally in a case-sensitive
> manner, it should never be receiving an initial urgency for a package
> which isn't in lower case.
>

Agreed from my PoV.

Neil
--
A. Because it breaks the logical sequence of discussion
Q. Why is top posting bad?
gpg key - http://www.halon.org.uk/pubkey.txt ; the.earth.li B345BDD3



--
To UNSUBSCRIBE, email to debian-policy-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...