Oddity in RNG generation with new @cert datatype

newer
Re: [tei-council] Oddity in RNG...

older
Peter's head in a lamp and other...

Martin Holmes

2 Nov 2015 2 Nov '15

3:19 p.m.

Hi all, The @cert in att.global.responsibility now has a janus value (att.certainty or att.probability). However, when I process one of my existing ODDs using the current dev tree, I get an empty value in my RNG: <define name="att.global.responsibility.attributes"> <ref name="att.global.responsibility.attribute.cert"/> <ref name="att.global.responsibility.attribute.resp"/> </define> <define name="att.global.responsibility.attribute.cert"> <optional> <attribute name="cert"> http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) signifies the degree of certainty associated with the intervention or interpretation. <empty/> </attribute> </optional> </define> <define name="att.global.responsibility.attribute.resp"> <optional> <attribute name="resp"> http://relaxng.org/ns/compatibility/annotations/1.0">(responsible party) indicates the agency responsible for the intervention or interpretation, for example an editor or transcriber. <list> <oneOrMore> <ref name="data.pointer"/> </oneOrMore> </list> </attribute> </optional> </define> Any idea why this might be happening? Cheers, Martin

Show replies by date

Raffaele Viglianti

3 Nov 3 Nov

7:13 a.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

Hi Martin, When I first made that change, I mistakenly pointed the ref to att.certainty and att.probability instead of data.certainty and data.probability. I fixed this after jenkins complained, but I imagine that the rng still got created and it might have looked like the one you included in your email. Is it possible that you're looking at an older RNG and/or need to rebuild? Also I've been trying to build locally to replicate your problem, but when I `make` in P5 I get this error: /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: stylesheet /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl doesn't exist. Have I missed a step? Raff On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes wrote:

...

Hi all,

The @cert in att.global.responsibility now has a janus value (att.certainty or att.probability). However, when I process one of my existing ODDs using the current dev tree, I get an empty value in my RNG:

<define name="att.global.responsibility.attributes"> <ref name="att.global.responsibility.attribute.cert"/> <ref name="att.global.responsibility.attribute.resp"/> </define> <define name="att.global.responsibility.attribute.cert"> <optional> <attribute name="cert"> http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) signifies the degree of certainty associated with the intervention or interpretation. <empty/> </attribute> </optional> </define> <define name="att.global.responsibility.attribute.resp"> <optional> <attribute name="resp"> http://relaxng.org/ns/compatibility/annotations/1.0">(responsible party) indicates the agency responsible for the intervention or interpretation, for example an editor or transcriber. <list> <oneOrMore> <ref name="data.pointer"/> </oneOrMore> </list> </attribute> </optional> </define>

Any idea why this might be happening?

Cheers, Martin -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

Martin Holmes

7:56 a.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

Hi Raff, No, this is freshly-generated from an existing ODD, using the latest bleeding-edge Oxygen plugin, meaning that it's the current stylesheets and the current P5 from dev. To build P5, you usually have to point it at the location of your stylesheet checkout: XSL=/home/rviglian/Projects/tei-code/TEI/Stylesheets or whatever. Cheers, Martin On 15-11-03 07:13 AM, Raffaele Viglianti wrote:

...

Hi Martin,

When I first made that change, I mistakenly pointed the ref to att.certainty and att.probability instead of data.certainty and data.probability. I fixed this after jenkins complained, but I imagine that the rng still got created and it might have looked like the one you included in your email.

Is it possible that you're looking at an older RNG and/or need to rebuild?

Also I've been trying to build locally to replicate your problem, but when I `make` in P5 I get this error: /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: stylesheet /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl doesn't exist. Have I missed a step?

Raff

On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes wrote:

...
Hi all,

The @cert in att.global.responsibility now has a janus value (att.certainty or att.probability). However, when I process one of my existing ODDs using the current dev tree, I get an empty value in my RNG:

<define name="att.global.responsibility.attributes"> <ref name="att.global.responsibility.attribute.cert"/> <ref name="att.global.responsibility.attribute.resp"/> </define> <define name="att.global.responsibility.attribute.cert"> <optional> <attribute name="cert"> http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) signifies the degree of certainty associated with the intervention or interpretation. <empty/> </attribute> </optional> </define> <define name="att.global.responsibility.attribute.resp"> <optional> <attribute name="resp"> http://relaxng.org/ns/compatibility/annotations/1.0">(responsible party) indicates the agency responsible for the intervention or interpretation, for example an editor or transcriber. <list> <oneOrMore> <ref name="data.pointer"/> </oneOrMore> </list> </attribute> </optional> </define>

Any idea why this might be happening?

Cheers, Martin -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

Martin Holmes

8:49 a.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

I'm very puzzled by this. Could someone else generate an RNG from this ODD: http://web.uvic.ca/~mholmes/mariage.odd using the dev P5 and Stylesheets, and see if they get the same empty content model for @cert? Can anyone suggest why this might be happening? It doesn't happen in the regular P5 build using teiall.xml in Test. Cheers, Martin On 15-11-03 07:56 AM, Martin Holmes wrote:

...

Hi Raff,

No, this is freshly-generated from an existing ODD, using the latest bleeding-edge Oxygen plugin, meaning that it's the current stylesheets and the current P5 from dev.

To build P5, you usually have to point it at the location of your stylesheet checkout:

XSL=/home/rviglian/Projects/tei-code/TEI/Stylesheets

or whatever.

Cheers, Martin

On 15-11-03 07:13 AM, Raffaele Viglianti wrote:

...
Hi Martin,

When I first made that change, I mistakenly pointed the ref to att.certainty and att.probability instead of data.certainty and data.probability. I fixed this after jenkins complained, but I imagine that the rng still got created and it might have looked like the one you included in your email.

Is it possible that you're looking at an older RNG and/or need to rebuild?

Also I've been trying to build locally to replicate your problem, but when I `make` in P5 I get this error: /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: stylesheet /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl doesn't exist. Have I missed a step?

Raff

On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes wrote:

...
Hi all,

The @cert in att.global.responsibility now has a janus value (att.certainty or att.probability). However, when I process one of my existing ODDs using the current dev tree, I get an empty value in my RNG:

<define name="att.global.responsibility.attributes"> <ref name="att.global.responsibility.attribute.cert"/> <ref name="att.global.responsibility.attribute.resp"/> </define> <define name="att.global.responsibility.attribute.cert"> <optional> <attribute name="cert"> http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) signifies the degree of certainty associated with the intervention or interpretation. <empty/> </attribute> </optional> </define> <define name="att.global.responsibility.attribute.resp"> <optional> <attribute name="resp"> http://relaxng.org/ns/compatibility/annotations/1.0">(responsible party) indicates the agency responsible for the intervention or interpretation, for example an editor or transcriber. <list> <oneOrMore> <ref name="data.pointer"/> </oneOrMore> </list> </attribute> </optional> </define>

Any idea why this might be happening?

Cheers, Martin -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

Hugh Cayless

9:07 a.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

Yeah, I see the same thing... Sent from my phone.

...

On Nov 3, 2015, at 11:49, Martin Holmes wrote:

I'm very puzzled by this. Could someone else generate an RNG from this ODD:

http://web.uvic.ca/~mholmes/mariage.odd

using the dev P5 and Stylesheets, and see if they get the same empty content model for @cert?

Can anyone suggest why this might be happening? It doesn't happen in the regular P5 build using teiall.xml in Test.

Cheers, Martin

...
On 15-11-03 07:56 AM, Martin Holmes wrote: Hi Raff,

No, this is freshly-generated from an existing ODD, using the latest bleeding-edge Oxygen plugin, meaning that it's the current stylesheets and the current P5 from dev.

To build P5, you usually have to point it at the location of your stylesheet checkout:

XSL=/home/rviglian/Projects/tei-code/TEI/Stylesheets

or whatever.

Cheers, Martin

...
On 15-11-03 07:13 AM, Raffaele Viglianti wrote: Hi Martin,

When I first made that change, I mistakenly pointed the ref to att.certainty and att.probability instead of data.certainty and data.probability. I fixed this after jenkins complained, but I imagine that the rng still got created and it might have looked like the one you included in your email.

Is it possible that you're looking at an older RNG and/or need to rebuild?

Also I've been trying to build locally to replicate your problem, but when I `make` in P5 I get this error: /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: stylesheet /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl doesn't exist. Have I missed a step?

Raff

...
On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes wrote:

Hi all,

The @cert in att.global.responsibility now has a janus value (att.certainty or att.probability). However, when I process one of my existing ODDs using the current dev tree, I get an empty value in my RNG:

<define name="att.global.responsibility.attributes"> <ref name="att.global.responsibility.attribute.cert"/> <ref name="att.global.responsibility.attribute.resp"/> </define> <define name="att.global.responsibility.attribute.cert"> <optional> <attribute name="cert"> http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) signifies the degree of certainty associated with the intervention or interpretation. <empty/> </attribute> </optional> </define> <define name="att.global.responsibility.attribute.resp"> <optional> <attribute name="resp"> http://relaxng.org/ns/compatibility/annotations/1.0">(responsible party) indicates the agency responsible for the intervention or interpretation, for example an editor or transcriber. <list> <oneOrMore> <ref name="data.pointer"/> </oneOrMore> </list> </attribute> </optional> </define>

Any idea why this might be happening?

Cheers, Martin -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

Martin Holmes

9:59 a.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

I bet Lou and Syd would be able to tell us why. It's got me completely stumped. Cheers, Martin On 15-11-03 09:07 AM, Hugh Cayless wrote:

...

Yeah, I see the same thing...

Sent from my phone.

...
On Nov 3, 2015, at 11:49, Martin Holmes wrote:

I'm very puzzled by this. Could someone else generate an RNG from this ODD:

http://web.uvic.ca/~mholmes/mariage.odd

using the dev P5 and Stylesheets, and see if they get the same empty content model for @cert?

Can anyone suggest why this might be happening? It doesn't happen in the regular P5 build using teiall.xml in Test.

Cheers, Martin

...
On 15-11-03 07:56 AM, Martin Holmes wrote: Hi Raff,

No, this is freshly-generated from an existing ODD, using the latest bleeding-edge Oxygen plugin, meaning that it's the current stylesheets and the current P5 from dev.

To build P5, you usually have to point it at the location of your stylesheet checkout:

XSL=/home/rviglian/Projects/tei-code/TEI/Stylesheets

or whatever.

Cheers, Martin

...
On 15-11-03 07:13 AM, Raffaele Viglianti wrote: Hi Martin,

When I first made that change, I mistakenly pointed the ref to att.certainty and att.probability instead of data.certainty and data.probability. I fixed this after jenkins complained, but I imagine that the rng still got created and it might have looked like the one you included in your email.

Is it possible that you're looking at an older RNG and/or need to rebuild?

Also I've been trying to build locally to replicate your problem, but when I `make` in P5 I get this error: /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: stylesheet /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl doesn't exist. Have I missed a step?

Raff

...
On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes wrote:

Hi all,

The @cert in att.global.responsibility now has a janus value (att.certainty or att.probability). However, when I process one of my existing ODDs using the current dev tree, I get an empty value in my RNG:

<define name="att.global.responsibility.attributes"> <ref name="att.global.responsibility.attribute.cert"/> <ref name="att.global.responsibility.attribute.resp"/> </define> <define name="att.global.responsibility.attribute.cert"> <optional> <attribute name="cert"> http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) signifies the degree of certainty associated with the intervention or interpretation. <empty/> </attribute> </optional> </define> <define name="att.global.responsibility.attribute.resp"> <optional> <attribute name="resp"> http://relaxng.org/ns/compatibility/annotations/1.0">(responsible party) indicates the agency responsible for the intervention or interpretation, for example an editor or transcriber. <list> <oneOrMore> <ref name="data.pointer"/> </oneOrMore> </list> </attribute> </optional> </define>

Any idea why this might be happening?

Cheers, Martin -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

Hugh Cayless

11:35 a.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

I’ve tracked it down, but I’m not 100% sure whether it’s actually a bug or whether we’re doing it wrong. The odd2odd.xsl checks to see if you have an rng:ref that’s the *child* of a <datatype>, but here we have a datatype that’s a choice between two datatypes (it’s the grandchild of <datatype>). The assumption seems to have been that datatypes couldn’t be defined by composition. Was that a conscious choice or just an oversight? If the latter, then I’ll check in a fix to the Stylesheets. If the former, then we need to define att.global.responsibility differently… Thoughts?

...

On Nov 3, 2015, at 12:59 , Martin Holmes wrote:

I bet Lou and Syd would be able to tell us why. It's got me completely stumped.

Cheers, Martin

On 15-11-03 09:07 AM, Hugh Cayless wrote:

...
Yeah, I see the same thing...

Sent from my phone.

...
On Nov 3, 2015, at 11:49, Martin Holmes wrote:

I'm very puzzled by this. Could someone else generate an RNG from this ODD:

http://web.uvic.ca/~mholmes/mariage.odd

using the dev P5 and Stylesheets, and see if they get the same empty content model for @cert?

Can anyone suggest why this might be happening? It doesn't happen in the regular P5 build using teiall.xml in Test.

Cheers, Martin

...
On 15-11-03 07:56 AM, Martin Holmes wrote: Hi Raff,

No, this is freshly-generated from an existing ODD, using the latest bleeding-edge Oxygen plugin, meaning that it's the current stylesheets and the current P5 from dev.

To build P5, you usually have to point it at the location of your stylesheet checkout:

XSL=/home/rviglian/Projects/tei-code/TEI/Stylesheets

or whatever.

Cheers, Martin

...
On 15-11-03 07:13 AM, Raffaele Viglianti wrote: Hi Martin,

When I first made that change, I mistakenly pointed the ref to att.certainty and att.probability instead of data.certainty and data.probability. I fixed this after jenkins complained, but I imagine that the rng still got created and it might have looked like the one you included in your email.

Is it possible that you're looking at an older RNG and/or need to rebuild?

Also I've been trying to build locally to replicate your problem, but when I `make` in P5 I get this error: /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: stylesheet /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl doesn't exist. Have I missed a step?

Raff

...
On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes wrote:

Hi all,

The @cert in att.global.responsibility now has a janus value (att.certainty or att.probability). However, when I process one of my existing ODDs using the current dev tree, I get an empty value in my RNG:

<define name="att.global.responsibility.attributes"> <ref name="att.global.responsibility.attribute.cert"/> <ref name="att.global.responsibility.attribute.resp"/> </define> <define name="att.global.responsibility.attribute.cert"> <optional> <attribute name="cert"> http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) signifies the degree of certainty associated with the intervention or interpretation. <empty/> </attribute> </optional> </define> <define name="att.global.responsibility.attribute.resp"> <optional> <attribute name="resp"> http://relaxng.org/ns/compatibility/annotations/1.0">(responsible party) indicates the agency responsible for the intervention or interpretation, for example an editor or transcriber. <list> <oneOrMore> <ref name="data.pointer"/> </oneOrMore> </list> </attribute> </optional> </define>

Any idea why this might be happening?

Cheers, Martin -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

Martin Holmes

11:40 a.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

What I don't understand, then, is why teiall.odd in the tests generates a correct teiall.rng. If you run make test, then look at the generated Test/teiall.rng, do you see a broken content model? Cheers, Martin On 15-11-03 11:35 AM, Hugh Cayless wrote:

...

I’ve tracked it down, but I’m not 100% sure whether it’s actually a bug or whether we’re doing it wrong. The odd2odd.xsl checks to see if you have an rng:ref that’s the *child* of a <datatype>, but here we have a datatype that’s a choice between two datatypes (it’s the grandchild of <datatype>). The assumption seems to have been that datatypes couldn’t be defined by composition. Was that a conscious choice or just an oversight? If the latter, then I’ll check in a fix to the Stylesheets. If the former, then we need to define att.global.responsibility differently…

Thoughts?

...
On Nov 3, 2015, at 12:59 , Martin Holmes wrote:

I bet Lou and Syd would be able to tell us why. It's got me completely stumped.

Cheers, Martin

On 15-11-03 09:07 AM, Hugh Cayless wrote:

...
Yeah, I see the same thing...

Sent from my phone.

...
On Nov 3, 2015, at 11:49, Martin Holmes wrote:

I'm very puzzled by this. Could someone else generate an RNG from this ODD:

http://web.uvic.ca/~mholmes/mariage.odd

using the dev P5 and Stylesheets, and see if they get the same empty content model for @cert?

Can anyone suggest why this might be happening? It doesn't happen in the regular P5 build using teiall.xml in Test.

Cheers, Martin

...
On 15-11-03 07:56 AM, Martin Holmes wrote: Hi Raff,

No, this is freshly-generated from an existing ODD, using the latest bleeding-edge Oxygen plugin, meaning that it's the current stylesheets and the current P5 from dev.

To build P5, you usually have to point it at the location of your stylesheet checkout:

XSL=/home/rviglian/Projects/tei-code/TEI/Stylesheets

or whatever.

Cheers, Martin

...
On 15-11-03 07:13 AM, Raffaele Viglianti wrote: Hi Martin,

When I first made that change, I mistakenly pointed the ref to att.certainty and att.probability instead of data.certainty and data.probability. I fixed this after jenkins complained, but I imagine that the rng still got created and it might have looked like the one you included in your email.

Is it possible that you're looking at an older RNG and/or need to rebuild?

Also I've been trying to build locally to replicate your problem, but when I `make` in P5 I get this error: /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: stylesheet /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl doesn't exist. Have I missed a step?

Raff

> On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes > wrote: > > Hi all, > > The @cert in att.global.responsibility now has a janus > value (att.certainty or att.probability). However, when I > process one of my existing ODDs using the current dev > tree, I get an empty value in my RNG: > > <define name="att.global.responsibility.attributes"> <ref > name="att.global.responsibility.attribute.cert"/> <ref > name="att.global.responsibility.attribute.resp"/> > </define> <define > name="att.global.responsibility.attribute.cert"> > <optional> <attribute name="cert"> xmlns:a=" > http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) > > signifies the degree of certainty associated with the intervention or > interpretation. <empty/> </attribute> > </optional> </define> <define > name="att.global.responsibility.attribute.resp"> > <optional> <attribute name="resp"> xmlns:a=" > http://relaxng.org/ns/compatibility/annotations/1.0">(responsible > party) indicates the agency responsible for the > intervention or interpretation, for example an editor or > transcriber. <list> <oneOrMore> <ref > name="data.pointer"/> </oneOrMore> </list> </attribute> > </optional> </define> > > Any idea why this might be happening? > > Cheers, Martin -- tei-council mailing list > tei-council@lists.tei-c.org > http://lists.lists.tei-c.org/mailman/listinfo/tei-council > > > PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

Hugh Cayless

11:42 a.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

My guess would be that it’s getting in another way with tei_all. Will check though.

...

On Nov 3, 2015, at 14:40 , Martin Holmes wrote:

What I don't understand, then, is why teiall.odd in the tests generates a correct teiall.rng. If you run make test, then look at the generated Test/teiall.rng, do you see a broken content model?

Cheers, Martin

On 15-11-03 11:35 AM, Hugh Cayless wrote:

...
I’ve tracked it down, but I’m not 100% sure whether it’s actually a bug or whether we’re doing it wrong. The odd2odd.xsl checks to see if you have an rng:ref that’s the *child* of a <datatype>, but here we have a datatype that’s a choice between two datatypes (it’s the grandchild of <datatype>). The assumption seems to have been that datatypes couldn’t be defined by composition. Was that a conscious choice or just an oversight? If the latter, then I’ll check in a fix to the Stylesheets. If the former, then we need to define att.global.responsibility differently…

Thoughts?

...
On Nov 3, 2015, at 12:59 , Martin Holmes wrote:

I bet Lou and Syd would be able to tell us why. It's got me completely stumped.

Cheers, Martin

On 15-11-03 09:07 AM, Hugh Cayless wrote:

...
Yeah, I see the same thing...

Sent from my phone.

...
On Nov 3, 2015, at 11:49, Martin Holmes wrote:

I'm very puzzled by this. Could someone else generate an RNG from this ODD:

http://web.uvic.ca/~mholmes/mariage.odd

using the dev P5 and Stylesheets, and see if they get the same empty content model for @cert?

Can anyone suggest why this might be happening? It doesn't happen in the regular P5 build using teiall.xml in Test.

Cheers, Martin

...
On 15-11-03 07:56 AM, Martin Holmes wrote: Hi Raff,

No, this is freshly-generated from an existing ODD, using the latest bleeding-edge Oxygen plugin, meaning that it's the current stylesheets and the current P5 from dev.

To build P5, you usually have to point it at the location of your stylesheet checkout:

XSL=/home/rviglian/Projects/tei-code/TEI/Stylesheets

or whatever.

Cheers, Martin

> On 15-11-03 07:13 AM, Raffaele Viglianti wrote: Hi Martin, > > When I first made that change, I mistakenly pointed the ref > to att.certainty and att.probability instead of > data.certainty and data.probability. I fixed this after > jenkins complained, but I imagine that the rng still got > created and it might have looked like the one you included > in your email. > > Is it possible that you're looking at an older RNG and/or > need to rebuild? > > Also I've been trying to build locally to replicate your > problem, but when I `make` in P5 I get this error: > /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: > stylesheet > /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl > doesn't exist. Have I missed a step? > > Raff > >> On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes >> wrote: >> >> Hi all, >> >> The @cert in att.global.responsibility now has a janus >> value (att.certainty or att.probability). However, when I >> process one of my existing ODDs using the current dev >> tree, I get an empty value in my RNG: >> >> <define name="att.global.responsibility.attributes"> <ref >> name="att.global.responsibility.attribute.cert"/> <ref >> name="att.global.responsibility.attribute.resp"/> >> </define> <define >> name="att.global.responsibility.attribute.cert"> >> <optional> <attribute name="cert"> > xmlns:a=" >> http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) >> >> signifies the degree of certainty associated with the intervention or >> interpretation. <empty/> </attribute> >> </optional> </define> <define >> name="att.global.responsibility.attribute.resp"> >> <optional> <attribute name="resp"> > xmlns:a=" >> http://relaxng.org/ns/compatibility/annotations/1.0">(responsible >> party) indicates the agency responsible for the >> intervention or interpretation, for example an editor or >> transcriber. <list> <oneOrMore> <ref >> name="data.pointer"/> </oneOrMore> </list> </attribute> >> </optional> </define> >> >> Any idea why this might be happening? >> >> Cheers, Martin -- tei-council mailing list >> tei-council@lists.tei-c.org >> http://lists.lists.tei-c.org/mailman/listinfo/tei-council >> >> >> PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

Hugh Cayless

11:52 a.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

Yeah, All works, but I think what’s happening is that in your ODD, the content model is being set to empty because the data.certainty and data.probability macros have been stripped, and that because they’re only getting included via the att.global.responsibility reference, in which, because the datatype is a choice between two datatypes, the references aren’t resolved, meaning the macros aren’t included. This does seem like a bug, doesn’t it? I’ll just go ahead and fix it...

...

On Nov 3, 2015, at 14:42 , Hugh Cayless wrote:

My guess would be that it’s getting in another way with tei_all. Will check though.

...
On Nov 3, 2015, at 14:40 , Martin Holmes wrote:

What I don't understand, then, is why teiall.odd in the tests generates a correct teiall.rng. If you run make test, then look at the generated Test/teiall.rng, do you see a broken content model?

Cheers, Martin

On 15-11-03 11:35 AM, Hugh Cayless wrote:

...
I’ve tracked it down, but I’m not 100% sure whether it’s actually a bug or whether we’re doing it wrong. The odd2odd.xsl checks to see if you have an rng:ref that’s the *child* of a <datatype>, but here we have a datatype that’s a choice between two datatypes (it’s the grandchild of <datatype>). The assumption seems to have been that datatypes couldn’t be defined by composition. Was that a conscious choice or just an oversight? If the latter, then I’ll check in a fix to the Stylesheets. If the former, then we need to define att.global.responsibility differently…

Thoughts?

...
On Nov 3, 2015, at 12:59 , Martin Holmes wrote:

I bet Lou and Syd would be able to tell us why. It's got me completely stumped.

Cheers, Martin

On 15-11-03 09:07 AM, Hugh Cayless wrote:

...
Yeah, I see the same thing...

Sent from my phone.

...
On Nov 3, 2015, at 11:49, Martin Holmes wrote:

I'm very puzzled by this. Could someone else generate an RNG from this ODD:

http://web.uvic.ca/~mholmes/mariage.odd

using the dev P5 and Stylesheets, and see if they get the same empty content model for @cert?

Can anyone suggest why this might be happening? It doesn't happen in the regular P5 build using teiall.xml in Test.

Cheers, Martin

> On 15-11-03 07:56 AM, Martin Holmes wrote: Hi Raff, > > No, this is freshly-generated from an existing ODD, using the > latest bleeding-edge Oxygen plugin, meaning that it's the > current stylesheets and the current P5 from dev. > > To build P5, you usually have to point it at the location of > your stylesheet checkout: > > XSL=/home/rviglian/Projects/tei-code/TEI/Stylesheets > > or whatever. > > Cheers, Martin > >> On 15-11-03 07:13 AM, Raffaele Viglianti wrote: Hi Martin, >> >> When I first made that change, I mistakenly pointed the ref >> to att.certainty and att.probability instead of >> data.certainty and data.probability. I fixed this after >> jenkins complained, but I imagine that the rng still got >> created and it might have looked like the one you included >> in your email. >> >> Is it possible that you're looking at an older RNG and/or >> need to rebuild? >> >> Also I've been trying to build locally to replicate your >> problem, but when I `make` in P5 I get this error: >> /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: >> stylesheet >> /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl >> doesn't exist. Have I missed a step? >> >> Raff >> >>> On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes >>> wrote: >>> >>> Hi all, >>> >>> The @cert in att.global.responsibility now has a janus >>> value (att.certainty or att.probability). However, when I >>> process one of my existing ODDs using the current dev >>> tree, I get an empty value in my RNG: >>> >>> <define name="att.global.responsibility.attributes"> <ref >>> name="att.global.responsibility.attribute.cert"/> <ref >>> name="att.global.responsibility.attribute.resp"/> >>> </define> <define >>> name="att.global.responsibility.attribute.cert"> >>> <optional> <attribute name="cert"> >> xmlns:a=" >>> http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) >>> >>> signifies the degree of certainty associated with the intervention or >>> interpretation. <empty/> </attribute> >>> </optional> </define> <define >>> name="att.global.responsibility.attribute.resp"> >>> <optional> <attribute name="resp"> >> xmlns:a=" >>> http://relaxng.org/ns/compatibility/annotations/1.0">(responsible >>> party) indicates the agency responsible for the >>> intervention or interpretation, for example an editor or >>> transcriber. <list> <oneOrMore> <ref >>> name="data.pointer"/> </oneOrMore> </list> </attribute> >>> </optional> </define> >>> >>> Any idea why this might be happening? >>> >>> Cheers, Martin -- tei-council mailing list >>> tei-council@lists.tei-c.org >>> http://lists.lists.tei-c.org/mailman/listinfo/tei-council >>> >>> >>> PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

Hugh Cayless

12:05 p.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

Pushed a fix. We’ll see if it breaks anything else...

...

On Nov 3, 2015, at 14:52 , Hugh Cayless wrote:

Yeah, All works, but I think what’s happening is that in your ODD, the content model is being set to empty because the data.certainty and data.probability macros have been stripped, and that because they’re only getting included via the att.global.responsibility reference, in which, because the datatype is a choice between two datatypes, the references aren’t resolved, meaning the macros aren’t included. This does seem like a bug, doesn’t it? I’ll just go ahead and fix it...

...
On Nov 3, 2015, at 14:42 , Hugh Cayless wrote:

My guess would be that it’s getting in another way with tei_all. Will check though.

...
On Nov 3, 2015, at 14:40 , Martin Holmes wrote:

What I don't understand, then, is why teiall.odd in the tests generates a correct teiall.rng. If you run make test, then look at the generated Test/teiall.rng, do you see a broken content model?

Cheers, Martin

On 15-11-03 11:35 AM, Hugh Cayless wrote:

...
I’ve tracked it down, but I’m not 100% sure whether it’s actually a bug or whether we’re doing it wrong. The odd2odd.xsl checks to see if you have an rng:ref that’s the *child* of a <datatype>, but here we have a datatype that’s a choice between two datatypes (it’s the grandchild of <datatype>). The assumption seems to have been that datatypes couldn’t be defined by composition. Was that a conscious choice or just an oversight? If the latter, then I’ll check in a fix to the Stylesheets. If the former, then we need to define att.global.responsibility differently…

Thoughts?

...
On Nov 3, 2015, at 12:59 , Martin Holmes wrote:

I bet Lou and Syd would be able to tell us why. It's got me completely stumped.

Cheers, Martin

On 15-11-03 09:07 AM, Hugh Cayless wrote:

...
Yeah, I see the same thing...

Sent from my phone.

> On Nov 3, 2015, at 11:49, Martin Holmes > wrote: > > I'm very puzzled by this. Could someone else generate an RNG > from this ODD: > > http://web.uvic.ca/~mholmes/mariage.odd > > using the dev P5 and Stylesheets, and see if they get the same > empty content model for @cert? > > Can anyone suggest why this might be happening? It doesn't > happen in the regular P5 build using teiall.xml in Test. > > Cheers, Martin > >> On 15-11-03 07:56 AM, Martin Holmes wrote: Hi Raff, >> >> No, this is freshly-generated from an existing ODD, using the >> latest bleeding-edge Oxygen plugin, meaning that it's the >> current stylesheets and the current P5 from dev. >> >> To build P5, you usually have to point it at the location of >> your stylesheet checkout: >> >> XSL=/home/rviglian/Projects/tei-code/TEI/Stylesheets >> >> or whatever. >> >> Cheers, Martin >> >>> On 15-11-03 07:13 AM, Raffaele Viglianti wrote: Hi Martin, >>> >>> When I first made that change, I mistakenly pointed the ref >>> to att.certainty and att.probability instead of >>> data.certainty and data.probability. I fixed this after >>> jenkins complained, but I imagine that the rng still got >>> created and it might have looked like the one you included >>> in your email. >>> >>> Is it possible that you're looking at an older RNG and/or >>> need to rebuild? >>> >>> Also I've been trying to build locally to replicate your >>> problem, but when I `make` in P5 I get this error: >>> /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: >>> stylesheet >>> /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl >>> doesn't exist. Have I missed a step? >>> >>> Raff >>> >>>> On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes >>>> wrote: >>>> >>>> Hi all, >>>> >>>> The @cert in att.global.responsibility now has a janus >>>> value (att.certainty or att.probability). However, when I >>>> process one of my existing ODDs using the current dev >>>> tree, I get an empty value in my RNG: >>>> >>>> <define name="att.global.responsibility.attributes"> <ref >>>> name="att.global.responsibility.attribute.cert"/> <ref >>>> name="att.global.responsibility.attribute.resp"/> >>>> </define> <define >>>> name="att.global.responsibility.attribute.cert"> >>>> <optional> <attribute name="cert"> >>> xmlns:a=" >>>> http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) >>>> >>>> signifies the degree of certainty associated with the intervention or >>>> interpretation. <empty/> </attribute> >>>> </optional> </define> <define >>>> name="att.global.responsibility.attribute.resp"> >>>> <optional> <attribute name="resp"> >>> xmlns:a=" >>>> http://relaxng.org/ns/compatibility/annotations/1.0">(responsible >>>> party) indicates the agency responsible for the >>>> intervention or interpretation, for example an editor or >>>> transcriber. <list> <oneOrMore> <ref >>>> name="data.pointer"/> </oneOrMore> </list> </attribute> >>>> </optional> </define> >>>> >>>> Any idea why this might be happening? >>>> >>>> Cheers, Martin -- tei-council mailing list >>>> tei-council@lists.tei-c.org >>>> http://lists.lists.tei-c.org/mailman/listinfo/tei-council >>>> >>>> >>>> PLEASE NOTE: postings to this list are publicly archived > -- tei-council mailing list tei-council@lists.tei-c.org > http://lists.lists.tei-c.org/mailman/listinfo/tei-council > > PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

Martin Holmes

12:42 p.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

Hi Hugh, On 15-11-03 11:52 AM, Hugh Cayless wrote:

...

Yeah, All works, but I think what’s happening is that in your ODD, the content model is being set to empty because the data.certainty and data.probability macros have been stripped,

How have they been stripped? That's what I don't really understand; I'm not doing anything to strip them (any more than teiall.odd is doing anything to include them).

...

and that because they’re only getting included via the att.global.responsibility reference, in which, because the datatype is a choice between two datatypes, the references aren’t resolved, meaning the macros aren’t included. This does seem like a bug, doesn’t it? I’ll just go ahead and fix it...

...

...
On Nov 3, 2015, at 14:42 , Hugh Cayless wrote:

My guess would be that it’s getting in another way with tei_all. Will check though.

...
On Nov 3, 2015, at 14:40 , Martin Holmes wrote:

What I don't understand, then, is why teiall.odd in the tests generates a correct teiall.rng. If you run make test, then look at the generated Test/teiall.rng, do you see a broken content model?

Cheers, Martin

On 15-11-03 11:35 AM, Hugh Cayless wrote:

...
I’ve tracked it down, but I’m not 100% sure whether it’s actually a bug or whether we’re doing it wrong. The odd2odd.xsl checks to see if you have an rng:ref that’s the *child* of a <datatype>, but here we have a datatype that’s a choice between two datatypes (it’s the grandchild of <datatype>). The assumption seems to have been that datatypes couldn’t be defined by composition. Was that a conscious choice or just an oversight? If the latter, then I’ll check in a fix to the Stylesheets. If the former, then we need to define att.global.responsibility differently…

Thoughts?

...
On Nov 3, 2015, at 12:59 , Martin Holmes wrote:

I bet Lou and Syd would be able to tell us why. It's got me completely stumped.

Cheers, Martin

On 15-11-03 09:07 AM, Hugh Cayless wrote:

...
Yeah, I see the same thing...

Sent from my phone.

> On Nov 3, 2015, at 11:49, Martin Holmes > wrote: > > I'm very puzzled by this. Could someone else generate an > RNG from this ODD: > > http://web.uvic.ca/~mholmes/mariage.odd > > using the dev P5 and Stylesheets, and see if they get the > same empty content model for @cert? > > Can anyone suggest why this might be happening? It > doesn't happen in the regular P5 build using teiall.xml > in Test. > > Cheers, Martin > >> On 15-11-03 07:56 AM, Martin Holmes wrote: Hi Raff, >> >> No, this is freshly-generated from an existing ODD, >> using the latest bleeding-edge Oxygen plugin, meaning >> that it's the current stylesheets and the current P5 >> from dev. >> >> To build P5, you usually have to point it at the >> location of your stylesheet checkout: >> >> XSL=/home/rviglian/Projects/tei-code/TEI/Stylesheets >> >> or whatever. >> >> Cheers, Martin >> >>> On 15-11-03 07:13 AM, Raffaele Viglianti wrote: Hi >>> Martin, >>> >>> When I first made that change, I mistakenly pointed >>> the ref to att.certainty and att.probability instead >>> of data.certainty and data.probability. I fixed this >>> after jenkins complained, but I imagine that the rng >>> still got created and it might have looked like the >>> one you included in your email. >>> >>> Is it possible that you're looking at an older RNG >>> and/or need to rebuild? >>> >>> Also I've been trying to build locally to replicate >>> your problem, but when I `make` in P5 I get this >>> error: >>> /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: >>> >>>

...

...
...
...
...
...
>>> /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl >>> >>> doesn't exist. Have I missed a step? >>> >>> Raff >>> >>>> On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes >>>> wrote: >>>> >>>> Hi all, >>>> >>>> The @cert in att.global.responsibility now has a >>>> janus value (att.certainty or att.probability). >>>> However, when I process one of my existing ODDs >>>> using the current dev tree, I get an empty value in >>>> my RNG: >>>> >>>> <define >>>> name="att.global.responsibility.attributes"> <ref >>>> name="att.global.responsibility.attribute.cert"/> >>>> <ref >>>> name="att.global.responsibility.attribute.resp"/> >>>> </define> <define >>>> name="att.global.responsibility.attribute.cert"> >>>> <optional> <attribute name="cert"> >>>> http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) >>>> >>>>

...
...
...
>>>> signifies the degree of certainty associated with the intervention or >>>> interpretation. <empty/> >>>> </attribute> </optional> </define> <define >>>> name="att.global.responsibility.attribute.resp"> >>>> <optional> <attribute name="resp"> >>>> http://relaxng.org/ns/compatibility/annotations/1.0">(responsible >>>> >>>>

Thanks for that. Cheers, Martin stylesheet party) indicates the agency responsible for the

...

...
...
...
...
...
>>>> intervention or interpretation, for example an >>>> editor or transcriber. <list> >>>> <oneOrMore> <ref name="data.pointer"/> </oneOrMore> >>>> </list> </attribute> </optional> </define> >>>> >>>> Any idea why this might be happening? >>>> >>>> Cheers, Martin -- tei-council mailing list >>>> tei-council@lists.tei-c.org >>>> http://lists.lists.tei-c.org/mailman/listinfo/tei-council >>>> >>>> >>>>

...
...
...
>>>> PLEASE NOTE: postings to this list are publicly archived > -- tei-council mailing list tei-council@lists.tei-c.org > http://lists.lists.tei-c.org/mailman/listinfo/tei-council > > > PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

Hugh Cayless

12:49 p.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

I think in All, they have multiple ways of getting in (e.g. via certainty/@cert). In your ODD, they have only one way, and that’s the way that doesn’t work (via its inclusion of att.global.responsibility). The datatype works properly if the macros are still around at the stage when it’s written out, and in All, they are, but in yours, they aren’t, and so @cert is set to empty as a fallback. So you aren’t doing anything to strip them. You aren’t doing anything wrong at all. The ODD expansion routine was assuming a datatype can only be defined with reference to a single macro rather than (e.g.) a choice between two macros.

...

On Nov 3, 2015, at 15:42 , Martin Holmes wrote:

Hi Hugh,

On 15-11-03 11:52 AM, Hugh Cayless wrote:

...
Yeah, All works, but I think what’s happening is that in your ODD, the content model is being set to empty because the data.certainty and data.probability macros have been stripped,

How have they been stripped? That's what I don't really understand; I'm not doing anything to strip them (any more than teiall.odd is doing anything to include them).

...
and that because they’re only getting included via the att.global.responsibility reference, in which, because the datatype is a choice between two datatypes, the references aren’t resolved, meaning the macros aren’t included. This does seem like a bug, doesn’t it? I’ll just go ahead and fix it...

Thanks for that.

Cheers, Martin

...
...
On Nov 3, 2015, at 14:42 , Hugh Cayless wrote:

My guess would be that it’s getting in another way with tei_all. Will check though.

...
On Nov 3, 2015, at 14:40 , Martin Holmes wrote:

What I don't understand, then, is why teiall.odd in the tests generates a correct teiall.rng. If you run make test, then look at the generated Test/teiall.rng, do you see a broken content model?

Cheers, Martin

...
I’ve tracked it down, but I’m not 100% sure whether it’s actually a bug or whether we’re doing it wrong. The odd2odd.xsl checks to see if you have an rng:ref that’s the *child* of a <datatype>, but here we have a datatype that’s a choice between two datatypes (it’s the grandchild of <datatype>). The assumption seems to have been that datatypes couldn’t be defined by composition. Was that a conscious choice or just an oversight? If the latter, then I’ll check in a fix to the Stylesheets. If the former, then we need to define att.global.responsibility differently…

Thoughts?

...
On Nov 3, 2015, at 12:59 , Martin Holmes wrote:

I bet Lou and Syd would be able to tell us why. It's got me completely stumped.

Cheers, Martin

On 15-11-03 09:07 AM, Hugh Cayless wrote: > Yeah, I see the same thing... > > Sent from my phone. > >> On Nov 3, 2015, at 11:49, Martin Holmes >> wrote: >> >> I'm very puzzled by this. Could someone else generate an >> RNG from this ODD: >> >> http://web.uvic.ca/~mholmes/mariage.odd >> >> using the dev P5 and Stylesheets, and see if they get the >> same empty content model for @cert? >> >> Can anyone suggest why this might be happening? It >> doesn't happen in the regular P5 build using teiall.xml >> in Test. >> >> Cheers, Martin >> >>> On 15-11-03 07:56 AM, Martin Holmes wrote: Hi Raff, >>> >>> No, this is freshly-generated from an existing ODD, >>> using the latest bleeding-edge Oxygen plugin, meaning >>> that it's the current stylesheets and the current P5 >>> from dev. >>> >>> To build P5, you usually have to point it at the >>> location of your stylesheet checkout: >>> >>> XSL=/home/rviglian/Projects/tei-code/TEI/Stylesheets >>> >>> or whatever. >>> >>> Cheers, Martin >>> >>>> On 15-11-03 07:13 AM, Raffaele Viglianti wrote: Hi >>>> Martin, >>>> >>>> When I first made that change, I mistakenly pointed >>>> the ref to att.certainty and att.probability instead >>>> of data.certainty and data.probability. I fixed this >>>> after jenkins complained, but I imagine that the rng >>>> still got created and it might have looked like the >>>> one you included in your email. >>>> >>>> Is it possible that you're looking at an older RNG >>>> and/or need to rebuild? >>>> >>>> Also I've been trying to build locally to replicate >>>> your problem, but when I `make` in P5 I get this >>>> error: >>>> /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: >>>> >>>>

On 15-11-03 11:35 AM, Hugh Cayless wrote: stylesheet

...
...
>>>> /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl >>>> >>>> doesn't exist. Have I missed a step? >>>> >>>> Raff >>>> >>>>> On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes >>>>> wrote: >>>>> >>>>> Hi all, >>>>> >>>>> The @cert in att.global.responsibility now has a >>>>> janus value (att.certainty or att.probability). >>>>> However, when I process one of my existing ODDs >>>>> using the current dev tree, I get an empty value in >>>>> my RNG: >>>>> >>>>> <define >>>>> name="att.global.responsibility.attributes"> <ref >>>>> name="att.global.responsibility.attribute.cert"/> >>>>> <ref >>>>> name="att.global.responsibility.attribute.resp"/> >>>>> </define> <define >>>>> name="att.global.responsibility.attribute.cert"> >>>>> <optional> <attribute name="cert"> >>>>> http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) >>>>> >>>>>

...
...
>>>>> signifies the degree of certainty associated with the intervention or >>>>> interpretation. <empty/> >>>>> </attribute> </optional> </define> <define >>>>> name="att.global.responsibility.attribute.resp"> >>>>> <optional> <attribute name="resp"> >>>>> http://relaxng.org/ns/compatibility/annotations/1.0">(responsible >>>>> >>>>> party) indicates the agency responsible for the >>>>> intervention or interpretation, for example an >>>>> editor or transcriber. <list> >>>>> <oneOrMore> <ref name="data.pointer"/> </oneOrMore> >>>>> </list> </attribute> </optional> </define> >>>>> >>>>> Any idea why this might be happening? >>>>> >>>>> Cheers, Martin -- tei-council mailing list >>>>> tei-council@lists.tei-c.org >>>>> http://lists.lists.tei-c.org/mailman/listinfo/tei-council >>>>> >>>>> >>>>>

...
...
>>>>> PLEASE NOTE: postings to this list are publicly archived >> -- tei-council mailing list tei-council@lists.tei-c.org >> http://lists.lists.tei-c.org/mailman/listinfo/tei-council >> >> >> PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org mailto:tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

Hugh Cayless

12:53 p.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

I should reiterate that some might consider the inability to define datatypes as composites of other datatypes might be considered by some to be a feature, not a bug. If anyone has any objections to that ability, they should start yelling about it now, before it’s too late to walk it back. I’m not sure I have an opinion about it.

...

On Nov 3, 2015, at 15:49 , Hugh Cayless wrote:

I think in All, they have multiple ways of getting in (e.g. via certainty/@cert). In your ODD, they have only one way, and that’s the way that doesn’t work (via its inclusion of att.global.responsibility). The datatype works properly if the macros are still around at the stage when it’s written out, and in All, they are, but in yours, they aren’t, and so @cert is set to empty as a fallback. So you aren’t doing anything to strip them. You aren’t doing anything wrong at all. The ODD expansion routine was assuming a datatype can only be defined with reference to a single macro rather than (e.g.) a choice between two macros.

...
On Nov 3, 2015, at 15:42 , Martin Holmes mailto:mholmes@uvic.ca> wrote:

Hi Hugh,

On 15-11-03 11:52 AM, Hugh Cayless wrote:

...
Yeah, All works, but I think what’s happening is that in your ODD, the content model is being set to empty because the data.certainty and data.probability macros have been stripped,

How have they been stripped? That's what I don't really understand; I'm not doing anything to strip them (any more than teiall.odd is doing anything to include them).

...
and that because they’re only getting included via the att.global.responsibility reference, in which, because the datatype is a choice between two datatypes, the references aren’t resolved, meaning the macros aren’t included. This does seem like a bug, doesn’t it? I’ll just go ahead and fix it...

Thanks for that.

Cheers, Martin

...
...
On Nov 3, 2015, at 14:42 , Hugh Cayless mailto:philomousos@gmail.com> wrote:

My guess would be that it’s getting in another way with tei_all. Will check though.

...
On Nov 3, 2015, at 14:40 , Martin Holmes mailto:mholmes@uvic.ca> wrote:

What I don't understand, then, is why teiall.odd in the tests generates a correct teiall.rng. If you run make test, then look at the generated Test/teiall.rng, do you see a broken content model?

Cheers, Martin

...
I’ve tracked it down, but I’m not 100% sure whether it’s actually a bug or whether we’re doing it wrong. The odd2odd.xsl checks to see if you have an rng:ref that’s the *child* of a <datatype>, but here we have a datatype that’s a choice between two datatypes (it’s the grandchild of <datatype>). The assumption seems to have been that datatypes couldn’t be defined by composition. Was that a conscious choice or just an oversight? If the latter, then I’ll check in a fix to the Stylesheets. If the former, then we need to define att.global.responsibility differently…

Thoughts?

> On Nov 3, 2015, at 12:59 , Martin Holmes mailto:mholmes@uvic.ca> > wrote: > > I bet Lou and Syd would be able to tell us why. It's got me > completely stumped. > > Cheers, Martin > > On 15-11-03 09:07 AM, Hugh Cayless wrote: >> Yeah, I see the same thing... >> >> Sent from my phone. >> >>> On Nov 3, 2015, at 11:49, Martin Holmes >>> mailto:mholmes@uvic.ca> wrote: >>> >>> I'm very puzzled by this. Could someone else generate an >>> RNG from this ODD: >>> >>> <http://web.uvic.ca/~mholmes/mariage.odd http://web.uvic.ca/~mholmes/mariage.odd> >>> >>> using the dev P5 and Stylesheets, and see if they get the >>> same empty content model for @cert? >>> >>> Can anyone suggest why this might be happening? It >>> doesn't happen in the regular P5 build using teiall.xml >>> in Test. >>> >>> Cheers, Martin >>> >>>> On 15-11-03 07:56 AM, Martin Holmes wrote: Hi Raff, >>>> >>>> No, this is freshly-generated from an existing ODD, >>>> using the latest bleeding-edge Oxygen plugin, meaning >>>> that it's the current stylesheets and the current P5 >>>> from dev. >>>> >>>> To build P5, you usually have to point it at the >>>> location of your stylesheet checkout: >>>> >>>> XSL=/home/rviglian/Projects/tei-code/TEI/Stylesheets >>>> >>>> or whatever. >>>> >>>> Cheers, Martin >>>> >>>>> On 15-11-03 07:13 AM, Raffaele Viglianti wrote: Hi >>>>> Martin, >>>>> >>>>> When I first made that change, I mistakenly pointed >>>>> the ref to att.certainty and att.probability instead >>>>> of data.certainty and data.probability. I fixed this >>>>> after jenkins complained, but I imagine that the rng >>>>> still got created and it might have looked like the >>>>> one you included in your email. >>>>> >>>>> Is it possible that you're looking at an older RNG >>>>> and/or need to rebuild? >>>>> >>>>> Also I've been trying to build locally to replicate >>>>> your problem, but when I `make` in P5 I get this >>>>> error: >>>>> /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: >>>>> >>>>>

On 15-11-03 11:35 AM, Hugh Cayless wrote: stylesheet

...
>>>>> /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl >>>>> >>>>> doesn't exist. Have I missed a step? >>>>> >>>>> Raff >>>>> >>>>>> On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes >>>>>> mailto:mholmes@uvic.ca> wrote: >>>>>> >>>>>> Hi all, >>>>>> >>>>>> The @cert in att.global.responsibility now has a >>>>>> janus value (att.certainty or att.probability). >>>>>> However, when I process one of my existing ODDs >>>>>> using the current dev tree, I get an empty value in >>>>>> my RNG: >>>>>> >>>>>> <define >>>>>> name="att.global.responsibility.attributes"> <ref >>>>>> name="att.global.responsibility.attribute.cert"/> >>>>>> <ref >>>>>> name="att.global.responsibility.attribute.resp"/> >>>>>> </define> <define >>>>>> name="att.global.responsibility.attribute.cert"> >>>>>> <optional> <attribute name="cert"> >>>>>> http://relaxng.org/ns/compatibility/annotations/1.0 http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) >>>>>> >>>>>>

...
>>>>>> signifies the degree of certainty associated with the intervention or >>>>>> interpretation. <empty/> >>>>>> </attribute> </optional> </define> <define >>>>>> name="att.global.responsibility.attribute.resp"> >>>>>> <optional> <attribute name="resp"> >>>>>> http://relaxng.org/ns/compatibility/annotations/1.0 http://relaxng.org/ns/compatibility/annotations/1.0">(responsible >>>>>> >>>>>> party) indicates the agency responsible for the >>>>>> intervention or interpretation, for example an >>>>>> editor or transcriber. <list> >>>>>> <oneOrMore> <ref name="data.pointer"/> </oneOrMore> >>>>>> </list> </attribute> </optional> </define> >>>>>> >>>>>> Any idea why this might be happening? >>>>>> >>>>>> Cheers, Martin -- tei-council mailing list >>>>>> tei-council@lists.tei-c.org mailto:tei-council@lists.tei-c.org >>>>>> http://lists.lists.tei-c.org/mailman/listinfo/tei-council >>>>>> >>>>>> >>>>>>

...
>>>>>> PLEASE NOTE: postings to this list are publicly archived >>> -- tei-council mailing list tei-council@lists.tei-c.org mailto:tei-council@lists.tei-c.org >>> http://lists.lists.tei-c.org/mailman/listinfo/tei-council http://lists.lists.tei-c.org/mailman/listinfo/tei-council >>> >>> >>> PLEASE NOTE: postings to this list are publicly archived > -- tei-council mailing list tei-council@lists.tei-c.org mailto:tei-council@lists.tei-c.org > http://lists.lists.tei-c.org/mailman/listinfo/tei-council http://lists.lists.tei-c.org/mailman/listinfo/tei-council > > PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org mailto:tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org mailto:tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

Lou Burnard

12:56 p.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

This shouldn't happen in pure odd of course because we don't allow composite datatypes Sent from my Samsung Galaxy Tab®|PRO -------- Original message -------- From: Martin Holmes Date:11/03/2015 20:42 (GMT+00:00) To: tei-council@lists.tei-c.org Subject: Re: [tei-council] Oddity in RNG generation with new @cert datatype Hi Hugh, On 15-11-03 11:52 AM, Hugh Cayless wrote:

...

Yeah, All works, but I think what’s happening is that in your ODD, the content model is being set to empty because the data.certainty and data.probability macros have been stripped,

How have they been stripped? That's what I don't really understand; I'm not doing anything to strip them (any more than teiall.odd is doing anything to include them).

...

and that because they’re only getting included via the att.global.responsibility reference, in which, because the datatype is a choice between two datatypes, the references aren’t resolved, meaning the macros aren’t included. This does seem like a bug, doesn’t it? I’ll just go ahead and fix it...

...

...
On Nov 3, 2015, at 14:42 , Hugh Cayless wrote:

My guess would be that it’s getting in another way with tei_all. Will check though.

...
On Nov 3, 2015, at 14:40 , Martin Holmes wrote:

What I don't understand, then, is why teiall.odd in the tests generates a correct teiall.rng. If you run make test, then look at the generated Test/teiall.rng, do you see a broken content model?

Cheers, Martin

On 15-11-03 11:35 AM, Hugh Cayless wrote:

...
I’ve tracked it down, but I’m not 100% sure whether it’s actually a bug or whether we’re doing it wrong. The odd2odd.xsl checks to see if you have an rng:ref that’s the *child* of a <datatype>, but here we have a datatype that’s a choice between two datatypes (it’s the grandchild of <datatype>). The assumption seems to have been that datatypes couldn’t be defined by composition. Was that a conscious choice or just an oversight? If the latter, then I’ll check in a fix to the Stylesheets. If the former, then we need to define att.global.responsibility differently…

Thoughts?

...
On Nov 3, 2015, at 12:59 , Martin Holmes wrote:

I bet Lou and Syd would be able to tell us why. It's got me completely stumped.

Cheers, Martin

On 15-11-03 09:07 AM, Hugh Cayless wrote:

...
Yeah, I see the same thing...

Sent from my phone.

> On Nov 3, 2015, at 11:49, Martin Holmes > wrote: > > I'm very puzzled by this. Could someone else generate an > RNG from this ODD: > > http://web.uvic.ca/~mholmes/mariage.odd > > using the dev P5 and Stylesheets, and see if they get the > same empty content model for @cert? > > Can anyone suggest why this might be happening? It > doesn't happen in the regular P5 build using teiall.xml > in Test. > > Cheers, Martin > >> On 15-11-03 07:56 AM, Martin Holmes wrote: Hi Raff, >> >> No, this is freshly-generated from an existing ODD, >> using the latest bleeding-edge Oxygen plugin, meaning >> that it's the current stylesheets and the current P5 >> from dev. >> >> To build P5, you usually have to point it at the >> location of your stylesheet checkout: >> >> XSL=/home/rviglian/Projects/tei-code/TEI/Stylesheets >> >> or whatever. >> >> Cheers, Martin >> >>> On 15-11-03 07:13 AM, Raffaele Viglianti wrote: Hi >>> Martin, >>> >>> When I first made that change, I mistakenly pointed >>> the ref to att.certainty and att.probability instead >>> of data.certainty and data.probability. I fixed this >>> after jenkins complained, but I imagine that the rng >>> still got created and it might have looked like the >>> one you included in your email. >>> >>> Is it possible that you're looking at an older RNG >>> and/or need to rebuild? >>> >>> Also I've been trying to build locally to replicate >>> your problem, but when I `make` in P5 I get this >>> error: >>> /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: >>> >>>

...

...
...
...
...
...
>>> /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl >>> >>> doesn't exist. Have I missed a step? >>> >>> Raff >>> >>>> On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes >>>> wrote: >>>> >>>> Hi all, >>>> >>>> The @cert in att.global.responsibility now has a >>>> janus value (att.certainty or att.probability). >>>> However, when I process one of my existing ODDs >>>> using the current dev tree, I get an empty value in >>>> my RNG: >>>> >>>> <define >>>> name="att.global.responsibility.attributes"> <ref >>>> name="att.global.responsibility.attribute.cert"/> >>>> <ref >>>> name="att.global.responsibility.attribute.resp"/> >>>> </define> <define >>>> name="att.global.responsibility.attribute.cert"> >>>> <optional> <attribute name="cert"> >>>> http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) >>>> >>>>

...
...
...
>>>> signifies the degree of certainty associated with the intervention or >>>> interpretation. <empty/> >>>> </attribute> </optional> </define> <define >>>> name="att.global.responsibility.attribute.resp"> >>>> <optional> <attribute name="resp"> >>>> http://relaxng.org/ns/compatibility/annotations/1.0">(responsible >>>> >>>>

Thanks for that. Cheers, Martin stylesheet party) indicates the agency responsible for the

...

...
...
...
...
...
>>>> intervention or interpretation, for example an >>>> editor or transcriber. <list> >>>> <oneOrMore> <ref name="data.pointer"/> </oneOrMore> >>>> </list> </attribute> </optional> </define> >>>> >>>> Any idea why this might be happening? >>>> >>>> Cheers, Martin -- tei-council mailing list >>>> tei-council@lists.tei-c.org >>>> http://lists.lists.tei-c.org/mailman/listinfo/tei-council >>>> >>>> >>>>

...
...
...
>>>> PLEASE NOTE: postings to this list are publicly archived > -- tei-council mailing list tei-council@lists.tei-c.org > http://lists.lists.tei-c.org/mailman/listinfo/tei-council > > > PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council PLEASE NOTE: postings to this list are publicly archived

Raffaele Viglianti

12:59 p.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

If we don't allow composite datatypes in pure ODD, how are we going to convert this specific class? R On Tue, Nov 3, 2015 at 3:56 PM, Lou Burnard wrote:

...

This shouldn't happen in pure odd of course because we don't allow composite datatypes

Sent from my Samsung Galaxy Tab®|PRO

-------- Original message -------- From: Martin Holmes Date:11/03/2015 20:42 (GMT+00:00) To: tei-council@lists.tei-c.org Subject: Re: [tei-council] Oddity in RNG generation with new @cert datatype

Hi Hugh,

On 15-11-03 11:52 AM, Hugh Cayless wrote:

...
Yeah, All works, but I think what’s happening is that in your ODD, the content model is being set to empty because the data.certainty and data.probability macros have been stripped,

How have they been stripped? That's what I don't really understand; I'm not doing anything to strip them (any more than teiall.odd is doing anything to include them).

...
and that because they’re only getting included via the att.global.responsibility reference, in which, because the datatype is a choice between two datatypes, the references aren’t resolved, meaning the macros aren’t included. This does seem like a bug, doesn’t it? I’ll just go ahead and fix it...

Thanks for that.

Cheers, Martin

...
...
On Nov 3, 2015, at 14:42 , Hugh Cayless wrote:

My guess would be that it’s getting in another way with tei_all. Will check though.

...
On Nov 3, 2015, at 14:40 , Martin Holmes wrote:

What I don't understand, then, is why teiall.odd in the tests generates a correct teiall.rng. If you run make test, then look at the generated Test/teiall.rng, do you see a broken content model?

Cheers, Martin

...
I’ve tracked it down, but I’m not 100% sure whether it’s actually a bug or whether we’re doing it wrong. The odd2odd.xsl checks to see if you have an rng:ref that’s the *child* of a <datatype>, but here we have a datatype that’s a choice between two datatypes (it’s the grandchild of <datatype>). The assumption seems to have been that datatypes couldn’t be defined by composition. Was that a conscious choice or just an oversight? If the latter, then I’ll check in a fix to the Stylesheets. If the former, then we need to define att.global.responsibility differently…

Thoughts?

...
On Nov 3, 2015, at 12:59 , Martin Holmes wrote:

I bet Lou and Syd would be able to tell us why. It's got me completely stumped.

Cheers, Martin

On 15-11-03 09:07 AM, Hugh Cayless wrote: > Yeah, I see the same thing... > > Sent from my phone. > >> On Nov 3, 2015, at 11:49, Martin Holmes >> wrote: >> >> I'm very puzzled by this. Could someone else generate an >> RNG from this ODD: >> >> http://web.uvic.ca/~mholmes/mariage.odd >> >> using the dev P5 and Stylesheets, and see if they get the >> same empty content model for @cert? >> >> Can anyone suggest why this might be happening? It >> doesn't happen in the regular P5 build using teiall.xml >> in Test. >> >> Cheers, Martin >> >>> On 15-11-03 07:56 AM, Martin Holmes wrote: Hi Raff, >>> >>> No, this is freshly-generated from an existing ODD, >>> using the latest bleeding-edge Oxygen plugin, meaning >>> that it's the current stylesheets and the current P5 >>> from dev. >>> >>> To build P5, you usually have to point it at the >>> location of your stylesheet checkout: >>> >>> XSL=/home/rviglian/Projects/tei-code/TEI/Stylesheets >>> >>> or whatever. >>> >>> Cheers, Martin >>> >>>> On 15-11-03 07:13 AM, Raffaele Viglianti wrote: Hi >>>> Martin, >>>> >>>> When I first made that change, I mistakenly pointed >>>> the ref to att.certainty and att.probability instead >>>> of data.certainty and data.probability. I fixed this >>>> after jenkins complained, but I imagine that the rng >>>> still got created and it might have looked like the >>>> one you included in your email. >>>> >>>> Is it possible that you're looking at an older RNG >>>> and/or need to rebuild? >>>> >>>> Also I've been trying to build locally to replicate >>>> your problem, but when I `make` in P5 I get this >>>> error: >>>> /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: >>>> >>>>

On 15-11-03 11:35 AM, Hugh Cayless wrote: stylesheet

...
...
>>>> /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl >>>> >>>> doesn't exist. Have I missed a step? >>>> >>>> Raff >>>> >>>>> On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes >>>>> wrote: >>>>> >>>>> Hi all, >>>>> >>>>> The @cert in att.global.responsibility now has a >>>>> janus value (att.certainty or att.probability). >>>>> However, when I process one of my existing ODDs >>>>> using the current dev tree, I get an empty value in >>>>> my RNG: >>>>> >>>>> <define >>>>> name="att.global.responsibility.attributes"> <ref >>>>> name="att.global.responsibility.attribute.cert"/> >>>>> <ref >>>>> name="att.global.responsibility.attribute.resp"/> >>>>> </define> <define >>>>> name="att.global.responsibility.attribute.cert"> >>>>> <optional> <attribute name="cert"> >>>>> http://relaxng.org/ns/compatibility/annotations/1.0 ">(certainty) >>>>> >>>>>

...
...
>>>>> signifies the degree of certainty associated with the intervention or >>>>> interpretation. <empty/> >>>>> </attribute> </optional> </define> <define >>>>> name="att.global.responsibility.attribute.resp"> >>>>> <optional> <attribute name="resp"> >>>>> http://relaxng.org/ns/compatibility/annotations/1.0 ">(responsible >>>>> >>>>> party) indicates the agency responsible for the >>>>> intervention or interpretation, for example an >>>>> editor or transcriber. <list> >>>>> <oneOrMore> <ref name="data.pointer"/> </oneOrMore> >>>>> </list> </attribute> </optional> </define> >>>>> >>>>> Any idea why this might be happening? >>>>> >>>>> Cheers, Martin -- tei-council mailing list >>>>> tei-council@lists.tei-c.org >>>>> http://lists.lists.tei-c.org/mailman/listinfo/tei-council >>>>> >>>>> >>>>>

...
...
>>>>> PLEASE NOTE: postings to this list are publicly archived >> -- tei-council mailing list tei-council@lists.tei-c.org >> http://lists.lists.tei-c.org/mailman/listinfo/tei-council >> >> >> PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

Hugh Cayless

1 p.m.

New subject: [tei-council] Oddity in RNG generation with new @cert datatype

Does that mean we’re setting ourselves up for pain in the near future by doing this now? What’s the right answer for something like global @cert?

...

On Nov 3, 2015, at 15:56 , Lou Burnard wrote:

This shouldn't happen in pure odd of course because we don't allow composite datatypes

Sent from my Samsung Galaxy Tab®|PRO

-------- Original message -------- From: Martin Holmes Date:11/03/2015 20:42 (GMT+00:00) To: tei-council@lists.tei-c.org Subject: Re: [tei-council] Oddity in RNG generation with new @cert datatype

Hi Hugh,

On 15-11-03 11:52 AM, Hugh Cayless wrote:

...
Yeah, All works, but I think what’s happening is that in your ODD, the content model is being set to empty because the data.certainty and data.probability macros have been stripped,

How have they been stripped? That's what I don't really understand; I'm not doing anything to strip them (any more than teiall.odd is doing anything to include them).

...
and that because they’re only getting included via the att.global.responsibility reference, in which, because the datatype is a choice between two datatypes, the references aren’t resolved, meaning the macros aren’t included. This does seem like a bug, doesn’t it? I’ll just go ahead and fix it...

Thanks for that.

Cheers, Martin

...
...
On Nov 3, 2015, at 14:42 , Hugh Cayless wrote:

My guess would be that it’s getting in another way with tei_all. Will check though.

...
On Nov 3, 2015, at 14:40 , Martin Holmes wrote:

What I don't understand, then, is why teiall.odd in the tests generates a correct teiall.rng. If you run make test, then look at the generated Test/teiall.rng, do you see a broken content model?

Cheers, Martin

...
I’ve tracked it down, but I’m not 100% sure whether it’s actually a bug or whether we’re doing it wrong. The odd2odd.xsl checks to see if you have an rng:ref that’s the *child* of a <datatype>, but here we have a datatype that’s a choice between two datatypes (it’s the grandchild of <datatype>). The assumption seems to have been that datatypes couldn’t be defined by composition. Was that a conscious choice or just an oversight? If the latter, then I’ll check in a fix to the Stylesheets. If the former, then we need to define att.global.responsibility differently…

Thoughts?

...
On Nov 3, 2015, at 12:59 , Martin Holmes wrote:

I bet Lou and Syd would be able to tell us why. It's got me completely stumped.

Cheers, Martin

On 15-11-03 09:07 AM, Hugh Cayless wrote: > Yeah, I see the same thing... > > Sent from my phone. > >> On Nov 3, 2015, at 11:49, Martin Holmes >> wrote: >> >> I'm very puzzled by this. Could someone else generate an >> RNG from this ODD: >> >> http://web.uvic.ca/~mholmes/mariage.odd >> >> using the dev P5 and Stylesheets, and see if they get the >> same empty content model for @cert? >> >> Can anyone suggest why this might be happening? It >> doesn't happen in the regular P5 build using teiall.xml >> in Test. >> >> Cheers, Martin >> >>> On 15-11-03 07:56 AM, Martin Holmes wrote: Hi Raff, >>> >>> No, this is freshly-generated from an existing ODD, >>> using the latest bleeding-edge Oxygen plugin, meaning >>> that it's the current stylesheets and the current P5 >>> from dev. >>> >>> To build P5, you usually have to point it at the >>> location of your stylesheet checkout: >>> >>> XSL=/home/rviglian/Projects/tei-code/TEI/Stylesheets >>> >>> or whatever. >>> >>> Cheers, Martin >>> >>>> On 15-11-03 07:13 AM, Raffaele Viglianti wrote: Hi >>>> Martin, >>>> >>>> When I first made that change, I mistakenly pointed >>>> the ref to att.certainty and att.probability instead >>>> of data.certainty and data.probability. I fixed this >>>> after jenkins complained, but I imagine that the rng >>>> still got created and it might have looked like the >>>> one you included in your email. >>>> >>>> Is it possible that you're looking at an older RNG >>>> and/or need to rebuild? >>>> >>>> Also I've been trying to build locally to replicate >>>> your problem, but when I `make` in P5 I get this >>>> error: >>>> /home/rviglian/Projects/tei-code/TEI/P5/antbuilder.xml:186: >>>> >>>>

On 15-11-03 11:35 AM, Hugh Cayless wrote: stylesheet

...
...
>>>> /home/rviglian/Projects/tei-code/TEI/P5/p5odds.isosch.xsl >>>> >>>> doesn't exist. Have I missed a step? >>>> >>>> Raff >>>> >>>>> On Mon, Nov 2, 2015 at 6:19 PM, Martin Holmes >>>>> wrote: >>>>> >>>>> Hi all, >>>>> >>>>> The @cert in att.global.responsibility now has a >>>>> janus value (att.certainty or att.probability). >>>>> However, when I process one of my existing ODDs >>>>> using the current dev tree, I get an empty value in >>>>> my RNG: >>>>> >>>>> <define >>>>> name="att.global.responsibility.attributes"> <ref >>>>> name="att.global.responsibility.attribute.cert"/> >>>>> <ref >>>>> name="att.global.responsibility.attribute.resp"/> >>>>> </define> <define >>>>> name="att.global.responsibility.attribute.cert"> >>>>> <optional> <attribute name="cert"> >>>>> http://relaxng.org/ns/compatibility/annotations/1.0">(certainty) >>>>> >>>>>

...
...
>>>>> signifies the degree of certainty associated with the intervention or >>>>> interpretation. <empty/> >>>>> </attribute> </optional> </define> <define >>>>> name="att.global.responsibility.attribute.resp"> >>>>> <optional> <attribute name="resp"> >>>>> http://relaxng.org/ns/compatibility/annotations/1.0">(responsible >>>>> >>>>> party) indicates the agency responsible for the >>>>> intervention or interpretation, for example an >>>>> editor or transcriber. <list> >>>>> <oneOrMore> <ref name="data.pointer"/> </oneOrMore> >>>>> </list> </attribute> </optional> </define> >>>>> >>>>> Any idea why this might be happening? >>>>> >>>>> Cheers, Martin -- tei-council mailing list >>>>> tei-council@lists.tei-c.org >>>>> http://lists.lists.tei-c.org/mailman/listinfo/tei-council >>>>> >>>>> >>>>>

...
...
>>>>> PLEASE NOTE: postings to this list are publicly archived >> -- tei-council mailing list tei-council@lists.tei-c.org >> http://lists.lists.tei-c.org/mailman/listinfo/tei-council >> >> >> PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

-- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived -- tei-council mailing list tei-council@lists.tei-c.org http://lists.lists.tei-c.org/mailman/listinfo/tei-council

PLEASE NOTE: postings to this list are publicly archived

3337

Age (days ago)

3338

Last active (days ago)

List overview

Download

16 comments

4 participants

participants (4)

Hugh Cayless
Lou Burnard
Martin Holmes
Raffaele Viglianti

Oddity in RNG generation with new @cert datatype

tags

participants (4)