Mariage

At LSPW's request, added a handler for the <gloss> tag, and for <quote> tags not parented by< cit>s. Updated all files in the db, and discussed some of many markup issues with La Forest nuptiale.

EDIT by MDH 2011-12-22: Implemented as documented in this post. Setting to Completed.

In working on the Forest nuptiale, I've noticed an interesting issue that we may need to spend some serious time working through:

After the "Avant-propos" and a sonnet dedicated to the reader from the author, we come to a "Catalogue des autheurs desquels on a tiré ce discours" page that lists all the authors that inspired this work. The first one, "Agathie" has a drop-cap "A". The list is in two columns.

In her encoding of the list, emurphy used a <list type="simple">; each name was made an <item>. She also attempted to make a column break by using a <cb/>.

emurphy has used this same format for the "Tables des chapitres" that occurs on the page following the "Catalogue des autheurs". In the "Tables des chapitres", there are also two columns, and the first item in the left column, "Romains", has a drop-cap.

Please let me know if you need any more details :)

There's sometimes a need to define a text-align setting to something which can only be enclosed in a <hi> tag, because of the surrounding context (for instance, where you cannot use a head tag because the parent element is a <quote> tag). hi elements give rise to span elements in the XHTML, and these default to display: inline, so I've put in a trap which detects when a text-align setting is applied to a hi tag and adds display: block to the style value. This appears to be working for the examples I know of. At some point there may be more CSS properties that need to act as triggers for this behaviour, and there may also be situations where display: block is not appropriate (such as when an element is floated), so I'll need to watch out for reports of oddities with this.

LSPW reported the Sonnet 1609 was not rendering, due to an ambiguous template match; this turned out to be caused by a link to an external source, whose path mimicked one of our internal paths. I've now made the internal path match much more specific, so the match is much more unlikely. It's still not impossible, though -- I should probably come back to this.

Small change on CC's instructions.

Further to my last email, these are the files that need to be updated because they contain the old <note type="link">.

Image files:

• agresseure
• balance
• bonheur_des_uns
• chat
• corrige_si_tu_peux
• invention_femmes_1
• invention_femmes_2
• grande_destruction
• le_bon_menage
• martires
• massacre_lagniet
• massacre_larmessin
• operateur_cephalique

Text documents:

• advis_salutaire
• purgatoire
• reconfort_des_femmes

Way back at the beginning of the project, we had a system for creating multiple links between documents. That's no longer functioning, because we now have a more sophisticated linking system based on the new db structure. However, the old links still persist in some of the original documents, and they're not working properly. This is what needs to be done:

• Find all instances of <note type="link">. Typically, these contain a block of text surrounded by a <ref> tag with multiple targets. The text needs to be rebuilt with more logical links made from ref tags around specific mentions of individual targets in the text.
• Many of the original links pointed to specific locations in the target documents based on ids. These ids were assigned to the ref tags originally, but that really doesn't make much sense in the case of engravings; we should be linking to the target annotation using the formula ann_imtArea_10. Those targets should be updated appropriately. In the case of items in the text documents, it's OK to give a unique id to the <ref> tag, but that will need to be passed through by the XSLT (which I don't think is happening right now).
• Finally, we need to document this on the blog with step-by-step instructions.

I've started this process with some edits to <il_cherche.xml> and <la_femme_battant.xml>, but haven't finished them yet.

LCC reported a problem with the Candale texts, which showed up recently -- we don't know what actually caused it to appear now, since the relevant part of the text has apparently not changed, but his is the bug profile:

The content of the document has this:

<fw type="sig" place="bot-center" rend="text-align: center; 
           margin-left: auto; margin-right: auto;">I. Part.<space quantity="4" unit="em"/>A</fw>

This is the XSLT which creates the error:

<xsl:variable name="fwContent" select="normalize-space(./text())"/>

The input to normalize-space() becomes a sequence of text nodes, which is not allowed; you get the error "A sequence of more than one item is not allowed as the first argument of normalize-space()". This is a working solution:

<xsl:variable name="fwContent"><xsl:value-of separator=" " select="normalize-space()"/></xsl:variable>

"Wrong-Headed Spouses" is now on the site, and has been added to the articles.xml TOC file. Also added some tweaks to heading display in the bibliography.

Fixed a couple of broken links to biblio items and documents in the articles. Also added a title attribute to biblio items which shows the xml:id of the item, to make it easier for encoders to link to a biblio item without having to read the XML of the biblio file.