Re: [xsl] Ungready Flag on Regex

First, I've try to change all xsl:non-matching-substring TVT withxsl:sequence, and that was worst.

Then, I've changed only pt:replaceTextDollar xsl:non-matcing-substringTVT to a xsl:sequence, without success.

The problem was my test was not correctly written. I now have writtenall test cases, testing direct calls, nested calls, and so on. Thanks alot, it now works.

@Martin : I apologize for being too quick, and not precise when applyingyour suggestion ; it was late, and the curfew news had a bad effect on me.

@Gerrit, nesting these markups have real use cases. For example, in aGIT course, if I want to explain the bisect's steps, I'll write sometinglike :


$ git bisect *start*
$ git bisect *bad*
$ git bisect *good*905d48bb
$ git bisect *run*run-xspec.sh
...
$ git bisect *reset*

So a $$\$ git bisect **start**$$ ...

Or, if I explain the merge-request process, I want to highlight the URLreturn by the server :


$ *git push origin issues/3*
Énumération des objets: 37, fait.
Décompte des objets: 100% (37/37), fait.
Compression par delta en utilisant jusqu'à 4 fils d'exécution
Compression des objets: 100% (11/11), fait.
Écriture des objets: 100% (20/20), 3.21 Kio | 1.60 Mio/s, fait.
Total 20 (delta 5), réutilisés 0 (delta 0), réutilisés du pack 0
remote:
remote: To create a merge request for issues/3, visit:

remote:https://gitlab.oxiane.org/formation/tools/oxslides-java/-/merge_requests/new?merge_request%5Bsource_branch%5D=issues%2F3

remote:
To https://gitlab.oxiane.org/formation/tools/oxslides-java.git
 * [new branch]      issues/3 -> issues/3

I'll probably never have a code or a bold nested inside a link, but maybe...

Be safe,
Christophe
**

Le 23/10/2020 à 03:01, Imsieke, Gerrit, le-tex gerrit(_dot_)imsieke(_at_)le-tex(_dot_)dea écrit :

Christophe,

Have you tried Martin's solution?
It solves the issue and it doesn't create atomic values instead oftext nodes.
You need to see that you've been smashing the results of the innerfunction invocations to text nodes, no matter whether elements havebeen created by these functions in the first place.
Martin hasn't changed xsl:non-matching-substring in the innermostfunction pt:replaceTextArobase() so it still produces a text node.xsl:matching-substring in pt:replaceTextArobase() is unchanged. Itcreates a text:span element with a text node in it.
Previously, whatever came out of pt:replaceTextArobase() was cast to atext node by the text value template in pt:replaceTextDollar().The use of xsl:sequence instead of the text value template just makessure that whatever pt:replaceTextArobase() returned will be returnedin the xsl:non-matching substring branch. If it is a text:spanelement, the result will be a text:span element, if it is a text node,the result will also be a text node. No casting to atomic values occurs.
Maybe you picture xsl:sequence as doing some reverse smashingoperation: While xsl:value-of smashes atomic values to text nodes, youmight think that xsl:sequence smashes text nodes to atomic values. Itdoes not.
Gerrit
P.S.: Looking forward to the future requirement that the ** $$ @@pseudo markup should be able to nest.
On 23.10.2020 00:50, Christophe Marchand cmarchand(_at_)oxiane(_dot_)com wrote:
That's not the problem. Changing to <xsl:sequence .../> is worse, itcreates some pseudo-atomic-values instead of text() nodes.
Christophe

Le 22/10/2020 à 23:48, Martin Honnen martin(_dot_)honnen(_at_)gmx(_dot_)de a 
écrit :
On 22.10.2020 23:14, Christophe Marchand cmarchand(_at_)oxiane(_dot_)com wrote:
For sure, regex are strange to me, even if I think they are reallyusefull.
So, when a regex king give me the right regex to do exactly what Iwant,
I take this regex, and I adapt it to other similar use-cases.
My first problem was with comments, and I've adapted it to otherkind of
markers :

**bold**
$$code$$
@@link@@

The first two are Ok, but I can not make it run for the third case.
I join the XSL, a XSpec file, I have one scenario that fails. Ifsomeone
could :

 1. give me the solution
 2. explain me why it was failing
 3. explain me why the solution works

it would be really a good thing.
In France, we now have a curfew at 9pm. There a lot of beer thatcan not
be sold ; i may have a barrel for the guy who help me !

Best regards,
Christophe
Le 08/10/2020 à 16:54, Imsieke, Gerrit, le-texgerrit(_dot_)imsieke(_at_)le-tex(_dot_)de
a écrit :
On 08.10.2020 16:26, Liam R. E. Quin liam(_at_)fromoldbooks(_dot_)org wrote:
You want [^*]*?
Thinking of the brilliant potential for misunderstanding that this
answer entails.

Christophe: I don't understand why you are asking me whether I want
[^*]*.

Jokes aside, I think you need the non-greedy question mark modifier,
but not on [^*]*. You use it to make .* non-greedy. Otherwise it will
stop at the first '*' that it encounters.

This works:

<xsl:template name="xsl:initial-template">
  <doc>
    <xsl:analyze-string
      select="'Comments are delimited by **/* */** or prefixed by
**//**'"
      regex="\*\*(.*?)\*\*">
      <xsl:matching-substring>
        <b>
          <xsl:value-of select="regex-group(1)"/>
        </b>
      </xsl:matching-substring>
      <xsl:non-matching-substring>
        <xsl:value-of select="."/>
      </xsl:non-matching-substring>
    </xsl:analyze-string>
  </doc>
</xsl:template>
I think the issue is your use of text value templates
<xsl:non-matching-substring>{pt:replaceTextArobase(.)}</xsl:non-matching-substring>
that strips any elements; you need to use
  <xsl:non-matching-substring>
    <xsl:sequence select="pt:replaceTextArobase(.)"/>
  </xsl:non-matching-substring>

--~----------------------------------------------------------------
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
EasyUnsubscribe: http://lists.mulberrytech.com/unsub/xsl-list/1167547
or by email: xsl-list-unsub(_at_)lists(_dot_)mulberrytech(_dot_)com
--~--