xslide-list — Discussion list for xslide XSL mode for Emacs

(xslide) xslide lives!

[Tony G. <tg...@me...>]

FYI, and assuming both that your subscription still works and you're still
interested in xslide, I've just checked in a copy of my current version of
xslide which is now a derived mode derived from `nxml-mode'.

I'm looking to use the SourceForge-provided Trac for tracking issues with
this resuscitated xslide, though the way the Trac config is locked down
may mean its going to turn out to be unusable.  For the moment, the place
to report issues with xslide is:

https://sourceforge.net/apps/trac/xslide/newticket

Regards,


Tony.

(xslide) svn checkout fails

[Andreas R. <and...@on...>]

Hi,

thanks a lot developing xslide.

|svn co https://xslide.svn.sourceforge.net/svnroot/xslide xslide

results in an empty .svn repo, just some directories without content.

while

||cvs -z3 
-d:pserver:ano...@xs...:/cvsroot/xslide co xslide

did it.

Just to mention the issue. Thanks again.

Andreas
|

Re: (xslide) xslide-list on Gmane?

[Dave P. <dav...@gm...>]

2008/6/6 xslide Support <xsl...@me...>:
> I received a message asking if xslide-list could be archived on Gmane,
> but I managed to accidentally lose the message.
>
> The answer, if Steinar is listening, is that I have no objection, though
> we all know that the xslide-list has been less than active for several
> years.

You didn't introduce enough bugs Tony, then we could
have all shouted at you to keep the list busy :-)

It does what it says on the tin!
Still in use here very frequently!

Thanks.



-- 
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk

(xslide) xslide-list on Gmane?

[xslide S. <xsl...@me...>]

I received a message asking if xslide-list could be archived on Gmane,
but I managed to accidentally lose the message.

The answer, if Steinar is listening, is that I have no objection, though
we all know that the xslide-list has been less than active for several
years.

Regards,


Tony.

Re: (xslide) Selecting xslt 2 or 1

[Dave P. <dav...@gm...>]

A starter for ten.

(if (/=3D (point-min) (point-max))
    (if (y-or-n-p "Warning: loading file into non-empty buffer.
               yes to continue")
=09(progn
=09  (erase-buffer)
=09  (insert-file-contents xsl-initial-stylesheet-file)
=09  )
      )
  )


It should be relatively easy to mod the xsl-initial-stylesheet-file
from 1.0 to 2.0. format.
  Just leaves selecting one from 'tother.



regards
--
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk

Re: (xslide) Selecting xslt 2 or 1

[Dave P. <dav...@gm...>]

On 13/12/05, xslide Support <xsl...@me...> wrote:

> It seems to me that there are two separate problems that need solving:
>
>  - Whether to use XSLT 1.0 or XSLT 2.0 font-lock, abbrevs, etc. for a
>    particular .xsl file.
>
>  - What content to insert in a new .xsl file.
>
> The first can be solved by a regexp or function that identifies the
> "version" attribute (or its absence).

+1
Though there is a 90% overlap.

>
> The second can be solved by setting `xsl-initial-stylesheet-file' to
> point to the appropriate file.  Which would be fine for a person
> making a clean change from creating new XSLT 1.0 files to new XSLT 2.0
> files, but still not so good if you still need to create both.

This is the problem I was addressing.
I open emacs and want to create a new file.
I need to be able to choose 1 or 2.
Current xslide does it by file name extension.
I use .xsl for both 1.0 and 2.0



>
> Would it be better to support a list of template files, similar to the
> list of XSLT processors available to `xsl-process'?  The user would
> then have the option of choosing the best initial file for the job,
> e.g., XSLT 1.0 producing HTML, XSLT 2.0 producing XML, etc.

My *personal* choice is for output method=3D'xml'
yet with full html element, head content and body content,
also with a default template.
It's easier to delete content than to add it?

The real differentiator is
a) the xslt 2.0 namespaces
b) the version.
For myself, I'd be happy to have the remainder remain the same.

Thinks.
  How about a command line that makes the modifications to the header
of the file, rather than loading a choice in at load time?
 It's brittle, in that if I change the header, it could break.

I.e. Always load xslt 1.0 template.
M-x ??? something updates the file to reflect the 2.0 template
(or perhaps even loads a new template instead of the 1.0 one)
The latter sounds more user friendly.

Thoughts?





> > Is M-x xslt1-mode and M-x xslt2-mode
> > a reasonable approach?
>
> No, because it gets in the way of using the extension to recognise the
> file mode.

> p.s. I am on vacation from 15th December to 10th January, inclusive,
> and won't be responding to email.

Noted. Hope you have a nice break Tony.



regards
--
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk

Re: (xslide) Selecting xslt 2 or 1

[xslide S. <xsl...@me...>]

Dave Pawson <dav...@gm...> writes:
> I'm stuck trying to determine whether to start in 1.0 or 2.0.
>  My only solution is to use a different mode (minor mode perhaps?)
> for each.

It seems to me that there are two separate problems that need solving:

 - Whether to use XSLT 1.0 or XSLT 2.0 font-lock, abbrevs, etc. for a
   particular .xsl file.

 - What content to insert in a new .xsl file.

The first can be solved by a regexp or function that identifies the
"version" attribute (or its absence).

The second can be solved by setting `xsl-initial-stylesheet-file' to
point to the appropriate file.  Which would be fine for a person
making a clean change from creating new XSLT 1.0 files to new XSLT 2.0
files, but still not so good if you still need to create both.

Would it be better to support a list of template files, similar to the
list of XSLT processors available to `xsl-process'?  The user would
then have the option of choosing the best initial file for the job,
e.g., XSLT 1.0 producing HTML, XSLT 2.0 producing XML, etc.

> I.e. No point in searching within the initial file, since it has been
> determined at
> that time?
>
> I don't really want to use xyz.xsl1 and xyz.xsl2 i.e. depend on the filename.
>
> Is M-x xslt1-mode and M-x xslt2-mode
> a reasonable approach?

No, because it gets in the way of using the extension to recognise the
file mode.

Regards,


Tony.

p.s. I am on vacation from 15th December to 10th January, inclusive,
and won't be responding to email.

(xslide) Selecting xslt 2 or 1

[Dave P. <dav...@gm...>]

I'm stuck trying to determine whether to start in 1.0 or 2.0.
 My only solution is to use a different mode (minor mode perhaps?)
for each.


I.e. No point in searching within the initial file, since it has been
determined at
that time?

I don't really want to use xyz.xsl1 and xyz.xsl2 i.e. depend on the filenam=
e.


Is M-x xslt1-mode and M-x xslt2-mode
a reasonable approach?

TIA
--
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk

Re: (xslide) xslt 2.0. Determine version

[xslide S. <xsl...@me...>]

Dave Pawson <dav...@gm...> writes:
>> Please provide your empty XSLT 2.0 stylesheet.

Checked in as xslt2-initial.xsl.

Regards,


Tony.

Re: (xslide) xslt 2.0. Determine version

[Dave P. <dav...@gm...>]

On 09/12/05, xslide Support <xsl...@me...> wrote:
> Please provide your empty XSLT 2.0 stylesheet.


<?xml version=3D"1.0" encoding=3D"utf-8"?>
<xsl:stylesheet xmlns:xsl=3D"http://www.w3.org/1999/XSL/Transform"
xmlns:xs=3D"http://www.w3.org/2001/XMLSchema"
xmlns:xsi=3D"http://www.w3.org/2001/XMLSchema-instance"
exclude-result-prefixes=3D"xs xsi"
 version=3D"2.0">


   <xsl:output method=3D"xml" indent=3D"yes" encoding=3D"utf-8"/>

   <xsl:template name=3D"*">
<xsl:message>
*** Failed to process <xsl:value-of select=3D"name()"/> ***
</xsl:message>
   </xsl:template>

</xsl:stylesheet>




--
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk

Re: (xslide) xslt 2.0. Determine version

[xslide S. <xsl...@me...>]

Dave Pawson <dav...@gm...> writes:
...
> I'll try and use this to load either a 1.0 or 2.0 'empty' stylesheet,
> then set the current variables to a list of either 1.0 or 2.0
> elements and attributes.

Please provide your empty XSLT 2.0 stylesheet.

Regards,


Tony.

Re: (xslide) xslt 2.0. Determine version

[Dave P. <dav...@gm...>]

Tested with and without the xml decl in the file.
Defaults to 1.0 (if missing too)

I'll try and use this to load either a 1.0 or 2.0 'empty' stylesheet,
then set the current variables to a list of either 1.0 or 2.0
elements and attributes.

(defvar xsl-ver 1.0)

=09
(defun fv()
"Find the xslt version"
(interactive)
(save-excursion
  (goto-char (point-min))
  (skip-chars-forward "\\s-")
  (cond
   ((looking-at "<\\?xml");; 2nd time if xmldcl present
    (setq p (re-search-forward "version=3D\"" nil t 2))
    (buffer-substring-no-properties p (+ p 3)))
    (t
    (setq p (re-search-forward "version=3D\"" nil t 1))
    (buffer-substring-no-properties p (+ p 3)))
    )
  )
)
(setq xsl-ver (if (fv) (fv) 1.0 ))

regards

--
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk

(xslide) xslt 2.0

[Dave P. <dav...@gm...>]

I've started working on a version of xslide for xslt 2.0

Is anyone interested in helping please?
regards
--
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk

Re: (xslide) xslide update to glens_branch in CVS

[Glen P. <gl...@or...>]

I found a bug in the <script> tag indentaion that I checked in a few days a=
go and I fixed it.  It's ready for someone besides me to try out.=0A---
Glen Peterson
Senior Software Engineer, Web Consultant
South Hamilton, MA USA


=0A

(xslide) Screenshot

[Glen P. <gl...@or...>]

Tony:  I'm attaching a screenshot of XSL and another of JSP for the screens=
hots page of this project.  I don't have rights to post them.

They are big.  I can make smaller ones.  What's the chance of seeing these =
on the Screenshots page?=0A---
Glen Peterson
Senior Software Engineer, Web Consultant
South Hamilton, MA USA


=0A

(xslide) xslide update to glens_branch in CVS

[Glen P. <gl...@or...>]

- Generalized fontification to work almost as well in jsp as in xsl.

- Fixed bug where a script tag within a comment could mess up indentation.

Not much else.  I made the second fix right before checking these files in.=
  Otherwise, I have been using this to edit jsp and xsl for about 3 months =
without finding any other issues that I cared about.  I'll have to test-dri=
ve the "script tag in comment" fix a bit and try this out in GNU emacs to m=
ake sure I haven't broken anything.

=0A---
Glen Peterson
Senior Software Engineer, Web Consultant
South Hamilton, MA USA


=0A

Re: (xslide) JSP support anyone?

[xslide S. <xsl...@me...>]

Wow, xslide-list traffic!

At 24 Jul 2004 10:49 -0400, Glen Peterson wrote:
 > Why add JSP support to xslide?  Because I started writing JSP at work
 > and crackhead-mode, though interestingly named, was further from what I
 > wanted than xslide.  So I did it.  You can check it out in glens_branch
 > of the CVS tree (under xslide).  If you think this is a bad idea, I can
 > start a separate project.  I think they actually go very well together. 
 > Just the name, xslide, doesn't really cover JSP.

I wrote an 'Ant' mode, currently called 'ante', by hacking a copy of
xslide.  I did it mostly for the imenu support so I could find targets
and subtargets by looking them up in a menu (or in a window when uses
the Emacs Code Browser).

 > Anyway, I added some fontification, and some auto-completion.  Tony, I
 > know you wanted me to use a data structure for auto-completion, but I
 > tried and it was really difficult.  Especially since there are so many
 > different combinations: optional parameters that when used with a tag
 > require it to be a quick close, or optional parameters that require
 > other parameters, or values of certain parameters that require other
 > parameters.  So I've given up on that approach and just hard-coded
 > everything.  Code is a data structure... of sorts.  :-)
 > 
 > I don't think it's quite ready for an official release, but what would
 > be required to do a release, Tony?

I'll try to find time to look at it.  Right now, I have to work on the
monthly release of xmlroff (after missing last month's all together).

I think the answer is to pick a day within the month and ship whatever
we have on or about that date.  I was going a long time between
xmlroff releases because I was trying to get it somehow more stable,
but the end result was that I seldom made releases.  Up until last
month, I made four or so monthly releases in a row because I said I'd
do a release on or about the first of the month.  Consider me an early
convert but a late practitioner of 'release early, release often'.

Regards,


Tony.

(xslide) JSP support anyone?

[Glen P. <gl...@or...>]

Why add JSP support to xslide?  Because I started writing JSP at work
and crackhead-mode, though interestingly named, was further from what I
wanted than xslide.  So I did it.  You can check it out in glens_branch
of the CVS tree (under xslide).  If you think this is a bad idea, I can
start a separate project.  I think they actually go very well together. 
Just the name, xslide, doesn't really cover JSP.

Anyway, I added some fontification, and some auto-completion.  Tony, I
know you wanted me to use a data structure for auto-completion, but I
tried and it was really difficult.  Especially since there are so many
different combinations: optional parameters that when used with a tag
require it to be a quick close, or optional parameters that require
other parameters, or values of certain parameters that require other
parameters.  So I've given up on that approach and just hard-coded
everything.  Code is a data structure... of sorts.  :-)

I don't think it's quite ready for an official release, but what would
be required to do a release, Tony?

(xslide) Comments in XSL files

[Patrick P. <pp...@fr...>]

Hi,

I have  uncommented and modified  those lines in xslide.el  (they were
nice for Lisp code, but not for XSL).

;;  (make-local-variable 'comment-start)
;;  (setq comment-start "; ")
;;  (make-local-variable 'comment-end)
;;  (setq comment-end "\n")

  (make-local-variable 'comment-start)
  (setq comment-start "<!-- ")
  (make-local-variable 'comment-end)
  (setq comment-end " -->")

It enables quick commenting/uncommenting of portions of code using M-;
(usually mapped to comment-dwim or comment-region). 

A+
PP
-- 
Groupe Morbihannais d'Utilisateurs de Logiciels Libres http://www.tuxbihan.org
Identifiant Jabber: pp...@am...
GPG fingerprint = 1A4F E154 3D2C A20E E4CA  A543 7951 C5C2 E44A A0B5

Patrick Percot.

Re: (xslide) Bug in my new indenting code

[xslide S. <xsl...@me...>]

At 26 Nov 2003 20:29 -0400, xslide Support wrote:
...
 > It appears that texinfo has its own XML format, but it also appears
 > that's only an output format from makeinfo, not an input format, and
 > the sample that I looked at looked closer to real texinfo and,
 > therefore, harder to generate than the DocBook2X intermediate format.

Yep, only an output format (and no mention of a DTD/schema):

http://www.gnu.org/software/texinfo/manual/texinfo/html_node/Output-Formats.html#Output%20Formats

Regards,


Tony Graham.

Re: (xslide) Bug in my new indenting code

[xslide S. <xsl...@me...>]

At 19 Nov 2003 22:14 -0500, Glen Peterson wrote:
...
 > slow on one-line xml files over 500K.  That's fixed now too.  I use
 > xsl-invasive-indent several times a day most days, but I doubt anyone
 > will even know it exists if they don't read this post.  How should I
 > advertise it (or any other function of xslide for that matter)?  Is
 > there any kind of xslide manual?  Where will users look first for this
 > sort of information?

A little prodding now and then is a wonderful thing.

I've finally looked at DocBook2X, and it seems to do a good job of
turning DocBook XML into texinfo files.

Since it uses an intermediate close-to-texinfo XML format, I'm also
making good progress on a stylesheet for transforming the XSL spec XML
into the intermediate XML format.  It's unfortunate that the
intermediate format is undocumented.

The stylesheet should also be useable for the XSLT and XPath specs as
well.

The other alternative was a Wiki, specifically an EmacsWiki Wiki, that
could be turned into texinfo, but that wouldn't have also allowed the
transforming of the W3C specs into texinfo the way that DocBook2X
does.

It appears that texinfo has its own XML format, but it also appears
that's only an output format from makeinfo, not an input format, and
the sample that I looked at looked closer to real texinfo and,
therefore, harder to generate than the DocBook2X intermediate format.

Regards,


Tony Graham.

(xslide) Bug in my new indenting code

[Glen P. <gl...@or...>]

Bug #760248 by pbw is relevant to my new indenting code:

A line that begins with an improperly indented </xsl:text> outside a
comment or CDATA section will mess up subsequent indenting.

I fixed this in my working copy, but I have made a bunch of other
changes (new table of tags) and reworked the indenting code for
efficiency without changing it's function (except by fixing this bug). 
Why efficiency in indenting?  xsl-invasive-indent used to get unusably
slow on one-line xml files over 500K.  That's fixed now too.  I use
xsl-invasive-indent several times a day most days, but I doubt anyone
will even know it exists if they don't read this post.  How should I
advertise it (or any other function of xslide for that matter)?  Is
there any kind of xslide manual?  Where will users look first for this
sort of information?

I want to make a branch for the changes I have made since my last
check-in so that I have a backup of them before checking out a working
copy to fix this bug.  I have added a file (for auto-completion
functions) and will probably first add it to the branch.  Will CVS work
properly when I want to merge this new file into the trunk?

Maybe I'll post my backup as a patch (as I used to do) and fix this
problem, then worry about CVS afterward.  Ugly, but effective.

Re: (xslide) xslide Patch #813124: Add msxsl.exe support in xsl-process.el

[xslide S. <xsl...@me...>]

At 7 Nov 2003 21:39 -0500, Glen Peterson wrote:
 > To: Ville Skytt=E4 (scop) or anyone familiar with xsl-process.el
 >=20
 > Looking in the changelog, it seems that you (Ville) did a bunch of the
 > work on xslide-process.el.  Doug Bloebaum (blabes) submitted a patch t=
o

Ville contributed a patch, which I installed.

 > this file and I am confused about some of the code that he changed.  H=
is
 > patch is at:
 >=20
 > https://sourceforge.net/tracker/index.php?func=3Ddetail&aid=3D813124&g=
roup_id=3D22964&atid=3D377145
 >=20
 > He deleted and replaced the following code with something that works o=
n
 > msxml but he *may* have broken some other processors in the process.=20
 > I'm not sure how I would go about regression testing this because I
 > don't know which processors use this code.  Do some of the processors

That is the problem with trying to get xslide to support multiple XSLT
processors.  I usually just install any patches for XSLT processors
that I don't use myself since I have no easy way of testing them.  If
it breaks using xslide with a processor that I do use, then I'll fix
it.

 > that have their own specific code also use this "generic" code?  The
 > section of code I am talking about looks like this (before his patch)
 >=20
 > ;; Generic
 > (".*\\<file:\\(\\(/[A-Za-z]:\\)?[^:]+\\):[ \t]*\\(\\([0-9]+\\):[
 > \t]*\\(\\([0-9]+\\):\\)?\\)?"
 > 1 4 6)
 > (".*\\([^ ]+\\.x[ms]l\\):[ \t]*\\([0-9]+\\):[ \t]*\\(\\([0-9]+\\):\\)?=
"
 > 1 2 4)
 > ("^\\([^:]+\\): \\([0-9]+\\): error:"
 > 1 2)
 >=20
 > Also, I'm unclear about how the user selects which processor they want
 > to use.  Xt comes up by default.  If I want to use java apache xalan,

Hopefully C-h k C-c C-p explains it well enough.

 > what would you expect me to do?  I got it to work, but I'm sure I didn=
't
 > do it the recommended way.  It was really cool that it remembered what=
 I
 > wanted after I got it to work.

Command histories are your friend.

 > Do you have some time to review his patch and/or answer my questions,
 > and/or add comments or patches to the code to answer these questions? =
 I
 > will not merge Doug's patch until I get this figured out.  Anyone else
 > want to shed light on the subject?

I'll look at it.

Regards,


Tony Graham.

(xslide) Re: xslide Patch #813124: Add msxsl.exe support in xsl-process.el

[Ville <sc...@xe...>]

On Sat, 2003-11-08 at 04:39, Glen Peterson wrote:

> https://sourceforge.net/tracker/index.php?func=detail&aid=813124&group_id=22964&atid=377145
> 
> He deleted and replaced the following code with something that works on
> msxml but he *may* have broken some other processors in the process.

Yes, it is possible that this breaks something.

> I'm not sure how I would go about regression testing this because I
> don't know which processors use this code.  Do some of the processors
> that have their own specific code also use this "generic" code?

I don't know.

> Also, I'm unclear about how the user selects which processor they want
> to use.  Xt comes up by default.  If I want to use java apache xalan,
> what would you expect me to do?

The way I do is to invoke xsl-process, hit enter a couple of times to
accept the filenames, and then in the "Process command:" prompt hit
up-arrow until the one I prefer is there.  The order where the command
lines appear can be modified by customizing xsl-process-command.

> Do you have some time to review his patch and/or answer my questions,
> and/or add comments or patches to the code to answer these questions?  I
> will not merge Doug's patch until I get this figured out.  Anyone else
> want to shed light on the subject?

I took a look at the patch, some notes:
- Uses windows linebreaks (they show up as ^M here)
- For multiline regexps, I'd recommend taking a look how the jd.xslt 
  regexp is done instead of relying on the linebreaks.

While at it, I fixed some of the regexps so that they work here using
xsltproc and  Xalan from Java 1.4.2.  Patch attached (taken against
XEmacs CVS but I believe it should apply to vanilla xslide too).

(xslide) xslide Patch #813124: Add msxsl.exe support in xsl-process.el

[Glen P. <gl...@or...>]

To: Ville Skytt=E4 (scop) or anyone familiar with xsl-process.el

Looking in the changelog, it seems that you (Ville) did a bunch of the
work on xslide-process.el.  Doug Bloebaum (blabes) submitted a patch to
this file and I am confused about some of the code that he changed.  His
patch is at:

https://sourceforge.net/tracker/index.php?func=3Ddetail&aid=3D813124&grou=
p_id=3D22964&atid=3D377145

He deleted and replaced the following code with something that works on
msxml but he *may* have broken some other processors in the process.=20
I'm not sure how I would go about regression testing this because I
don't know which processors use this code.  Do some of the processors
that have their own specific code also use this "generic" code?  The
section of code I am talking about looks like this (before his patch)

;; Generic
(".*\\<file:\\(\\(/[A-Za-z]:\\)?[^:]+\\):[ \t]*\\(\\([0-9]+\\):[
\t]*\\(\\([0-9]+\\):\\)?\\)?"
1 4 6)
(".*\\([^ ]+\\.x[ms]l\\):[ \t]*\\([0-9]+\\):[ \t]*\\(\\([0-9]+\\):\\)?"
1 2 4)
("^\\([^:]+\\): \\([0-9]+\\): error:"
1 2)

Also, I'm unclear about how the user selects which processor they want
to use.  Xt comes up by default.  If I want to use java apache xalan,
what would you expect me to do?  I got it to work, but I'm sure I didn't
do it the recommended way.  It was really cool that it remembered what I
wanted after I got it to work.

Do you have some time to review his patch and/or answer my questions,
and/or add comments or patches to the code to answer these questions?  I
will not merge Doug's patch until I get this figured out.  Anyone else
want to shed light on the subject?

- Glen Peterson

P.S.  No-one responded to my previous email about when to release, so
unless anyone objects, I'll cut a release first, then merge this or any
other changes.