subject:"Re\: Patch Submission\: LyxBlogger Converters"

El Sat, 27 Mar 2010 18:07:04 -0500
Jack Desert jackdesert...@gmail.com escribió:
 El Sat, 27 Mar 2010 18:03:45 -0400
 rgheck rgh...@bobjweil.com escribió:
  On 03/27/2010 03:46 PM, Jack Desert wrote:
  
   If eLyXer is installed, it also adds:
\converter html   blog   lyxblogger $$i   
  
  
  You can allow it to add this even if elyxer isn't installed. Or is there 
  some reason we want to make sure that lyxblogger only runs on elyxer's 
  output and not, say, on latex2html's?
 
 LyxBlogger is specifically set up to parse the LyXHTML and eLyXer formats. 
 Some features may turn out strange using TtH, Hevea, or latex2html, simply 
 because the regular expressions may not find what they're looking for.
 
 So yes, let's leave the html - blog converter inside the if block so a 
 newbie doesn't get unexpected results right out of the gate.
 
 -Jack
 

Richard,

Would you consider adding a META tag to the LyXHTML output? That will make it 
more predictable to tell which xhtml format is being used. Some examples from 
other converters:

META name=GENERATOR content=hevea 1.10

meta name=GENERATOR content=TtH 3.85

-Jack


-- 
~~~
Jack Desert --Writer, Entrepeneur
Author and Spokesman: www.LetsEATalready.com
Software Developer:   http://GrooveTask.org
Email: jackdesert...@gmail.com
~~~

Re: Patch Submission: LyxBlogger Converters

2010-04-05 Thread rgheck


On 04/05/2010 05:14 PM, Jack Desert wrote:

Would you consider adding a META tag to the LyXHTML output? That will make it 
more predictable to tell which xhtml format is being used. Some examples from 
other converters:

META name=GENERATOR content=hevea 1.10

meta name=GENERATOR content=TtH 3.85

   

Good idea. It'll output something like:
meta name=GENERATOR content=LyX 2.0.0svn /
where the content is determined by the PACKAGE_STRING macro in config.h 
and so can change. I assume it will always begin LyX, though.


rh

Re: Patch Submission: LyxBlogger Converters


 Good idea. It'll output something like:
 meta name=GENERATOR content=LyX 2.0.0svn /
 where the content is determined by the PACKAGE_STRING macro in config.h 
 and so can change. I assume it will always begin LyX, though.
 
 rh
 

Sounds good. Let me know when it's finalized and I'll add it to LyXBlogger.

-Jack

-- 
~~~
Jack Desert --Writer, Entrepeneur
Author and Spokesman: www.LetsEATalready.com
Software Developer:   http://GrooveTask.org
Email: jackdesert...@gmail.com
~~~

Re: Patch Submission: LyxBlogger Converters

2010-04-05 Thread Richard Heck


On 04/05/2010 06:37 PM, Jack Desert wrote:
   

Good idea. It'll output something like:
meta name=GENERATOR content=LyX 2.0.0svn /
where the content is determined by the PACKAGE_STRING macro in config.h
and so can change. I assume it will always begin LyX, though.

rh

 

Sounds good. Let me know when it's finalized and I'll add it to LyXBlogger.

   
It's in. The exact string will change with different releases, e.g., at 
some point it will be LyX 2.0.0, etc.


rh

Re: Patch Submission: LyxBlogger Converters

El Sat, 27 Mar 2010 18:07:04 -0500
Jack Desert  escribió:
> El Sat, 27 Mar 2010 18:03:45 -0400
> rgheck  escribió:
> > On 03/27/2010 03:46 PM, Jack Desert wrote:
> > >
> > > If eLyXer is installed, it also adds:
> > >  \converter html   blog   "lyxblogger $$i"   ""
> > >
> > >
> > You can allow it to add this even if elyxer isn't installed. Or is there 
> > some reason we want to make sure that lyxblogger only runs on elyxer's 
> > output and not, say, on latex2html's?
> 
> LyxBlogger is specifically set up to parse the LyXHTML and eLyXer formats. 
> Some features may turn out strange using TtH, Hevea, or latex2html, simply 
> because the regular expressions may not find what they're looking for.
> 
> So yes, let's leave the html -> blog converter inside the "if" block so a 
> newbie doesn't get unexpected results right out of the gate.
> 
> -Jack
> 

Richard,

Would you consider adding a META tag to the LyXHTML output? That will make it 
more predictable to tell which xhtml format is being used. Some examples from 
other converters:





-Jack


-- 
~~~
Jack Desert --Writer, Entrepeneur
Author and Spokesman: www.LetsEATalready.com
Software Developer:   http://GrooveTask.org
Email: jackdesert...@gmail.com
~~~

Re: Patch Submission: LyxBlogger Converters

2010-04-05 Thread rgheck


On 04/05/2010 05:14 PM, Jack Desert wrote:

Would you consider adding a META tag to the LyXHTML output? That will make it 
more predictable to tell which xhtml format is being used. Some examples from 
other converters:





   

Good idea. It'll output something like:

where the content is determined by the PACKAGE_STRING macro in config.h 
and so can change. I assume it will always begin "LyX", though.


rh

Re: Patch Submission: LyxBlogger Converters


> Good idea. It'll output something like:
> 
> where the content is determined by the PACKAGE_STRING macro in config.h 
> and so can change. I assume it will always begin "LyX", though.
> 
> rh
> 

Sounds good. Let me know when it's finalized and I'll add it to LyXBlogger.

-Jack

-- 
~~~
Jack Desert --Writer, Entrepeneur
Author and Spokesman: www.LetsEATalready.com
Software Developer:   http://GrooveTask.org
Email: jackdesert...@gmail.com
~~~

Re: Patch Submission: LyxBlogger Converters

2010-04-05 Thread Richard Heck


On 04/05/2010 06:37 PM, Jack Desert wrote:
   

Good idea. It'll output something like:

where the content is determined by the PACKAGE_STRING macro in config.h
and so can change. I assume it will always begin "LyX", though.

rh

 

Sounds good. Let me know when it's finalized and I'll add it to LyXBlogger.

   
It's in. The exact string will change with different releases, e.g., at 
some point it will be "LyX 2.0.0", etc.


rh

Re: Patch Submission: LyxBlogger Converters

2010-04-02 Thread Jürgen Spitzmüller

rgheck wrote:
 On 03/30/2010 04:15 PM, Jack Desert wrote:
  Actually, the first issue is that, especially since 2.0 will (hopefully)
  let the user choose conversion paths, etc, we can go ahead and install
  both the xhtml-blog and html-blog converters. This is particularly
  important in 2.0, it seems to me, since otherwise people who happen to
  have had elyxer installed for use with 1.6.x would not then see that
  they have a new option in 2.0.x.
  
  When can we reasonably expect to see this user chooses his own
  conversion paths code? Which contributors are currently working on it?
  Is there a branch somewhere that at least has the GUI so we can see what
  it looks like?
 
 Jurgen did a bunch of work on it but ran into a problem. I've told him
 I'll help him sort it out and did some of the necessary background work.
 But then he got busy and otherwise distracted, so it's kind of on hold
 for the moment. 

That's right. I tried to revive work on that several times, and always got 
distracted again. Just too much to do ATM ...

But I still hope to get at least to describe Richard the missing bits 
eventually.

 But I think we both really want to get it into 2.0. 

Definitely.

 If
 not, then it would almost certainly have to wait for 2.0, due to file
 format issues.

You mean 2.1.

Jürgen

Re: Patch Submission: LyxBlogger Converters

2010-04-02 Thread rgheck


On 04/02/2010 04:51 AM, Jürgen Spitzmüller wrote:

rgheck wrote:
   

On 03/30/2010 04:15 PM, Jack Desert wrote:
 

Actually, the first issue is that, especially since 2.0 will (hopefully)
let the user choose conversion paths, etc, we can go ahead and install
both the xhtml-blog and html-blog converters. This is particularly
important in 2.0, it seems to me, since otherwise people who happen to
have had elyxer installed for use with 1.6.x would not then see that
they have a new option in 2.0.x.
 

When can we reasonably expect to see this user chooses his own
conversion paths code? Which contributors are currently working on it?
Is there a branch somewhere that at least has the GUI so we can see what
it looks like?
   

Jurgen did a bunch of work on it but ran into a problem. I've told him
I'll help him sort it out and did some of the necessary background work.
But then he got busy and otherwise distracted, so it's kind of on hold
for the moment.
 

That's right. I tried to revive work on that several times, and always got
distracted again. Just too much to do ATM ...

But I still hope to get at least to describe Richard the missing bits
eventually.

   

OK, I will wait.


But I think we both really want to get it into 2.0.
 

Definitely.

   

If not, then it would almost certainly have to wait for 2.0, due to file
format issues.
 

You mean 2.1.

   

Yes, indeed.

rh

Re: Patch Submission: LyxBlogger Converters

2010-04-02 Thread Jürgen Spitzmüller

rgheck wrote:
> On 03/30/2010 04:15 PM, Jack Desert wrote:
> >> Actually, the first issue is that, especially since 2.0 will (hopefully)
> >> let the user choose conversion paths, etc, we can go ahead and install
> >> both the xhtml->blog and html->blog converters. This is particularly
> >> important in 2.0, it seems to me, since otherwise people who happen to
> >> have had elyxer installed for use with 1.6.x would not then see that
> >> they have a new option in 2.0.x.
> > 
> > When can we reasonably expect to see this "user chooses his own
> > conversion paths" code? Which contributors are currently working on it?
> > Is there a branch somewhere that at least has the GUI so we can see what
> > it looks like?
> 
> Jurgen did a bunch of work on it but ran into a problem. I've told him
> I'll help him sort it out and did some of the necessary background work.
> But then he got busy and otherwise distracted, so it's kind of on hold
> for the moment. 

That's right. I tried to revive work on that several times, and always got 
distracted again. Just too much to do ATM ...

But I still hope to get at least to describe Richard the missing bits 
eventually.

> But I think we both really want to get it into 2.0. 

Definitely.

> If
> not, then it would almost certainly have to wait for 2.0, due to file
> format issues.

You mean 2.1.

Jürgen

Re: Patch Submission: LyxBlogger Converters

2010-04-02 Thread rgheck


On 04/02/2010 04:51 AM, Jürgen Spitzmüller wrote:

rgheck wrote:
   

On 03/30/2010 04:15 PM, Jack Desert wrote:
 

Actually, the first issue is that, especially since 2.0 will (hopefully)
let the user choose conversion paths, etc, we can go ahead and install
both the xhtml->blog and html->blog converters. This is particularly
important in 2.0, it seems to me, since otherwise people who happen to
have had elyxer installed for use with 1.6.x would not then see that
they have a new option in 2.0.x.
 

When can we reasonably expect to see this "user chooses his own
conversion paths" code? Which contributors are currently working on it?
Is there a branch somewhere that at least has the GUI so we can see what
it looks like?
   

Jurgen did a bunch of work on it but ran into a problem. I've told him
I'll help him sort it out and did some of the necessary background work.
But then he got busy and otherwise distracted, so it's kind of on hold
for the moment.
 

That's right. I tried to revive work on that several times, and always got
distracted again. Just too much to do ATM ...

But I still hope to get at least to describe Richard the missing bits
eventually.

   

OK, I will wait.


But I think we both really want to get it into 2.0.
 

Definitely.

   

If not, then it would almost certainly have to wait for 2.0, due to file
format issues.
 

You mean 2.1.

   

Yes, indeed.

rh

Re: Patch Submission: LyxBlogger Converters



Thanks for sending the license statement. I've added you to the credits. 
I've also committed part of this patch, but as I was about to commit the 
whole thing another issue occurred to me. Sorry!


Actually, the first issue is that, especially since 2.0 will (hopefully) 
let the user choose conversion paths, etc, we can go ahead and install 
both the xhtml-blog and html-blog converters. This is particularly 
important in 2.0, it seems to me, since otherwise people who happen to 
have had elyxer installed for use with 1.6.x would not then see that 
they have a new option in 2.0.x.


Or we could, except that the check for elyxer does NOT guarantee that 
lyxblogger will only see output from elyxer. On the contrary, even if 
elyxer is installed, the user might, in 2.0.x, again, override LyX's 
default choice of elyxer as HTML converter. Given that lyxblogger 
expects HTML from elyxer, this could lead to export failures and 
confused users.


That makes me wonder whether lyxblogger might not operate a different 
way, namely, as a LyX--Blog converter that, if it is given a LyX file 
as input, calls elyxer itself. Then we COULD do the check for elyxer 
before registering it as a LyX--Blog converter and all would go 
according to plan.


rh

Re: Patch Submission: LyxBlogger Converters

2010-03-30 Thread Jack Desert

 Thanks for sending the license statement. I've added you to the credits. 

Excellent.
 
 Actually, the first issue is that, especially since 2.0 will (hopefully) 
 let the user choose conversion paths, etc, we can go ahead and install 
 both the xhtml-blog and html-blog converters. This is particularly 
 important in 2.0, it seems to me, since otherwise people who happen to 
 have had elyxer installed for use with 1.6.x would not then see that 
 they have a new option in 2.0.x.

When can we reasonably expect to see this user chooses his own conversion 
paths code? Which contributors are currently working on it? Is there a branch 
somewhere that at least has the GUI so we can see what it looks like?

You do have a valid point about making sure eLyXer users know there's a new kid 
on the block (LyXHTML) in 2.0. I think LyX 2.0 makes that pretty obvious with 
it's big, beautiful preview button that lists LyXHTML as one of its few 
options. Even so, the LyxBlogger xterm window already tells the user which file 
format is being used. It would be trivial yet informative to expand that 
message to: 

'''Based on your input file, you are using the eLyXer format.
LyXBlogger also supports LyX 2.0's internal LyXHTML format.
For more information, see the user's guide at ...'''

and its converse:

'''Based on your input file, you are using the LyXHTML format.
LyXBlogger also supports the eLyXer format.
For more information, see the user's guide at ...'''

 
 Or we could, except that [snip]

Or we could what? Is there a phrase missing here? 

 
 [snip]the check for elyxer does NOT guarantee that lyxblogger will only 
 see output from elyxer. This could lead to export failures and confused users.


To keep latex2html and other converters from being used, LyXBlogger could 
simply parse the html header, looking for the eLyXer style sheet, which always 
points to 
http://www.nongnu.org/elyxer/lyx.css

And  drop an error if it's not found.

'''ERROR: Unsupported Input Type 
Supported inputs are eLyXer and LyXHTML
Your input format appears to be neither of these. 
Proceed anyway? Y (N)'''

That seems quite reasonable. Easy to implement, and gives immediate feedback. 
In fact, I like it so much I'm putting it on my to do list now. 


 That makes me wonder whether lyxblogger might not operate a different 
 way, namely, as a LyX--Blog converter that, if it is given a LyX file 
 as input, calls elyxer itself. Then we COULD do the check for elyxer 
 before registering it as a LyX--Blog converter and all would go 
 according to plan.


Ah, but this would ideed bypass the beauty of the LyX chained converters, and 
add complexity to the front end of LyxBlogger. LyxBlogger would have to deal 
with three input types: xhtml, html, and .lyx. Which brings another question to 
mind: Can the LyX - LyXHTML converter be called from command line? If it can, 
then both converters could theoretically be LyX - blog. I'm not sure I'm 
interested in pursuing that route, but it's something to think about.


CONCLUSION

All said, I still support the idea of only installing one converter by default, 
as in the latest patch I provided (LyxBlogger_Formats_Patch3.diff). With the 
addition of the FORMAT messages and ERROR checking I introduced above, it keeps 
the user informed of his/her options, prevents abuse by TtH, Hevea, and others, 
and maintains elegantly uncluttered export menus. And it would be functional 
_now_, without waiting for the code from the converter-path-chooser which may 
or may not be ready for 2.0. 


-Jack


-- 
~~~
Jack Desert --Writer, Entrepeneur
Author and Spokesman: www.LetsEATalready.com
Software Developer:   http://GrooveTask.org
Email: jackdesert...@gmail.com
~~~

Re: Patch Submission: LyxBlogger Converters


On 03/30/2010 04:15 PM, Jack Desert wrote:



Actually, the first issue is that, especially since 2.0 will (hopefully)
let the user choose conversion paths, etc, we can go ahead and install
both the xhtml-blog and html-blog converters. This is particularly
important in 2.0, it seems to me, since otherwise people who happen to
have had elyxer installed for use with 1.6.x would not then see that
they have a new option in 2.0.x.
 

When can we reasonably expect to see this user chooses his own conversion 
paths code? Which contributors are currently working on it? Is there a branch 
somewhere that at least has the GUI so we can see what it looks like?

   
Jurgen did a bunch of work on it but ran into a problem. I've told him 
I'll help him sort it out and did some of the necessary background work. 
But then he got busy and otherwise distracted, so it's kind of on hold 
for the moment. But I think we both really want to get it into 2.0. If 
not, then it would almost certainly have to wait for 2.0, due to file 
format issues.



Or we could, except that [snip]
 

Or we could what? Is there a phrase missing here?

   
I think I meant: We could install lyxblogger as an html-blog converter, 
but...


   

[snip]the check for elyxer does NOT guarantee that lyxblogger will only
see output from elyxer. This could lead to export failures and confused users.
 


To keep latex2html and other converters from being used, LyXBlogger could 
simply parse the html header, looking for the eLyXer style sheet, which always 
points to
 http://www.nongnu.org/elyxer/lyx.css

And  drop an error if it's not found.

 '''ERROR: Unsupported Input Type
 Supported inputs are eLyXer and LyXHTML
 Your input format appears to be neither of these.
 Proceed anyway? Y (N)'''

That seems quite reasonable. Easy to implement, and gives immediate feedback. 
In fact, I like it so much I'm putting it on my to do list now.

   
This is all true, though note that, from within LyX itself, you can't 
ask the question Proceed anyway? There's no mechanism for LyX to 
communicate with the running script. Indeed, if you tried to ask such a 
question, the script would freeze, waiting for input. So you either 
proceed or you don't.


I do not mean any of this as a criticism of lyxblogger. My point is just 
that the way LyX's whole converter mechanism is designed presently makes 
integrating it somewhat messy. Each converter defines a path in a graph, 
and the question what conversions are possible? then becomes the 
question: Is there a path from format A to format B? You can get some 
pretty wild conversion paths sometimes---there are actually some weird 
issues that result from elyxer's being a LyX--HTML converter---but it 
is a very flexible system. My ponit is that depends upon converters 
being able to convert between the formats that they advertise as being 
able to convert between. So, to totally overstate the point, from LyX's 
point of view what this means is that lyxblogger isn't really an html to 
blog converter. It doesn't even purport to work with arbitrary HTML 
files. So installing it as an html to blog converter means installing an 
html to blog converter that we know won't work if someone chooses to 
override the choice of elyxer and instead to go lyx-latex-html.


Right now, that isn't possible, so the sort of patch you provided would 
work perfectly fine in 1.6.6. I'd suggest you prepare a patch, doing 
just the elyxer part, for branch, post it, and see if we get approval 
from Jurgen. For 2.0, it would also work as things presently are. But 
trunk isn't really usable right now, except by the very brave or very 
stupid, so there's no rush. I'd propose that we wait on this for the 
time being and see what happens with the choose your converter stuff. 
Once we get near beta, we will do something.



That makes me wonder whether lyxblogger might not operate a different
way, namely, as a LyX--Blog converter that, if it is given a LyX file
as input, calls elyxer itself. Then we COULD do the check for elyxer
before registering it as a LyX--Blog converter and all would go
according to plan.
 


Ah, but this would ideed bypass the beauty of the LyX chained converters,

   
Yes, I know, but maybe the previous explanation makes the point of the 
suggestion clearer.



and add complexity to the front end of LyxBlogger. LyxBlogger would have to 
deal with three input types: xhtml, html, and .lyx.

   

Not pretty, I know.


Which brings another question to mind: Can the LyX -  LyXHTML converter be 
called from command line?

   

Yes, this is possible: lyx -e xhtml file.lyx.


If it can, then both converters could theoretically be LyX -  blog. I'm not 
sure I'm interested in pursuing that route, but it's something to think about.

   
Worth thinking about, yes. Not clear how much benefit there is there, 
but maybe some.


rh

Re: Patch Submission: LyxBlogger Converters



Thanks for sending the license statement. I've added you to the credits. 
I've also committed part of this patch, but as I was about to commit the 
whole thing another issue occurred to me. Sorry!


Actually, the first issue is that, especially since 2.0 will (hopefully) 
let the user choose conversion paths, etc, we can go ahead and install 
both the xhtml->blog and html->blog converters. This is particularly 
important in 2.0, it seems to me, since otherwise people who happen to 
have had elyxer installed for use with 1.6.x would not then see that 
they have a new option in 2.0.x.


Or we could, except that the check for elyxer does NOT guarantee that 
lyxblogger will only see output from elyxer. On the contrary, even if 
elyxer is installed, the user might, in 2.0.x, again, override LyX's 
default choice of elyxer as HTML converter. Given that lyxblogger 
expects HTML from elyxer, this could lead to export failures and 
confused users.


That makes me wonder whether lyxblogger might not operate a different 
way, namely, as a LyX-->Blog converter that, if it is given a LyX file 
as input, calls elyxer itself. Then we COULD do the check for elyxer 
before registering it as a LyX-->Blog converter and all would go 
according to plan.


rh

Re: Patch Submission: LyxBlogger Converters

2010-03-30 Thread Jack Desert

> Thanks for sending the license statement. I've added you to the credits. 

Excellent.
 
> Actually, the first issue is that, especially since 2.0 will (hopefully) 
> let the user choose conversion paths, etc, we can go ahead and install 
> both the xhtml->blog and html->blog converters. This is particularly 
> important in 2.0, it seems to me, since otherwise people who happen to 
> have had elyxer installed for use with 1.6.x would not then see that 
> they have a new option in 2.0.x.

When can we reasonably expect to see this "user chooses his own conversion 
paths" code? Which contributors are currently working on it? Is there a branch 
somewhere that at least has the GUI so we can see what it looks like?

You do have a valid point about making sure eLyXer users know there's a new kid 
on the block (LyXHTML) in 2.0. I think LyX 2.0 makes that pretty obvious with 
it's big, beautiful preview button that lists LyXHTML as one of its few 
options. Even so, the LyxBlogger xterm window already tells the user which file 
format is being used. It would be trivial yet informative to expand that 
message to: 

'''Based on your input file, you are using the eLyXer format.
LyXBlogger also supports LyX 2.0's internal LyXHTML format.
For more information, see the user's guide at ...'''

and its converse:

'''Based on your input file, you are using the LyXHTML format.
LyXBlogger also supports the eLyXer format.
For more information, see the user's guide at ...'''

 
> Or we could, except that [snip]

Or we could what? Is there a phrase missing here? 

 
> [snip]the check for elyxer does NOT guarantee that lyxblogger will only 
> see output from elyxer. This could lead to export failures and confused users.


To keep latex2html and other converters from being used, LyXBlogger could 
simply parse the html header, looking for the eLyXer style sheet, which always 
points to 
http://www.nongnu.org/elyxer/lyx.css

And  drop an error if it's not found.

'''ERROR: Unsupported Input Type 
Supported inputs are eLyXer and LyXHTML
Your input format appears to be neither of these. 
Proceed anyway? Y (N)'''

That seems quite reasonable. Easy to implement, and gives immediate feedback. 
In fact, I like it so much I'm putting it on my to do list now. 


> That makes me wonder whether lyxblogger might not operate a different 
> way, namely, as a LyX-->Blog converter that, if it is given a LyX file 
> as input, calls elyxer itself. Then we COULD do the check for elyxer 
> before registering it as a LyX-->Blog converter and all would go 
> according to plan.


Ah, but this would ideed bypass the beauty of the LyX chained converters, and 
add complexity to the front end of LyxBlogger. LyxBlogger would have to deal 
with three input types: xhtml, html, and .lyx. Which brings another question to 
mind: Can the LyX -> LyXHTML converter be called from command line? If it can, 
then both converters could theoretically be LyX -> blog. I'm not sure I'm 
interested in pursuing that route, but it's something to think about.


CONCLUSION

All said, I still support the idea of only installing one converter by default, 
as in the latest patch I provided (LyxBlogger_Formats_Patch3.diff). With the 
addition of the FORMAT messages and ERROR checking I introduced above, it keeps 
the user informed of his/her options, prevents abuse by TtH, Hevea, and others, 
and maintains elegantly uncluttered export menus. And it would be functional 
_now_, without waiting for the code from the converter-path-chooser which may 
or may not be ready for 2.0. 


-Jack


-- 
~~~
Jack Desert --Writer, Entrepeneur
Author and Spokesman: www.LetsEATalready.com
Software Developer:   http://GrooveTask.org
Email: jackdesert...@gmail.com
~~~

Re: Patch Submission: LyxBlogger Converters


On 03/30/2010 04:15 PM, Jack Desert wrote:



Actually, the first issue is that, especially since 2.0 will (hopefully)
let the user choose conversion paths, etc, we can go ahead and install
both the xhtml->blog and html->blog converters. This is particularly
important in 2.0, it seems to me, since otherwise people who happen to
have had elyxer installed for use with 1.6.x would not then see that
they have a new option in 2.0.x.
 

When can we reasonably expect to see this "user chooses his own conversion 
paths" code? Which contributors are currently working on it? Is there a branch 
somewhere that at least has the GUI so we can see what it looks like?

   
Jurgen did a bunch of work on it but ran into a problem. I've told him 
I'll help him sort it out and did some of the necessary background work. 
But then he got busy and otherwise distracted, so it's kind of on hold 
for the moment. But I think we both really want to get it into 2.0. If 
not, then it would almost certainly have to wait for 2.0, due to file 
format issues.



Or we could, except that [snip]
 

Or we could what? Is there a phrase missing here?

   
I think I meant: We could install lyxblogger as an html->blog converter, 
but...


   

[snip]the check for elyxer does NOT guarantee that lyxblogger will only
see output from elyxer. This could lead to export failures and confused users.
 


To keep latex2html and other converters from being used, LyXBlogger could 
simply parse the html header, looking for the eLyXer style sheet, which always 
points to
 http://www.nongnu.org/elyxer/lyx.css

And  drop an error if it's not found.

 '''ERROR: Unsupported Input Type
 Supported inputs are eLyXer and LyXHTML
 Your input format appears to be neither of these.
 Proceed anyway? Y (N)'''

That seems quite reasonable. Easy to implement, and gives immediate feedback. 
In fact, I like it so much I'm putting it on my to do list now.

   
This is all true, though note that, from within LyX itself, you can't 
ask the question "Proceed anyway?" There's no mechanism for LyX to 
communicate with the running script. Indeed, if you tried to ask such a 
question, the script would freeze, waiting for input. So you either 
proceed or you don't.


I do not mean any of this as a criticism of lyxblogger. My point is just 
that the way LyX's whole converter mechanism is designed presently makes 
integrating it somewhat messy. Each converter defines a path in a graph, 
and the question "what conversions are possible?" then becomes the 
question: Is there a path from format A to format B? You can get some 
pretty wild conversion paths sometimes---there are actually some weird 
issues that result from elyxer's being a LyX-->HTML converter---but it 
is a very flexible system. My ponit is that depends upon converters 
being able to convert between the formats that they advertise as being 
able to convert between. So, to totally overstate the point, from LyX's 
point of view what this means is that lyxblogger isn't really an html to 
blog converter. It doesn't even purport to work with arbitrary HTML 
files. So installing it as an html to blog converter means installing an 
html to blog converter that we know won't work if someone chooses to 
override the choice of elyxer and instead to go lyx->latex->html.


Right now, that isn't possible, so the sort of patch you provided would 
work perfectly fine in 1.6.6. I'd suggest you prepare a patch, doing 
just the elyxer part, for branch, post it, and see if we get approval 
from Jurgen. For 2.0, it would also work as things presently are. But 
trunk isn't really usable right now, except by the very brave or very 
stupid, so there's no rush. I'd propose that we wait on this for the 
time being and see what happens with the "choose your converter" stuff. 
Once we get near beta, we will do something.



That makes me wonder whether lyxblogger might not operate a different
way, namely, as a LyX-->Blog converter that, if it is given a LyX file
as input, calls elyxer itself. Then we COULD do the check for elyxer
before registering it as a LyX-->Blog converter and all would go
according to plan.
 


Ah, but this would ideed bypass the beauty of the LyX chained converters,

   
Yes, I know, but maybe the previous explanation makes the point of the 
suggestion clearer.



and add complexity to the front end of LyxBlogger. LyxBlogger would have to 
deal with three input types: xhtml, html, and .lyx.

   

Not pretty, I know.


Which brings another question to mind: Can the LyX ->  LyXHTML converter be 
called from command line?

   

Yes, this is possible: lyx -e xhtml file.lyx.


If it can, then both converters could theoretically be LyX ->  blog. I'm not 
sure I'm interested in pursuing that route, but it's something to think about.

   
Worth thinking about, yes. Not clear how much benefit there is there, 
but maybe some.


rh

Re: Patch Submission: LyxBlogger Converters

2010-03-29 Thread rgheck

On 03/28/2010 08:52 PM, Jack Desert wrote:

LyxBlogger is specifically set up to parse the LyXHTML and eLyXer formats. Some
features may turn out strange using TtH, Hevea, or latex2html, simply because
the regular expressions may not find what they're looking for.

So yes, let's leave the html - blog converter inside the if block so a
newbie doesn't get unexpected results right out of the gate.

-Jack

In presenting the user with a preconfigured format and converter for
LyxBlogger, I have two primary aims:

1. That we guess correctly most of the time which converter the user will
want to use, and enable it by default.

2. That the user be able to select the other converter as painlessly as
possible if we guessed wrong.

That said, of course a plain Jane system should use xhtml - blog by default because
that's all that will be available. But if someone has gone to the trouble of installing
eLyXer, it makes sense to present html - blog as the default instead. On the outside
chance that we guessed wrong on this one, reverting to LyXHTML is as simple as modifying the
LyxBlogger converter from html---blog to xhtml---blog in
Preferences-Converters.

I have put together a patch that illustrates this. The following lines are
added to lyxrc.defaults based on which programs are found on the machine:

A. Lyxblogger not installed
( no entry )

B. LyxBlogger installed, eLyXer not installed
\Formatblog blog LyxBlogger
document
\converter xhtml blog lyxblogger $$i

C. LyxBlogger and eLyXer installed
\Formatblog blog LyxBlogger
document
\converter html blog lyxblogger $$i

OK, this looks good. I'll commit shortly.

Could you send a standard GPL message to the list, such as:
http://marc.info/?l=lyx-develm=117501689204059
I'm not sure we need it for this, but it'll get you listed as a contributor.

Re: Patch Submission: LyxBlogger Converters

2010-03-29 Thread rgheck

On 03/28/2010 08:52 PM, Jack Desert wrote:

So yes, let's leave the html -> blog converter inside the "if" block so a
newbie doesn't get unexpected results right out of the gate.

-Jack

In presenting the user with a preconfigured format and converter for
LyxBlogger, I have two primary aims:

1. That we guess correctly most of the time which converter the user will
want to use, and enable it by default.

2. That the user be able to select the other converter as painlessly as
possible if we guessed wrong.

That said, of course a plain Jane system should use xhtml -> blog by default because
that's all that will be available. But if someone has gone to the trouble of installing
eLyXer, it makes sense to present html -> blog as the default instead. On the outside
chance that we guessed wrong on this one, reverting to LyXHTML is as simple as modifying the
LyxBlogger converter from html--->blog to xhtml--->blog in
Preferences->Converters.

I have put together a patch that illustrates this. The following lines are
added to lyxrc.defaults based on which programs are found on the machine:

A. Lyxblogger not installed
( no entry )

B. LyxBlogger installed, eLyXer not installed
\Formatblog blog "LyxBlogger" "" "" ""
"document"
\converter xhtml blog "lyxblogger $$i" ""

C. LyxBlogger and eLyXer installed
\Formatblog blog "LyxBlogger" "" "" ""
"document"
\converter html blog "lyxblogger $$i" ""

OK, this looks good. I'll commit shortly.

Could you send a standard GPL message to the list, such as:
http://marc.info/?l=lyx-devel=117501689204059
I'm not sure we need it for this, but it'll get you listed as a contributor.

Re: Patch Submission: LyxBlogger Converters

2010-03-28 Thread Jack Desert


 LyxBlogger is specifically set up to parse the LyXHTML and eLyXer formats. 
 Some features may turn out strange using TtH, Hevea, or latex2html, simply 
 because the regular expressions may not find what they're looking for.
 
 So yes, let's leave the html - blog converter inside the if block so a 
 newbie doesn't get unexpected results right out of the gate.
 
 -Jack
 

In presenting the user with a preconfigured format and converter for 
LyxBlogger, I have two primary aims:

   1. That we guess correctly most of the time which converter the user will 
want to use, and enable it by default.  

   2. That the user be able to select the other converter as painlessly as 
possible if we guessed wrong. 

That said, of course a plain Jane system should use xhtml - blog by default 
because that's all that will be available. But if someone has gone to the 
trouble of installing eLyXer, it makes sense to present html - blog as the 
default instead. On the outside chance that we guessed wrong on this one, 
reverting to LyXHTML is as simple as modifying the LyxBlogger converter from   
html---blog   to   xhtml---blog   in Preferences-Converters.  

I have put together a patch that illustrates this. The following lines are 
added to lyxrc.defaults based on which programs are found on the machine:

A. Lyxblogger not installed
( no entry ) 

B. LyxBlogger installed, eLyXer not installed
\Formatblog   blog   LyxBlogger   document
\converter xhtml  blog   lyxblogger $$i   

C. LyxBlogger and eLyXer installed
\Formatblog   blog   LyxBlogger   document
\converter html   blog   lyxblogger $$i   


-Jack

(attachment)

-- 
~~~
Jack Desert --Writer, Entrepeneur
Author and Spokesman: www.LetsEATalready.com
Software Developer:   http://GrooveTask.org
Email: jackdesert...@gmail.com
~~~
Index: configure.py
===
--- configure.py	(revisiÃ³n: 33901)
+++ configure.py	(copia de trabajo)
@@ -646,6 +646,17 @@
   else:
 addToRC(r'''\copierhtml   python -tt $$s/scripts/ext_copy.py $$i $$o''')
 
+# Check if LyxBlogger is installed.
+path, lyxblogger = checkProg('A LyX to WordPress Blog Publishing Tool',
+  ['lyxblogger $$i'], rc_entry = [])
+if lyxblogger.find('lyxblogger') = 0:
+  addToRC(r'\Formatblog   blog   LyxBlogger   document')
+  # Use either LyXHTML or eLyXer as input to LyxBlogger, but not both
+  if elyxerfound:
+addToRC(r'\converter html   blog   lyxblogger $$i   ')
+  else:
+addToRC(r'\converter xhtml  blog   lyxblogger $$i   ')
+
 # On SuSE the scripts have a .sh suffix, and on debian they are in /usr/share/tex4ht/
 path, htmlconv = checkProg('a LaTeX - MS Word converter', [htlatex $$i 'html,word' 'symbol/!' '-cvalidate', \
 htlatex.sh $$i 'html,word' 'symbol/!' '-cvalidate', \

Re: Patch Submission: LyxBlogger Converters

2010-03-28 Thread Jack Desert


> LyxBlogger is specifically set up to parse the LyXHTML and eLyXer formats. 
> Some features may turn out strange using TtH, Hevea, or latex2html, simply 
> because the regular expressions may not find what they're looking for.
> 
> So yes, let's leave the html -> blog converter inside the "if" block so a 
> newbie doesn't get unexpected results right out of the gate.
> 
> -Jack
> 

In presenting the user with a preconfigured format and converter for 
LyxBlogger, I have two primary aims:

   1. That we guess correctly most of the time which converter the user will 
want to use, and enable it by default.  

   2. That the user be able to select the other converter as painlessly as 
possible if we guessed wrong. 

That said, of course a plain Jane system should use xhtml -> blog by default 
because that's all that will be available. But if someone has gone to the 
trouble of installing eLyXer, it makes sense to present html -> blog as the 
default instead. On the outside chance that we guessed wrong on this one, 
reverting to LyXHTML is as simple as modifying the LyxBlogger converter from   
html--->blog   to   xhtml--->blog   in Preferences->Converters.  

I have put together a patch that illustrates this. The following lines are 
added to lyxrc.defaults based on which programs are found on the machine:

A. Lyxblogger not installed
( no entry ) 

B. LyxBlogger installed, eLyXer not installed
\Formatblog   blog   "LyxBlogger"   "" "" ""  "document"
\converter xhtml  blog   "lyxblogger $$i"   ""

C. LyxBlogger and eLyXer installed
\Formatblog   blog   "LyxBlogger"   "" "" ""  "document"
\converter html   blog   "lyxblogger $$i"   ""


-Jack

(attachment)

-- 
~~~
Jack Desert --Writer, Entrepeneur
Author and Spokesman: www.LetsEATalready.com
Software Developer:   http://GrooveTask.org
Email: jackdesert...@gmail.com
~~~
Index: configure.py
===
--- configure.py	(revisiÃ³n: 33901)
+++ configure.py	(copia de trabajo)
@@ -646,6 +646,17 @@
   else:
 addToRC(r'''\copierhtml   "python -tt $$s/scripts/ext_copy.py $$i $$o"''')
 
+# Check if LyxBlogger is installed.
+path, lyxblogger = checkProg('A LyX to WordPress Blog Publishing Tool',
+  ['lyxblogger $$i'], rc_entry = [])
+if lyxblogger.find('lyxblogger') >= 0:
+  addToRC(r'\Formatblog   blog   "LyxBlogger"   "" "" ""  "document"')
+  # Use either LyXHTML or eLyXer as input to LyxBlogger, but not both
+  if elyxerfound:
+addToRC(r'\converter html   blog   "lyxblogger $$i"   ""')
+  else:
+addToRC(r'\converter xhtml  blog   "lyxblogger $$i"   ""')
+
 # On SuSE the scripts have a .sh suffix, and on debian they are in /usr/share/tex4ht/
 path, htmlconv = checkProg('a LaTeX -> MS Word converter', ["htlatex $$i 'html,word' 'symbol/!' '-cvalidate'", \
 "htlatex.sh $$i 'html,word' 'symbol/!' '-cvalidate'", \

Re: Patch Submission: LyxBlogger Converters


 So what I'd suggest is adding the LyX blogger format and the two 
 converters. By default, the LyXHTML one would be chosen, since that will 
 be a shorter path. But, as I said, both may be available, if we can 
 finish the patch to which I alluded.
 
 rh
 

Sounds reasonable. Here is a new patch that does exactly that.


For the LyXHTML format, it adds:
\converter xhtml  blog   lyxblogger $$i   
\Formatblog   blog   LyxBlogger   document

If eLyXer is installed, it also adds:
\converter html   blog   lyxblogger $$i   


I look forward to your new feature that will allow choice of conversion paths.

Jack
(attachment)

-- 
~~~
Jack Desert --Writer, Entrepeneur
Author and Spokesman: www.LetsEATalready.com
Software Developer:   http://GrooveTask.org
Email: jackdesert...@gmail.com
~~~
Index: configure.py
===
--- configure.py	(revisiÃ³n: 33883)
+++ configure.py	(copia de trabajo)
@@ -646,6 +646,16 @@
   else:
 addToRC(r'''\copierhtml   python -tt $$s/scripts/ext_copy.py $$i $$o''')
 
+# LyxBlogger with LyXHTML:
+path, lyxblogger = checkProg('A LyX to WordPress Blog Publishing Tool',
+  ['lyxblogger $$i'],
+  rc_entry = [ r'\converter xhtml  blog   %%   '])
+if lyxblogger.find('lyxblogger') = 0:
+  addToRC(r'\Formatblog   blog   LyxBlogger   document')
+  # LyxBlogger with eLyXer
+  if elyxerfound:
+addToRC(r'\converter html   blog   lyxblogger $$i   ')
+
 # On SuSE the scripts have a .sh suffix, and on debian they are in /usr/share/tex4ht/
 path, htmlconv = checkProg('a LaTeX - MS Word converter', [htlatex $$i 'html,word' 'symbol/!' '-cvalidate', \
 htlatex.sh $$i 'html,word' 'symbol/!' '-cvalidate', \

Re: Patch Submission: LyxBlogger Converters

2010-03-27 Thread rgheck


On 03/27/2010 03:46 PM, Jack Desert wrote:


If eLyXer is installed, it also adds:
 \converter html   blog   lyxblogger $$i   

   
You can allow it to add this even if elyxer isn't installed. Or is there 
some reason we want to make sure that lyxblogger only runs on elyxer's 
output and not, say, on latex2html's?


rh

Re: Patch Submission: LyxBlogger Converters

El Sat, 27 Mar 2010 18:03:45 -0400
rgheck rgh...@bobjweil.com escribió:
 On 03/27/2010 03:46 PM, Jack Desert wrote:
 
  If eLyXer is installed, it also adds:
   \converter html   blog   lyxblogger $$i   
 
 
 You can allow it to add this even if elyxer isn't installed. Or is there 
 some reason we want to make sure that lyxblogger only runs on elyxer's 
 output and not, say, on latex2html's?

LyxBlogger is specifically set up to parse the LyXHTML and eLyXer formats. Some 
features may turn out strange using TtH, Hevea, or latex2html, simply because 
the regular expressions may not find what they're looking for.

So yes, let's leave the html - blog converter inside the if block so a 
newbie doesn't get unexpected results right out of the gate.

-Jack

-- 
~~~
Jack Desert --Writer, Entrepeneur
Author and Spokesman: www.LetsEATalready.com
Software Developer:   http://GrooveTask.org
Email: jackdesert...@gmail.com
~~~

Re: Patch Submission: LyxBlogger Converters


> So what I'd suggest is adding the LyX blogger format and the two 
> converters. By default, the LyXHTML one would be chosen, since that will 
> be a shorter path. But, as I said, both may be available, if we can 
> finish the patch to which I alluded.
> 
> rh
> 

Sounds reasonable. Here is a new patch that does exactly that.


For the LyXHTML format, it adds:
\converter xhtml  blog   "lyxblogger $$i"   ""
\Formatblog   blog   "LyxBlogger"   "" "" ""  "document"

If eLyXer is installed, it also adds:
\converter html   blog   "lyxblogger $$i"   ""


I look forward to your new feature that will allow choice of conversion paths.

Jack
(attachment)

-- 
~~~
Jack Desert --Writer, Entrepeneur
Author and Spokesman: www.LetsEATalready.com
Software Developer:   http://GrooveTask.org
Email: jackdesert...@gmail.com
~~~
Index: configure.py
===
--- configure.py	(revisiÃ³n: 33883)
+++ configure.py	(copia de trabajo)
@@ -646,6 +646,16 @@
   else:
 addToRC(r'''\copierhtml   "python -tt $$s/scripts/ext_copy.py $$i $$o"''')
 
+# LyxBlogger with LyXHTML:
+path, lyxblogger = checkProg('A LyX to WordPress Blog Publishing Tool',
+  ['lyxblogger $$i'],
+  rc_entry = [ r'\converter xhtml  blog   "%%"   ""'])
+if lyxblogger.find('lyxblogger') >= 0:
+  addToRC(r'\Formatblog   blog   "LyxBlogger"   "" "" ""  "document"')
+  # LyxBlogger with eLyXer
+  if elyxerfound:
+addToRC(r'\converter html   blog   "lyxblogger $$i"   ""')
+
 # On SuSE the scripts have a .sh suffix, and on debian they are in /usr/share/tex4ht/
 path, htmlconv = checkProg('a LaTeX -> MS Word converter', ["htlatex $$i 'html,word' 'symbol/!' '-cvalidate'", \
 "htlatex.sh $$i 'html,word' 'symbol/!' '-cvalidate'", \

Re: Patch Submission: LyxBlogger Converters

2010-03-27 Thread rgheck


On 03/27/2010 03:46 PM, Jack Desert wrote:


If eLyXer is installed, it also adds:
 \converter html   blog   "lyxblogger $$i"   ""

   
You can allow it to add this even if elyxer isn't installed. Or is there 
some reason we want to make sure that lyxblogger only runs on elyxer's 
output and not, say, on latex2html's?


rh

Re: Patch Submission: LyxBlogger Converters