Re: The fast-colorizing branch: a revolution in syntax coloring

[Edward K. Ream]

On Sun, Jan 15, 2017 at 11:44 AM, 'Terry Brown' via leo-editor <
leo-editor@googlegroups.com> wrote:


> > Once in a while I am seeing this, with both Python 3 and 2:
> >
> > Internal Leo error in bodyString
> > not unicode: (the actual string)
> > Called from writeFile,makeCacheList,__get_b,bodyString
>
> Not sure that that's new - e.g. I reported it here:
> https://github.com/leo-editor/leo-editor/issues/342
> which I think was before Pyzo came on the scene.
>

​Thanks Terry.  I have just merged the fast-colorizing branch into the
trunk.

Rev f8c68ea adds two calls to g.toUnicode ​in createOutlineFromCacheList.
That may eliminate a few calls to g.internal_error.

Edward

-- 
You received this message because you are subscribed to the Google Groups 
"leo-editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to leo-editor+unsubscr...@googlegroups.com.
To post to this group, send email to leo-editor@googlegroups.com.
Visit this group at https://groups.google.com/group/leo-editor.
For more options, visit https://groups.google.com/d/optout.

Re: The fast-colorizing branch: a revolution in syntax coloring

['Terry Brown' via leo-editor]

On Sun, 15 Jan 2017 04:44:11 -0800 (PST)
"Edward K. Ream"  wrote:

> On Sunday, January 15, 2017 at 6:53:32 AM UTC-5, Edward K. Ream wrote:
> 
> *I have not tested this code with Python 2*. 
> >  
> 
> It turns out that colorizer.colorize *is* called, but only with
> python 2.
> 
> Once in a while I am seeing this, with both Python 3 and 2:
> 
> Internal Leo error in bodyString
> not unicode: (the actual string)
> Called from writeFile,makeCacheList,__get_b,bodyString

Not sure that that's new - e.g. I reported it here:
https://github.com/leo-editor/leo-editor/issues/342
which I think was before Pyzo came on the scene.

Cheers -Terry

> This may be a subtle result of the changes made in the LeoTree.select 
> logic.  Off hand, I don't see why bodyString can't just convert to 
> unicode...
> 
> I don't have time to investigate this now. Not sure how serious it
> is, but I don't think the branch can be merged in its present state.
> 
> Edward
> 

-- 
You received this message because you are subscribed to the Google Groups 
"leo-editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to leo-editor+unsubscr...@googlegroups.com.
To post to this group, send email to leo-editor@googlegroups.com.
Visit this group at https://groups.google.com/group/leo-editor.
For more options, visit https://groups.google.com/d/optout.

Re: The fast-colorizing branch: a revolution in syntax coloring

[Edward K. Ream]

On Sunday, January 15, 2017 at 6:53:32 AM UTC-5, Edward K. Ream wrote:

*I have not tested this code with Python 2*. 
>

It turns out that colorizer.colorize *is* called, but only with python 2.

Once in a while I am seeing this, with both Python 3 and 2:

Internal Leo error in bodyString
not unicode: (the actual string)
Called from writeFile,makeCacheList,__get_b,bodyString

This may be a subtle result of the changes made in the LeoTree.select 
logic.  Off hand, I don't see why bodyString can't just convert to 
unicode...

I don't have time to investigate this now. Not sure how serious it is, but 
I don't think the branch can be merged in its present state.

Edward

-- 
You received this message because you are subscribed to the Google Groups 
"leo-editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to leo-editor+unsubscr...@googlegroups.com.
To post to this group, send email to leo-editor@googlegroups.com.
Visit this group at https://groups.google.com/group/leo-editor.
For more options, visit https://groups.google.com/d/optout.

Re: The fast-colorizing branch: a revolution in syntax coloring

[Edward K. Ream]

On Sunday, January 15, 2017 at 6:43:44 AM UTC-5, Edward K. Ream wrote:
>
>
> 3. From an engineering standpoint, it revolutionizes Leo's approach to 
> Qt's QSyntaxHighlighter class.  In essence, *everything happens 
> automatically*. 
>

*I have not tested this code with Python 2*. This must be done before 
merging the branch. 

Note: leo_h.rehighlight method appears *never* to be called.  This is 
weird, imo.

That's all for now.  Time to prepare for our trip back to Madison.  I'll be 
following your comments with interest along the way.

Edward

-- 
You received this message because you are subscribed to the Google Groups 
"leo-editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to leo-editor+unsubscr...@googlegroups.com.
To post to this group, send email to leo-editor@googlegroups.com.
Visit this group at https://groups.google.com/group/leo-editor.
For more options, visit https://groups.google.com/d/optout.

The fast-colorizing branch: a revolution in syntax coloring

[Edward K. Ream]

The fast-colorizing git branch is a revolution in syntax coloring, while 
using the bulk of Leo's existing coloring code. The new code supports all 
of the languages in leo/modes.

Please test this branch for at least a day or so before being merged into 
the main line. Please report any problems immediately.

This branch is revolutionary in several senses:

1. It fixes, at long last #273 
: Qt Bug involving 
large text nodes.

2. It completes #365 : 
Use pyzo syntax coloring.

3. From an engineering standpoint, it revolutionizes Leo's approach to Qt's 
QSyntaxHighlighter class.  In essence, *everything happens automatically*. 
When changing nodes, Leo need only re-init the underlying jEdit colorizing 
code so that the proper language is colorized.  

This is a stupendous advance. No need for any of the interface in the 
ColorizerMixin class.  No need to "lock out" coloring: it happens 
automatically whenever Leo's body pane changes! Heh.  As I write this I see 
that colorizer.colorize is no longer ever called!

I don't think I have ever misunderstood a class as badly as I have 
misunderstood QSyntaxHighlighter class.  Set up and used properly, it is an 
amazing helper. To be fair, Leo uses this class in a very unusual manner. 
The mistake I made was subtle, and the Qt documents don't warn against it.

The performance bug created by calling QSyntaxHighlighter *explicitly* 
caused a "series of unfortunate events" ;-) Seeing the speed pyzo's pure 
python syntax colorer was the key breakthrough.

Edward

-- 
You received this message because you are subscribed to the Google Groups 
"leo-editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to leo-editor+unsubscr...@googlegroups.com.
To post to this group, send email to leo-editor@googlegroups.com.
Visit this group at https://groups.google.com/group/leo-editor.
For more options, visit https://groups.google.com/d/optout.