[liblouis-liblouisxml] Re: House cleaning, i.e. pending patches

  • From: "Jonathan Sharp" <jonathans@xxxxxxxxxx>
  • To: <liblouis-liblouisxml@xxxxxxxxxxxxx>
  • Date: Thu, 20 May 2010 22:00:35 +1200

Hi Christian,

The problem with holes in outpos also occurs during forward translation, for 
example forward translating the string "Pride and Prejudice" using en-us-g2 
gives outpos values of 1 2 3 4 5 6 7 7 7 8 12 -1 -1 13 14 15 16 17 18

I wonder if it would be better to fix these problems at their source rather 
than initializing outpos and then checking it's values after translation?

Thanks

Jonathan Sharp


-----Original Message-----
From: liblouis-liblouisxml-bounce@xxxxxxxxxxxxx 
[mailto:liblouis-liblouisxml-bounce@xxxxxxxxxxxxx] On Behalf Of Christian Egli
Sent: Thursday, 20 May 2010 8:54 p.m.
To: liblouis-liblouisxml
Subject: [liblouis-liblouisxml] House cleaning, i.e. pending patches

Hi all

I have some time on my hands the next few days to do some liblouis house
cleaning, namely there are a number of pending patches that need to be
either incorporated or dropped.

1. The patch from Volker Bijewitz to do with cursorCompbrlLeftOnly. We
   asked Volker for a test case but he never sent one. The patch itself
   looks very minor and harmless. Should I incorporate the patch or
   should I just forget it?

2. Major changes to the danish table from Daniel K. Gartmann. I haven't
   done a lou_checktable on it and it is hard for me to tell wether this
   is a good change. It appears that he mostly just moved stuff from the
   danish grade 1 table into the da.ctb table. Also I wonder if he knew
   the existing grade 1 table, since he mostly deleted all from it. As I
   don't speak danish nor danish braille this is hard for me to judge.
   Should I just commit?

3. Two patches coming from the issue tracker from timothy.ty.lee: One is
   a documentation patch which fixes a problem where 
> The "outpos" and "inpos" parameters for lou_translate() and
> lou_backTranslate() are incorrectly described inside the
> documentation.
>
> By inspecting the source code and checking values returned by those two
> functions, it becomes evident that "outpos" should have a length of
> "inlen", and "inpos" a length of "outlen", not the other way around.
>
> This mix-up in array length is also present in lou_around.c when dumping
> inpos and outpos returned by lou_backTranslate().
  I don't use the screen reader features so I can't tell if this is this
  really so? Should I apply this patch?

4. The other patch by timothy.ty.lee fixes a problem with
   backtranslation and uninitialized ouput positions. It first
   initializes the outputpos with -1 and in the end replaces -1 values
   with the previous position. Does this make sense? Again should I
   apply? Here's the description:

> To reproduce the problem:
> 1. Before calling lou_backTranslate(), fill outpos with -1's
> 2. Back-translate ",numb} #c = ,,y6" using en-us-g2.ctb
> 
> When you examine the values returned in outpos, it can be seen that some of
> the -1 were not overwritten by the function.  What should happen instead is
> that outpos[0..inlen] should be set by lou_backTranslate().
> 
> This problem is present in the trunk, and affects all operating systems.
> 
> The attached patch aganist trunk overcomes the problem by filling outpos[]
> with -1's to mark array as uninitialized when lou_backTranslate() starts.
> Before returning to caller, the function examines outpos[] and fills in
> unused elements in the array.

5. Lastly there is my patch which increases the MAXSTRING to 512 instead
   of 256 to handle my problem with very long table lists. I tried to
   solve the problem by dynamically allocating the memory but got stuck
   halfway. John said once that we can "temporary" increase the value.
   Does that mean I can check my change in?

Thanks
Christian

-- 
Christian Egli
Swiss Library for the Blind, Visually Impaired and Print Disabled
Grubenstrasse 12, CH-8045 Zürich, Switzerland
For a description of the software and to download it go to
http://www.jjb-software.com
 

__________ Information from ESET NOD32 Antivirus, version of virus signature 
database 5130 (20100519) __________

The message was checked by ESET NOD32 Antivirus.

http://www.eset.com
 
 

__________ Information from ESET NOD32 Antivirus, version of virus signature 
database 5131 (20100520) __________

The message was checked by ESET NOD32 Antivirus.

http://www.eset.com
 


For a description of the software and to download it go to
http://www.jjb-software.com

Other related posts:

  1. Home
  2. liblouis-liblouisxml
  3. Archive
  4. 05-2010