Allow braille translation result to be larger than twice the size of the input.

Allow braille translation result to be larger than twice the size of the input.

The liblouis wrapper used to allocate an output buffer that was twice as big
as the number of input characters.  There are situations where this is
not enough.  The new approach is to retry with a larger buffer up to
a certain limit.

R=dtseng@chromium.org
BUG=none

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=243251

[David Tseng, 2014-01-06 18:59:45]

LGTM

https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis/nac...
File third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc (right):

https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis/nac...
third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc:130: int outlen;
nit: outbufsize

https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis/nac...
third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc:146: // retranslations
(see the coments above the second exit condition inside
s/coments/comments

https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis/nac...
third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc:171: outbuf.clear();
Is it true that 8-dot tables are always 1-1 mappings from text to braille?

[plundblad, 2014-01-06 19:35:27]

dtseng@chromium.org writes:
> LGTM
> 
> 
>
https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis/nac...
> File third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc (right):
> 
>
https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis/nac...
> third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc:130: int outlen;
> nit: outbufsize

Keeping consistent with inlen. These are 'in/out' parameters that get modified
by lou_translate and after the call they don't reflect the buffer size, but
rather
how much was consumed.  They are named after the parameters of lou_translate.


>
https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis/nac...
> third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc:146: //
> retranslations (see the coments above the second exit condition inside
> s/coments/comments

Done, 

>
https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis/nac...
> third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc:171:
> outbuf.clear();
> Is it true that 8-dot tables are always 1-1 mappings from text to
> braille?

No.  Any particular reason for asking?  The distinction between six and eight
dot
tables is a bit vague.  There's nothing preventing an 8 dot table to contain
things
like ellipsis -> ... (1 unicode character becomes 3 cells) for
example.  Also, currently, undefined characters are represented as
'\xhhhh', potentially making 1 character into 8 cells.  (crbug.com/321715).

> https://codereview.chromium.org/124403002/

-- 

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[commit-bot: I haz the power, 2014-01-06 19:42:48]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/plundblad@chromium.org/124403002/120001

[David Tseng, 2014-01-06 20:02:32]

On Mon, Jan 6, 2014 at 11:35 AM, Peter Nilsson Lundblad <
plundblad@google.com> wrote:

> dtseng@chromium.org writes:
> > LGTM
> >
> >
> >
>
https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis/nac...
> > File third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc (right):
> >
> >
>
https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis/nac...
> > third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc:130: int outlen;
> > nit: outbufsize
>
> Keeping consistent with inlen. These are 'in/out' parameters that get
> modified
> by lou_translate and after the call they don't reflect the buffer size,
> but rather
> how much was consumed.  They are named after the parameters of
> lou_translate.
>
> You renamed inLen to inbufsize above...

>
> >
>
https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis/nac...
> > third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc:146: //
> > retranslations (see the coments above the second exit condition inside
> > s/coments/comments
>
> Done,
>
> >
>
https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis/nac...
> > third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc:171:
> > outbuf.clear();
> > Is it true that 8-dot tables are always 1-1 mappings from text to
> > braille?
>
> No.  Any particular reason for asking?  The distinction between six and
> eight dot
> tables is a bit vague.  There's nothing preventing an 8 dot table to
> contain things
> like ellipsis -> ... (1 unicode character becomes 3 cells) for
> example.  Also, currently, undefined characters are represented as
> '\xhhhh', potentially making 1 character into 8 cells.  (crbug.com/321715
> ).
>
> Well, besides undefined characters, isn't US English computer braille one
to one? (therefore, avoiding the retry logic).



> > https://codereview.chromium.org/124403002/
>
> --
>
> To unsubscribe from this group and stop receiving emails from it, send an
> email to chromium-reviews+unsubscribe@chromium.org.
>

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[commit-bot: I haz the power, 2014-01-06 20:13:33]

Retried try job too often on mac_rel for step(s) base_unittests, browser_tests,
interactive_ui_tests, net_unittests, unit_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=mac_rel&nu...

[commit-bot: I haz the power, 2014-01-06 23:12:22]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/plundblad@chromium.org/124403002/120001

[Peter Lundblad, 2014-01-06 23:15:40]

David Tseng writes:
> 
> 
> 
> On Mon, Jan 6, 2014 at 11:35 AM, Peter Nilsson Lundblad
<plundblad@google.com>
> wrote:
> 
>     dtseng@chromium.org writes:
>     > LGTM
>     >
>     >
>     >
https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis
>     /nacl_wrapper/liblouis_wrapper.cc
>     > File third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc (right):
>     >
>     >
https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis
>     /nacl_wrapper/liblouis_wrapper.cc#newcode130
>     > third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc:130: int outlen;
>     > nit: outbufsize
> 
>     Keeping consistent with inlen. These are 'in/out' parameters that get
>     modified
>     by lou_translate and after the call they don't reflect the buffer size,
but
>     rather
>     how much was consumed.  They are named after the parameters of
>     lou_translate.
> 
> 
> You renamed inLen to inbufsize above... 

Nope, it was moved down into the loop since it is only needed in that
scope.  There was another variable added at its place;)


>     >
https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis
>     /nacl_wrapper/liblouis_wrapper.cc#newcode146
>     > third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc:146: //
>     > retranslations (see the coments above the second exit condition inside
>     > s/coments/comments
> 
>     Done,
> 
>     >
https://codereview.chromium.org/124403002/diff/40001/third_party/liblouis
>     /nacl_wrapper/liblouis_wrapper.cc#newcode171
>     > third_party/liblouis/nacl_wrapper/liblouis_wrapper.cc:171:
>     > outbuf.clear();
>     > Is it true that 8-dot tables are always 1-1 mappings from text to
>     > braille?
> 
>     No.  Any particular reason for asking?  The distinction between six and
>     eight dot
>     tables is a bit vague.  There's nothing preventing an 8 dot table to
>     contain things
>     like ellipsis -> ... (1 unicode character becomes 3 cells) for
>     example.  Also, currently, undefined characters are represented as
>     '\xhhhh', potentially making 1 character into 8 cells.  (crbug.com/321715
>     ).
> 
> 
> Well, besides undefined characters, isn't US English computer braille one to
> one? (therefore, avoiding the retry logic).

We avoid retranslation in many cases, especially since we start out
with a buffer that has twice the size of the input.


>  
> 
>     > https://codereview.chromium.org/124403002/
>    
>     --
>    
>     To unsubscribe from this group and stop receiving emails from it, send an
>     email to chromium-reviews+unsubscribe@chromium.org.
> 
> 

-- 

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[commit-bot: I haz the power, 2014-01-07 10:58:47]

Change committed as 243251