Disable JBIG2 cache; prevent data corruption

Disable JBIG2 cache; prevent data corruption

https://code.google.com/p/pdfium/issues/detail?id=207

BUG=pdfium:207

[jbreiden, 2015-09-23 23:47:03]

jbreiden@google.com changed reviewers:
+ thestig@chromium.org

[Lei Zhang, 2015-09-30 22:41:58]

Did you want me to review this? You added me as a reviewer but never sent any
messages.

[jbreiden, 2015-09-30 23:43:29]

Yes, we need to stop the possibility of data corruption. I 
would like you (or someone else?) to review and then I'd like
to submit it.

I did not realize that "sending a message" was part of the 
development flow of Chromium. Sorry.

[Lei Zhang, 2015-09-30 23:46:11]

On 2015/09/30 23:43:29, jbreiden wrote:
> Yes, we need to stop the possibility of data corruption. I 
> would like you (or someone else?) to review and then I'd like
> to submit it.

Will take a look.

> I did not realize that "sending a message" was part of the 
> development flow of Chromium. Sorry.

Yes, it's possible to hit publish, and then add a reviewer, or hit edit, add a
reviewer, and not send anything. You did the latter.

Feel free to ping me if I don't reply to your CL after 2-3 days. If I'm OOO, I
usually temporarily change my name to indicate that.

[Lei Zhang, 2015-09-30 23:47:40]

https://chromiumcodereview.appspot.com/1362133003/diff/1/core/src/fxcodec/jbi...
File core/src/fxcodec/jbig2/JBig2_Context.cpp (right):

https://chromiumcodereview.appspot.com/1362133003/diff/1/core/src/fxcodec/jbi...
core/src/fxcodec/jbig2/JBig2_Context.cpp:18: // Dislable until we can figure out
how to clear cache between documents.
typo

https://chromiumcodereview.appspot.com/1362133003/diff/1/core/src/fxcodec/jbi...
core/src/fxcodec/jbig2/JBig2_Context.cpp:18: // Dislable until we can figure out
how to clear cache between documents.
Can you file a public pdfium bug to roughly describe the issue, and reference it
here?

[jbreiden, 2015-10-01 00:34:04]

I'll hopefully remember to do this tomorrow. Heading home now.

[jbreiden, 2015-10-01 21:44:49]

Ready for review / submit

https://chromiumcodereview.appspot.com/1362133003/diff/1/core/src/fxcodec/jbi...
File core/src/fxcodec/jbig2/JBig2_Context.cpp (right):

https://chromiumcodereview.appspot.com/1362133003/diff/1/core/src/fxcodec/jbi...
core/src/fxcodec/jbig2/JBig2_Context.cpp:18: // Dislable until we can figure out
how to clear cache between documents.
On 2015/09/30 23:47:40, Lei Zhang wrote:
> Can you file a public pdfium bug to roughly describe the issue, and reference
it
> here?

Done.

https://chromiumcodereview.appspot.com/1362133003/diff/1/core/src/fxcodec/jbi...
core/src/fxcodec/jbig2/JBig2_Context.cpp:18: // Dislable until we can figure out
how to clear cache between documents.
On 2015/09/30 23:47:40, Lei Zhang wrote:
> typo

Done.

[Lei Zhang, 2015-10-01 21:46:16]

lgtm, do you need me to commit for you?

[Lei Zhang, 2015-10-01 21:46:43]

Oh, and the convention is BUG=pdfium:207.

[jbreiden, 2015-10-01 21:47:35]

The CQ bit was checked by jbreiden@google.com

[jbreiden, 2015-10-01 21:49:35]

Yes, please commit if you are able. I ran the following command but it didn't do
anything:

git cl set_commit

[Lei Zhang, 2015-10-01 21:50:46]

On 2015/10/01 21:49:35, jbreiden wrote:
> Yes, please commit if you are able. I ran the following command but it didn't
do
> anything:
> 
> git cl set_commit

That's not the right command. It should be: git cl land

[jbreiden, 2015-10-01 21:57:50]

$ git cl land
...
fatal: remote error: Access denied to [my-personal-email]

[Lei Zhang, 2015-10-01 22:05:26]

On 2015/10/01 21:57:50, jbreiden wrote:
> $ git cl land
> ...
> fatal: remote error: Access denied to [my-personal-email]

Ok, CL for landing: https://codereview.chromium.org/1382073002/

[Adrian Kuegel, 2015-10-12 07:56:10]

On 2015/10/01 22:05:26, Lei Zhang wrote:
> On 2015/10/01 21:57:50, jbreiden wrote:
> > $ git cl land
> > ...
> > fatal: remote error: Access denied to [my-personal-email]
> 
> Ok, CL for landing: https://codereview.chromium.org/1382073002/

Using git cl set_commit made the CQ try to land it, but it couldn't because it
doesn't know which project it belongs to. So it is currently still processing
this CL in each loop (CQ is not too smart about that :-( ).
I think closing this CL probably will help, so I am doing this now.

[Adrian Kuegel, 2015-10-12 08:20:36]

On 2015/10/12 07:56:10, Adrian Kuegel wrote:
> On 2015/10/01 22:05:26, Lei Zhang wrote:
> > On 2015/10/01 21:57:50, jbreiden wrote:
> > > $ git cl land
> > > ...
> > > fatal: remote error: Access denied to [my-personal-email]
> > 
> > Ok, CL for landing: https://codereview.chromium.org/1382073002/
> 
> Using git cl set_commit made the CQ try to land it, but it couldn't because it
> doesn't know which project it belongs to. So it is currently still processing
> this CL in each loop (CQ is not too smart about that :-( ).
> I think closing this CL probably will help, so I am doing this now.

Ok, closing the issue helped. CQ is now not processing this CL anymore. And also
since recently, git cl set_commit will fail if you are trying it on a CL without
CQ support.