[libfuzzer] Add or update dictionaries for //net fuzzers.

[libfuzzer] Add or update dictionaries for //net fuzzers.

BUG=624752
Committed: https://crrev.com/4a561d36be6ea398dd54313ed552e1d154b21661
Cr-Commit-Position: refs/heads/master@{#404175}

[mmoroz, 2016-07-07 14:21:00]

Description was changed from

==========
[libfuzzer] Add or update dictionaries for //net fuzzers.

BUG=624752
==========

to

==========
[libfuzzer] Add or update dictionaries for //net fuzzers.

BUG=624752
==========

[mmoroz, 2016-07-07 14:21:00]

mmoroz@chromium.org changed reviewers:
+ eroman@chromium.org, inferno@chromium.org

[eroman, 2016-07-07 16:23:50]

lgtm

[mmoroz, 2016-07-07 16:25:50]

The CQ bit was checked by mmoroz@chromium.org

[mmoroz, 2016-07-07 16:25:50]

The CQ bit was unchecked by mmoroz@chromium.org

[mmoroz, 2016-07-07 16:26:09]

The CQ bit was checked by mmoroz@chromium.org

[commit-bot: I haz the power, 2016-07-07 16:26:55]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[aarya, 2016-07-07 16:31:34]

aarya@google.com changed reviewers:
+ aarya@google.com, aizatsky@chromium.org, kcc@chromium.org, ochang@chromium.org

[aarya, 2016-07-07 16:31:35]

https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
File net/data/fuzzer_dictionaries/net_url_request_fuzzer.dict (right):

https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
net/data/fuzzer_dictionaries/net_url_request_fuzzer.dict:347: ")"
Should we ignore one letter stuff ?

https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
File net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict
(right):

https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict:152:
"[RFC4270]"
Are these parsed properly ? Looks like incorrect keywords that should be
filtered.

https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict:419:
"\"SHOULD\","
Also should we be stripping like ", ', [, ], etc on both left and right ?

[aarya, 2016-07-07 16:34:51]

More comments.

https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
File net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict
(right):

https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict:108:
"|I|S|S|S|"
Probably needs ignoring.

https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict:455: "L.,"
I see a lot of letter followed by .(dot) which seem bad, probably needs
ignoring.

[mmoroz, 2016-07-07 16:35:14]

The CQ bit was unchecked by mmoroz@chromium.org

[aarya, 2016-07-07 16:35:24]

The CQ bit was unchecked by aarya@google.com

[aarya, 2016-07-07 16:35:52]

Removed CQ, i think we should fixup dict generation algorithm as per previous
comments first.

[mmoroz, 2016-07-07 16:44:37]

Thanks for comments Abhishek, really good suggestions.

Do you mind to land this for now? Then I will update the generation script in a
separate CL, and then replace these dicts.

I agree that these dicts are not perfect and like your comments, but these dicts
are much better than nothing or old dicts, I would like to use them for a couple
of days :)

https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
File net/data/fuzzer_dictionaries/net_url_request_fuzzer.dict (right):

https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
net/data/fuzzer_dictionaries/net_url_request_fuzzer.dict:347: ")"
On 2016/07/07 16:31:35, aarya wrote:
> Should we ignore one letter stuff ?

Good point, probably it's worth to remove for the most cases.

https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
File net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict
(right):

https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict:108:
"|I|S|S|S|"
On 2016/07/07 16:34:51, aarya wrote:
> Probably needs ignoring.

+1, or may be parse '|' like OR and generate multiple elements. Though this
one-symbol elements may be ignored.

https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict:152:
"[RFC4270]"
On 2016/07/07 16:31:35, aarya wrote:
> Are these parsed properly ? Looks like incorrect keywords that should be
> filtered.

I'll write some additional filters to ignore stuff like that. May be also it's
worth to remove '[', ']' and other brackets, but sometimes may be not.

https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict:419:
"\"SHOULD\","
On 2016/07/07 16:31:35, aarya wrote:
> Also should we be stripping like ", ', [, ], etc on both left and right ?

I need to think a bit about this. For the most cases we may strip them, but
would be great to have when it's important.

https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict:455: "L.,"
On 2016/07/07 16:34:51, aarya wrote:
> I see a lot of letter followed by .(dot) which seem bad, probably needs
> ignoring.

Interesting. Probably worth to remove.

[inferno, 2016-07-07 16:46:26]

On 2016/07/07 16:44:37, mmoroz wrote:
> Thanks for comments Abhishek, really good suggestions.
> 
> Do you mind to land this for now? Then I will update the generation script in
a
> separate CL, and then replace these dicts.
> 
> I agree that these dicts are not perfect and like your comments, but these
dicts
> are much better than nothing or old dicts, I would like to use them for a
couple
> of days :)
> 
>
https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
> File net/data/fuzzer_dictionaries/net_url_request_fuzzer.dict (right):
> 
>
https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
> net/data/fuzzer_dictionaries/net_url_request_fuzzer.dict:347: ")"
> On 2016/07/07 16:31:35, aarya wrote:
> > Should we ignore one letter stuff ?
> 
> Good point, probably it's worth to remove for the most cases.
> 
>
https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
> File net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict
> (right):
> 
>
https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
> net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict:108:
> "|I|S|S|S|"
> On 2016/07/07 16:34:51, aarya wrote:
> > Probably needs ignoring.
> 
> +1, or may be parse '|' like OR and generate multiple elements. Though this
> one-symbol elements may be ignored.
> 
>
https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
> net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict:152:
> "[RFC4270]"
> On 2016/07/07 16:31:35, aarya wrote:
> > Are these parsed properly ? Looks like incorrect keywords that should be
> > filtered.
> 
> I'll write some additional filters to ignore stuff like that. May be also it's
> worth to remove '[', ']' and other brackets, but sometimes may be not.
> 
>
https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
> net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict:419:
> "\"SHOULD\","
> On 2016/07/07 16:31:35, aarya wrote:
> > Also should we be stripping like ", ', [, ], etc on both left and right ?
> 
> I need to think a bit about this. For the most cases we may strip them, but
> would be great to have when it's important.
> 
>
https://codereview.chromium.org/2128583006/diff/90001/net/data/fuzzer_diction...
> net/data/fuzzer_dictionaries/net_websocket_frame_parser_fuzzer.dict:455: "L.,"
> On 2016/07/07 16:34:51, aarya wrote:
> > I see a lot of letter followed by .(dot) which seem bad, probably needs
> > ignoring.
> 
> Interesting. Probably worth to remove.

Sure sounds good. Let me readd cq.

[inferno, 2016-07-07 16:46:32]

The CQ bit was checked by inferno@chromium.org

[commit-bot: I haz the power, 2016-07-07 16:47:27]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-07 17:45:24]

Description was changed from

==========
[libfuzzer] Add or update dictionaries for //net fuzzers.

BUG=624752
==========

to

==========
[libfuzzer] Add or update dictionaries for //net fuzzers.

BUG=624752
==========

[commit-bot: I haz the power, 2016-07-07 17:45:25]

Committed patchset #6 (id:90001)

[commit-bot: I haz the power, 2016-07-07 17:47:34]

Description was changed from

==========
[libfuzzer] Add or update dictionaries for //net fuzzers.

BUG=624752
==========

to

==========
[libfuzzer] Add or update dictionaries for //net fuzzers.

BUG=624752

Committed: https://crrev.com/4a561d36be6ea398dd54313ed552e1d154b21661
Cr-Commit-Position: refs/heads/master@{#404175}
==========

[commit-bot: I haz the power, 2016-07-07 17:47:35]

Patchset 6 (id:??) landed as
https://crrev.com/4a561d36be6ea398dd54313ed552e1d154b21661
Cr-Commit-Position: refs/heads/master@{#404175}