877302 - [Dialer] when searching for a number in contacts, display partial matches (matching the end of the number)

Status: RESOLVED FIXED

(Firefox OS Graveyard :: Gaia::Dialer, defect, P2)

Description

[Andreas Gal :gal]

+++ This bug was initially created as a clone of Bug #871939 +++

For a precise description please see Bug #871939. This bug is a placeholder in case Bug #871939 doesn't capture all of the issues.

Comment 1

[Andreas Gal :gal]

leo team, please provide instructions in this bug how we should handle these cases and why they should match by default area code 11 (Sao Paulo):

87654321
909087654321

987654321
9090987654321

Comment 2

[Andreas Gal :gal]

Any updates here?

Comment 3

[Leo]

We ask scenario to tester. When I receive answer, I tell as soon as possible

Comment 4

[Leo]

Unfortunetely tester doesn't reply. But we already said scenario in Bug 871939, only care about below algorithm

Step 1. Find the matching number same as the back of 7-digits number in contacts
Step 2. If device find several matching numbers, shows number that get the first contact found.

So, you shouldn't care about other cases. Brazil operator only follow this algorithm.

87654321
909087654321

987654321
9090987654321

In these cases, number should be matched though number isn't really same.(Doesn't care about area code)

Comment 5

[Andreas Gal :gal]

Ok, comment 4 is only about the searching-for-a-number case.

Comment 6

[Wayne Chang [:wchang]]

Etienne,
can you look at this bug? 
Essentially what the bug is looking for is that we only match the last 7 digits and not take consideration of any prefix. And in terms of displaying, the first hit from the search is displayed.

Comment 7

[Wayne Chang [:wchang]]

That's if Bug #871939 doesn't already cover what leo is looking for.

Comment 8

[Andreas Gal :gal]

bug 871939 doesn't do partial matching. It just makes sure the collect call prefix is stripped off for exact matches.

Comment 9

[Andreas Gal :gal]

We need an owner for this bug.

Comment 10

[Andreas Gal :gal]

Ok, so what we need here is as follows:

- create a new index telMatch8 that contains the last 8 digits of every number >= 8 digits (always).
- when doing a match and mcc == Brazil (change PhoneNumberUtils._getCountryName -> getCurrentCountry() == 'BR'), and telMatch didn't hit, and the number is >= digits, then look into telMatch8 with the last 8 digits and return that match.

Comment 11

[Andreas Gal :gal]

There is another way of fixing this actually, at least potentially. In Brazil (where we need this strange hack) there is a mandatory requirement to send the local area code via cell broadcast on channel 50 (which we display on the lock screen). We could intercept such cell broadcast on channel 50 and extract the area code in Brazil. The format is "VIVO RJ 21", so operator space state space area-code. When in Brazil (mcc == Brazil) we could then successfully transform local numbers into national numbers.

Comment 12

[Andreas Gal :gal]

Ok, I got some more vendor feedback here. Comment #11 won't work since we need this to work for Venezuela and Columbia as well, and there is no CB there. Ugh.

Comment 13

[Gregor Wagner [:gwagner]]

Attachment: v1

Comment 14

[Reuben Morais [:reuben]]

FWIW, the reason this is needed is reverse charge calls, see http://en.wikipedia.org/wiki/Collect_call#Brazil

Comment 15

[Leo]

We teseted this patch to leo device, bug is still reproduced.

Comment 16

[Leo]

Attachment: Checking list

Comment 17

[Leo]

(In reply to Leo from comment #16)
> Created attachment 757811 [details]
Checking list

This excel file is not current status. So, please see only checking list(Don't care about Pass/Fail)

Current status : saved - 01187654321

87654321 : Fail
01187654321 : Pass
909087654321 : Fail
0411187654321 : Fail
90411187654321 : Fail
[phone] : Pass

Comment 18

[Gregor Wagner [:gwagner]]

(In reply to Leo from comment #17)
> (In reply to Leo from comment #16)
> > Created attachment 757811 [details]
> Checking list
> 
> This excel file is not current status. So, please see only checking
> list(Don't care about Pass/Fail)
> 
> Current status : saved - 01187654321
> 
> 87654321 : Fail
> 01187654321 : Pass
> 909087654321 : Fail
> 0411187654321 : Fail
> 90411187654321 : Fail
> [phone] : Pass

Yes this is expected. This patch only adds the functionality but it's not enabled by default. Do we want this feature enabled by default, country specific, network specific or maybe branch specific? Is it always 7 digits or sometimes 8 digits? I will add a followup patch that enables the feature by default so you can run the tests.

Comment 19

[Gregor Wagner [:gwagner]]

Attachment: patch

Comment 20

[Gregor Wagner [:gwagner]]

Could you try with the new patch?

Comment 21

[Leo]

I'll test this patch.

Comment 22

[Leo]

I tested this patch.

Current status : saved - 01187654321

87654321 : Pass
01187654321 : Pass
909087654321 : Fail
0411187654321 : Fail
90411187654321 : Fail
[phone] : Pass

But, Fail cases are not shown any informatioin.
I think it's problem of "SimplePhoneMatcher.bestMatch(variants, matches)" function.
When inputing "909087654321", gecko give correct result, but "SimplePhoneMatcher.bestMatch(variants, matches)" returns "null"
So, I input below code in contact.js(gaia/apps/communications/dialer, for test)
====================
var matchResult = SimplePhoneMatcher.bestMatch(variants, matches);

+if(matchResult.localIndex == null || matchResult.bestMatchIndex == null) {
+  matchResult.localIndex = 0;
+  matchResult.bestMatchIndex = 0;
+}
====================
When I input code, device can find numbers.
I request you check "contacts.js" file.

Comment 23

[Leo]

Attachment: Test Result - new issue

I attach issue image

Comment 24

[Etienne Segonzac (:etienne)]

Attachment: SimplePhoneMatcher patch

Comment 25

[Etienne Segonzac (:etienne)]

Gaia part of the fix: https://github.com/mozilla-b2g/gaia/commit/7b9ab1632c6036b00fd8795bf583cd581c10ba0c

(feel free to close once the gecko part lands)

Comment 26

[Gregor Wagner [:gwagner]]

(In reply to Leo from comment #22)
> I tested this patch.
> 
> Current status : saved - 01187654321
> 
> 87654321 : Pass
> 01187654321 : Pass
> 909087654321 : Fail
> 0411187654321 : Fail
> 90411187654321 : Fail
> [phone] : Pass
> 
> But, Fail cases are not shown any informatioin.
> I think it's problem of "SimplePhoneMatcher.bestMatch(variants, matches)"
> function.
> When inputing "909087654321", gecko give correct result, but
> "SimplePhoneMatcher.bestMatch(variants, matches)" returns "null"
> So, I input below code in contact.js(gaia/apps/communications/dialer, for
> test)
> ====================
> var matchResult = SimplePhoneMatcher.bestMatch(variants, matches);
> 
> +if(matchResult.localIndex == null || matchResult.bestMatchIndex == null) {
> +  matchResult.localIndex = 0;
> +  matchResult.bestMatchIndex = 0;
> +}
> ====================
> When I input code, device can find numbers.
> I request you check "contacts.js" file.

This should be fixed now with etiennes patch.
Please test again.

Comment 27

[Leo]

It's perfectly running!

Please lend this patch to v1-train.

Thanks for your help!

Comment 28

[Ben Turner (not reading bugmail, use the needinfo flag!)]

Comment on attachment 757864 [details] [diff] [review]
patch

Review of attachment 757864 [details] [diff] [review]:
-----------------------------------------------------------------

::: b2g/app/b2g.js
@@ +388,5 @@
>  pref("dom.global-constructor.disable.mozContact", false);
>  
> +// Shortnumber matching needed for e.g. Brazil:
> +// 01187654321 can be found with 7654321
> +pref("dom.phonenumber.substringmatching", 7);

This seems wrong. Shouldn't we only enable this on a per-country basis? As I understand it this patch was for testing only.

::: dom/contacts/fallback/ContactDB.jsm
@@ +918,5 @@
>    },
>  
> +  observe: function(aSubject, aTopic, aData) {
> +    if (aTopic === 'nsPref:changed' && aData === "dom.phonenumber.substringmatching") {
> +       this.substringMatching = Services.prefs.getIntPref("dom.phonenumber.substringmatching");

You need to check that it's > 0 here too.

@@ +927,5 @@
> +    this.initDBHelper(DB_NAME, DB_VERSION, [STORE_NAME, SAVED_GETALL_STORE_NAME, REVISION_STORE], aGlobal);
> +    try {
> +      let val = Services.prefs.getIntPref("dom.phonenumber.substringmatching");
> +      if (val && val > 0)
> +        this.substringMatching = val;

Nit: braces

::: modules/libpref/src/init/all.js
@@ +4099,5 @@
>  pref("dom.mozContacts.enabled", false);
>  pref("dom.navigator-property.disable.mozContacts", true);
>  pref("dom.global-constructor.disable.mozContact", true);
>  
> +pref("dom.phonenumber.substringmatching", -1);

Should comment here that <=0 means "disabled"

Comment 29

[Ben Turner (not reading bugmail, use the needinfo flag!)]

Leo, can you provide us with a list of the countries where this special behavior is needed?

Comment 30

[Gregor Wagner [:gwagner]]

(In reply to ben turner [:bent] from comment #28)
> Comment on attachment 757864 [details] [diff] [review]
> patch
> 
> Review of attachment 757864 [details] [diff] [review]:
> -----------------------------------------------------------------
> 
> ::: b2g/app/b2g.js
> @@ +388,5 @@
> >  pref("dom.global-constructor.disable.mozContact", false);
> >  
> > +// Shortnumber matching needed for e.g. Brazil:
> > +// 01187654321 can be found with 7654321
> > +pref("dom.phonenumber.substringmatching", 7);
> 
> This seems wrong. Shouldn't we only enable this on a per-country basis? As I
> understand it this patch was for testing only.

Right we need some input here from leo. I wanted to get an OK for the rest of it :)
Thanks!

Comment 31

[Gregor Wagner [:gwagner]]

Attachment: patch

Comment 32

[Ben Turner (not reading bugmail, use the needinfo flag!)]

Comment on attachment 759112 [details] [diff] [review]
patch

Review of attachment 759112 [details] [diff] [review]:
-----------------------------------------------------------------

::: dom/contacts/fallback/ContactService.jsm
@@ +22,5 @@
>                                     "@mozilla.org/parentprocessmessagemanager;1",
>                                     "nsIMessageListenerManager");
>  
>  let myGlobal = this;
> +let substringSearchCountryList = ['BR', 'CO', 'VE'];

Each of these could have a different number of required digits, right? 7 for Brazil, but others were greater iirc.

I think you're going to want to have a pref for each country where this is needed, e.g.:

  pref("dom.phonenumber.substringmatching.BR", 7);
  pref("dom.phonenumber.substringmatching.CO", 245); // kidding

@@ +45,5 @@
>      this._db.init(myGlobal);
>  
> +    let countryName = PhoneNumberUtils.getCountryName();
> +    if (substringSearchCountryList.indexOf(countryName) >= 0) {
> +      debug("Enable Substring Matching for Phone Numbers!");

if (DEBUG)

@@ +74,5 @@
> +      this._children = null;
> +      this._cursors = null;
> +    } else if (aTopic === 'nsPref:changed' && aData === "dom.phonenumber.substringmatching") {
> +      let countryName = PhoneNumberUtils.getCountryName();
> +      if (substringSearchCountryList.indexOf(countryName) >= 0) {

Should you always call 'this._db.enableSubstringMatching(0)' if the country isn't in the list? I'm not sure if that can change during runtime (if you swap SIMs for instance?)

Comment 33

[Gregor Wagner [:gwagner]]

(In reply to ben turner [:bent] from comment #32)
> Comment on attachment 759112 [details] [diff] [review]
> patch
> 
> Review of attachment 759112 [details] [diff] [review]:
> -----------------------------------------------------------------
> 
> ::: dom/contacts/fallback/ContactService.jsm
> @@ +22,5 @@
> >                                     "@mozilla.org/parentprocessmessagemanager;1",
> >                                     "nsIMessageListenerManager");
> >  
> >  let myGlobal = this;
> > +let substringSearchCountryList = ['BR', 'CO', 'VE'];
> 
> Each of these could have a different number of required digits, right? 7 for
> Brazil, but others were greater iirc.
> 
> I think you're going to want to have a pref for each country where this is
> needed, e.g.:
> 
>   pref("dom.phonenumber.substringmatching.BR", 7);
>   pref("dom.phonenumber.substringmatching.CO", 245); // kidding


The documents we got suggest 7 digits afaik.

> 
> @@ +74,5 @@
> > +      this._children = null;
> > +      this._cursors = null;
> > +    } else if (aTopic === 'nsPref:changed' && aData === "dom.phonenumber.substringmatching") {
> > +      let countryName = PhoneNumberUtils.getCountryName();
> > +      if (substringSearchCountryList.indexOf(countryName) >= 0) {
> 
> Should you always call 'this._db.enableSubstringMatching(0)' if the country
> isn't in the list? I'm not sure if that can change during runtime (if you
> swap SIMs for instance?)

I don't think we support sim swapping during runtime.

Comment 34

[Leo]

This action is needed in Brazil.

And other country, I already send email to Mr. Andreas([e-mail]) and Mr. Wayne ([e-mail]).

But, this matching algorithm is similar with Android matching algorithm

Comment 35

[Reuben Morais [:reuben]]

I'm not so sure Android does matching on the last 7 digits only. Brazilian phone numbers are 8 digits in all states but Sao Paulo, which recently moved to 9 digit numbers (by prepending "9" to all existing numbers). I'm not saying doing this is wrong, but there is a compromise, and I'm not sure if it's obvious to everyone involved.

If I type "4123 1234", and have contacts with numbers "4123 1234" and "3123 1234", I expect to see the former first, but it'll always be the latter with the current patch. Right now numbers in Sao Paulo all have "9" as their first digit, so it doesn't matter, but that will change soon.

Also, the patch is only adding the entire (length-7, length) substring to the "match" index, which means the matched contacts will not be shown until the person has entered all 7 digits. Why is the search not using contains? That would fix this problem without changing the behavior of "match", adding new prefs for every country, etc.

Comment 36

[Leo]

I agree with you. This operation miss some cases. 


But, each operator(each country) have different matching algorithm.
Let's have one default algorithm, which is for Brazil operator. And depending on the operator, a different algorithm can be provided to override the default one. Because Brazil operator algorithm is very simple and easy to expand

Comment 37

[Gregor Wagner [:gwagner]]

Attachment: patch

Comment 38

[Gregor Wagner [:gwagner]]

Attachment: patch

Comment 39

[Reuben Morais [:reuben]]

Comment on attachment 761735 [details] [diff] [review]
patch

Review of attachment 761735 [details] [diff] [review]:
-----------------------------------------------------------------

::: b2g/app/b2g.js
@@ +387,5 @@
>  pref("dom.navigator-property.disable.mozContacts", false);
>  pref("dom.global-constructor.disable.mozContact", false);
>  
> +// Shortnumber matching needed for e.g. Brazil:
> +// 01187654321 can be found with 7654321

nit: it says "e.g. Brazil", but then the example uses 7 digits. Either change it to "e.g. Venezuela" or make it 8 digits long.

::: dom/contacts/fallback/ContactDB.jsm
@@ +886,5 @@
> +
> +        // Some countries need special handling for number matching. Bug 877302
> +        if (this.substringMatching && normalized.length > this.substringMatching) {
> +          let length = normalized.length;
> +          normalized = normalized.substring(length - this.substringMatching, length);

You can do |normalized = normalized.slice(-this.substringMatching);| and avoid the length check and the arithmetic.

@@ +931,5 @@
>        }
>      }.bind(this);
>    },
>  
> +  enableSubstringMatching: function enableSubstringMatching(aDigits) {

Maybe add a comment explaining that this is here so we can test it, but it will not work at all if called after a database exists and contacts have been saved.

::: dom/contacts/fallback/ContactService.jsm
@@ +22,5 @@
>                                     "@mozilla.org/parentprocessmessagemanager;1",
>                                     "nsIMessageListenerManager");
>  
>  let myGlobal = this;
> +let substringSearchCountryList = ['BR', 'CO', 'VE'];

Why hardcode this list here? Can't we simply check if "dom.phonenumber.substringmatching." + countryName is set?

@@ +52,5 @@
> +        if (val && val > 0) {
> +          this._db.enableSubstringMatching(val);
> +        }
> +      } catch (ex) {
> +        dump("No substringmatching entry found for: " + countryName);

+ "\n"

@@ +62,4 @@
>    },
>  
>    observe: function(aSubject, aTopic, aData) {
> +    debug("OBSERVE: " + aTopic);

if (DEBUG) …

@@ +80,5 @@
> +    } else if (aTopic === 'nsPref:changed' && aData.contains("dom.phonenumber.substringmatching")) {
> +      let countryName = PhoneNumberUtils.getCountryName();
> +      if (substringSearchCountryList.indexOf(countryName) >= 0) {
> +        let val = Services.prefs.getIntPref("dom.phonenumber.substringmatching." + countryName);
> +        debug("Set Val" + val);

if (DEBUG) …

::: dom/contacts/tests/test_contacts_substringmatching.html
@@ +10,5 @@
> +  <link rel="stylesheet" type="text/css" href="/tests/SimpleTest/test.css" />
> +</head>
> +<body>
> +
> +<a target="_blank" href="https://bugzilla.mozilla.org/show_bug.cgi?id=674720">Mozilla Bug 674720</a>

nit: Points to the wrong bug.

@@ +25,5 @@
> +  SpecialPowers.Cu.import("resource://gre/modules/ContactService.jsm");
> +}
> +
> +var substringLength = 7;
> +SpecialPowers.setIntPref("dom.phonenumber.substringmatching.BR", substringLength);

Why are we testing with a different value here? This should be 8, and should use SpecialPowers.pushPrefEnv.

@@ +79,5 @@
> +  },
> +  function () {
> +    ok(true, "Retrieving by substring 1");
> +    var length = prop.tel[0].value.length;
> +    var num = prop.tel[0].value.substring(length - substringLength, length);

Slice with negative index works for this too.

@@ +89,5 @@
> +    req.onsuccess = function () {
> +      is(req.result.length, 1, "Found exactly 1 contact.");
> +      findResult1 = req.result[0];
> +      ok(findResult1.id == sample_id1, "Same ID");
> +      dump("findResult: " + JSON.stringify(findResult1) + "\n");

This test has a leftover debugging spew, please remove.

@@ +239,5 @@
> +    req.onerror = onFailure;
> +  },
> +  function () {
> +    ok(true, "all done!\n");
> +    SpecialPowers.setIntPref("dom.phonenumber.substringmatching", -1);

Wrong pref?

::: modules/libpref/src/init/all.js
@@ +4105,5 @@
>  pref("dom.global-constructor.disable.mozContact", true);
>  
> +// WebContacts: Perform substring matching for the last N digits.
> +// -1 is disabled
> +pref("dom.phonenumber.substringmatching.BR", -1);

What about CO and VE? Instead, leave it unset and catch the exception in ContactService.

Comment 40

[Gregor Wagner [:gwagner]]

(In reply to Reuben Morais [:reuben] from comment #39)
> Comment on attachment 761735 [details] [diff] [review]
> patch
> 
> Review of attachment 761735 [details] [diff] [review]:

thx! I guess i missed the final qref for this patch.


> >  let myGlobal = this;
> > +let substringSearchCountryList = ['BR', 'CO', 'VE'];
> 
> Why hardcode this list here? Can't we simply check if
> "dom.phonenumber.substringmatching." + countryName is set?
>

I was worried about performance since this always runs during startup but prefs are loaded into memory so it shouldn't matter.

Comment 41

[Gregor Wagner [:gwagner]]

Attachment: patch

Comment 42

[Gregor Wagner [:gwagner]]

Attachment: patch

Comment 43

[Ben Turner (not reading bugmail, use the needinfo flag!)]

Comment on attachment 761995 [details] [diff] [review]
patch

Review of attachment 761995 [details] [diff] [review]:
-----------------------------------------------------------------

::: dom/contacts/fallback/ContactService.jsm
@@ +44,5 @@
>      this._db.init(myGlobal);
>  
> +    let countryName = PhoneNumberUtils.getCountryName();
> +    try {
> +      if (Services.prefs.getPrefType("dom.phonenumber.substringmatching." + countryName)) {

You should either:

  1. test that getPrefType() == PREF_INT and lose the try/catch
  2. stop checking getPrefType() and just call getIntPref(), relying on try/catch

Exceptions can be expensive I think so doing 1 sounds better to me.

You'll need the same code below in the pref observer function so maybe move this to a helper function?

@@ +79,5 @@
> +      let countryName = PhoneNumberUtils.getCountryName();
> +      if (Services.prefs.getPrefType("dom.phonenumber.substringmatching." + countryName)) {
> +        let val = Services.prefs.getIntPref("dom.phonenumber.substringmatching." + countryName);
> +        if (val && val > 0) {
> +          this._db.enableSubstringMatching(val);

Wait, shouldn't this support disabling too?

::: modules/libpref/src/init/all.js
@@ +4116,5 @@
>  pref("dom.global-constructor.disable.mozContact", true);
>  
> +// WebContacts: Perform substring matching for the last N digits.
> +// -1 is disabled
> +pref("dom.phonenumber.substringmatching.BR", -1);

You shouldn't need this. Having an entry for "disabled" should be the same as not having an entry.

Comment 44

[Ben Turner (not reading bugmail, use the needinfo flag!)]

Comment on attachment 761995 [details] [diff] [review]
patch

r=me with those addressed.

Comment 45

[Gregor Wagner [:gwagner]]

https://hg.mozilla.org/projects/birch/rev/ab100d18b100

Comment 46

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/ab100d18b100

Comment 47

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-b2g18/rev/fdc97d987049

Comment 48

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-b2g18_v1_1_0_hd/rev/fdc97d987049

Comment 49

[leo.bugzilla.gecko]

Thanks for this patch.
But, I have to apply another country "ES"(Spain).
Please check this.

Thanks.

Comment 50

[leo.bugzilla.gecko]

If the 'substringMatching' is enabled, problem occurs.
If the 'filtervalue' is not number, Gecko returns incorrect value.
-------------------------------------------------
1. Gaia Calls 'mozContacts.find(options)'
   options = {filterBy: ['tel'], filterOp: 'equals', filterValue: number};
   In this case, 'filterValue' is string.(ex. number = "Withheld number")
2. Gaia receive incorrect result from Gecko.
-------------------------------------------------

This issue depends on 'substringMatching'.
The value of 'substringMatching' is 'undefined' -> OK.
The value of 'substringMatching' is set -> Reproduced.

I'll attach log file.
Please check log.

Thanks.

Comment 51

[leo.bugzilla.gecko]

I'm sorry.
I can't upload file.
I'll copy important log.

----------------------------------------------
1. ContactManager: find! {"sortBy":"undefined","sortOrder":"undefined","filterValue":"Withheldnumber","filterOp":"match","filterBy":["tel"],"filterLimit":0}

2. Fallback ContactService component: Fallback DOMContactManager::receiveMessage Contacts:Find

3. ContactDB component: ContactDB:find val:Withheldnumber by: tel op: match
ContactDB component: _findWithIndex: Withheldnumber match tel 
ContactDB component: key: tel
ContactDB component: match

4. PhoneNumberutils: normalize(Withheldnumber): , true

5. ContactDB component: substringMatching: 8 (<- I was added)

6. ContactDB component: Request successful. Record count: 4

7. Fallback ContactService component: result:[{"properties":{"name":["Activar Buzon Movistar"],"honorificPrefix":[],"givenName":["Activar Buzon Movistar"],"additionalName":[],"familyName":[],"honorificSuffix":[],"nickname":[],"email":[],"photo":[],"url":[],"category":[],"adr":[],"tel":[{"type":"Other","value":"22500"}],"org":[],"jobTitle":[],"bday":null,"note":[],"impp":[],"anniversary":null,"sex":null,"genderIdentity":null},"updated":"1970-01-01T00:38:48.608Z","published":"1970-01-01T00:38:48.608Z","id":"921b0c5dffbf4b5fb31784eaddc4c690"},{"properties":{"name":["Atencion al Cliente"],"honorificPrefix":[],"givenName":["Atencion al Cliente"],"additionalName":[],"familyName":[],"honorificSuffix":[],"nickname":[],"email":[],"photo":[],"url":[],"category":[],"adr":[],"tel":[{"type":"Other","value":"1004"}],"org":[],"jobTitle":[],"bday":null,"note":[],"impp":[],"anniversary":null,"sex":null,"genderIdentity":null},"updated":"1970-01-01T00:38:48.643Z","published":"1970-01-01T00:38:48.643Z","id":"a73119

Comment 52

[Wayne Chang [:wchang]]

Ryan, with the reopen in comment 50, should we modify the branch fix information as well?

(In reply to Ryan VanderMeulen [:RyanVM UTC-4] from comment #47)
> https://hg.mozilla.org/releases/mozilla-b2g18/rev/fdc97d987049

Comment 53

[Ryan VanderMeulen [:RyanVM]]

Yes

Comment 54

[Alex Keybl [:akeybl]]

(In reply to leo.bugzilla.gecko from comment #50)
> If the 'substringMatching' is enabled, problem occurs.
> If the 'filtervalue' is not number, Gecko returns incorrect value.

Please open a new blocking bug.