Commit Graph

59 Commits

Author SHA1 Message Date
Tooru Fujisawa
328d63ce7f Bug 1777486 - Part 2: Migrate XPCOMUtils.jsm consumers with automatic migration. r=webdriver-reviewers,geckoview-reviewers,preferences-reviewers,application-update-reviewers,pip-reviewers,kmag,owlish,whimboo
Differential Revision: https://phabricator.services.mozilla.com/D151214
2022-07-12 04:21:34 +00:00
Andreea Pavel
f009ba2983 Backed out 11 changesets (Bug 1777486) for failing bc at browser_startup.js on a CLOSED TREE
Backed out changeset b6c4c386f1a6 (Bug 1777486)
Backed out changeset 195cc2de8433 (Bug 1777486)
Backed out changeset 20c746fb1648 (Bug 1777486)
Backed out changeset d5fd8173d62d (Bug 1777486)
Backed out changeset 6d758fab5a3e (Bug 1777486)
Backed out changeset e938b601ba15 (Bug 1777486)
Backed out changeset 0c4ea0b9416b (Bug 1777486)
Backed out changeset 0559c53cc668 (Bug 1777486)
Backed out changeset eea573d3a9f9 (Bug 1777486)
Backed out changeset 9ce3a6496a49 (bug 1777486)
Backed out changeset b0867652fc48 (bug 1777486)
2022-07-11 22:24:40 +03:00
Tooru Fujisawa
bf0a8b781b Bug 1777486 - Part 2: Migrate XPCOMUtils.jsm consumers with automatic migration. r=webdriver-reviewers,geckoview-reviewers,preferences-reviewers,application-update-reviewers,pip-reviewers,kmag,owlish,whimboo
Differential Revision: https://phabricator.services.mozilla.com/D151214
2022-07-11 15:09:13 +00:00
Tooru Fujisawa
b10d0289af Bug 1667455 - Part 4: Stop importing Services.jsm from JSM. r=kmag,webdriver-reviewers,perftest-reviewers,webcompat-reviewers,geckoview-reviewers,application-update-reviewers,pip-reviewers,twisniewski,devtools-reviewers,m_kato,jdescottes,ochameau,mconley,sfoster,AlexandruIonescu
Differential Revision: https://phabricator.services.mozilla.com/D150893
2022-07-11 12:41:50 +00:00
Tooru Fujisawa
9aa6514fb2 Bug 1772100 - Part 2: Use plain object for lazy getter in browser/components/attribution/. r=Gijs
Differential Revision: https://phabricator.services.mozilla.com/D148012
2022-06-09 16:21:47 +00:00
Tooru Fujisawa
ebcb6dfd9c Bug 1772360 - Remove unnecessary call for XPCOMUtils.defineLazyGlobalGetters. r=florian,webdriver-reviewers,webcompat-reviewers,denschub,whimboo
Differential Revision: https://phabricator.services.mozilla.com/D148168
2022-06-06 04:42:10 +00:00
Tooru Fujisawa
bf94464e91 Bug 1772351 - Part 1: Stop calling ChromeUtils.defineModuleGetter for AppConstants. r=florian
Differential Revision: https://phabricator.services.mozilla.com/D148164
2022-06-06 04:42:08 +00:00
Tooru Fujisawa
c56eb55371 Bug 1772313 - Part 5: Stop calling ChromeUtils.defineModuleGetter for Services. r=kmag,webcompat-reviewers,pip-reviewers,denschub,mhowell
Differential Revision: https://phabricator.services.mozilla.com/D148158
2022-06-06 04:42:06 +00:00
Molnar Sandor
4c2b06df41 Backed out 15 changesets (bug 1772313, bug 1772351, bug 1772360) for causing xpc failures in telemetry/tests/unit/test_SocketScalars.js CLOSED TREE
Backed out changeset c8b0a2ed239e (bug 1772360)
Backed out changeset 1922adad6abe (bug 1772351)
Backed out changeset 6d3634cce489 (bug 1772351)
Backed out changeset f03968a9d053 (bug 1772351)
Backed out changeset e7a38ef90fe4 (bug 1772351)
Backed out changeset b389e7654771 (bug 1772313)
Backed out changeset 59a33598ff51 (bug 1772313)
Backed out changeset ab60885a8a93 (bug 1772313)
Backed out changeset aad8efac9d05 (bug 1772313)
Backed out changeset 4378e519a8e7 (bug 1772313)
Backed out changeset 302ac54741f8 (bug 1772313)
Backed out changeset 74c59f9fd51c (bug 1772313)
Backed out changeset 575fb877c56c (bug 1772313)
Backed out changeset 2eaa68f10b19 (bug 1772313)
Backed out changeset 4a8b3ba193dc (bug 1772313)
2022-06-06 07:04:25 +03:00
Tooru Fujisawa
81758ebac5 Bug 1772360 - Remove unnecessary call for XPCOMUtils.defineLazyGlobalGetters. r=florian,webdriver-reviewers,webcompat-reviewers,denschub,whimboo
Differential Revision: https://phabricator.services.mozilla.com/D148168
2022-06-06 02:49:14 +00:00
Tooru Fujisawa
6ed0357d29 Bug 1772351 - Part 1: Stop calling ChromeUtils.defineModuleGetter for AppConstants. r=florian
Differential Revision: https://phabricator.services.mozilla.com/D148164
2022-06-06 02:49:12 +00:00
Tooru Fujisawa
dae44f3ce8 Bug 1772313 - Part 5: Stop calling ChromeUtils.defineModuleGetter for Services. r=kmag,webcompat-reviewers,pip-reviewers,denschub,mhowell
Differential Revision: https://phabricator.services.mozilla.com/D148158
2022-06-06 02:49:10 +00:00
Ben Hearsum
bbfdf38c8b Bug 1616355: attribution postSigningData should be per installation r=nalexander
about:newtab#asrouter actually depends on the ability to be able to write to postSigningData at the moment, so this will break that in any circumstance where the running Firefox cannot write to the installation directory. This code is only used for dev & qa testing though, and I've been told this is OK (and we may change how it works to avoid writing the file at all).

Differential Revision: https://phabricator.services.mozilla.com/D144167
2022-05-03 20:04:28 +00:00
Ben Hearsum
c3b0ed7d3d Bug 1756209: send Windows Store Campaign ID as attribution code r=nalexander
Nearly all the complexity here is abstracted away in the previous patch, so this is quite straightforward other than the one oddity noted in the comments.

Differential Revision: https://phabricator.services.mozilla.com/D143112
2022-04-26 15:55:34 +00:00
Kagami Sascha Rosylight
d6ec873f16 Bug 1703953 - Part 2: Apply mozilla/use-isInstance rules for .jsm files r=webdriver-reviewers,pip-reviewers,mhowell,Gijs,whimboo
This replaces all `instanceof` uses for DOM interfaces, since the operator in priviliged context works same as .isInstance().

Differential Revision: https://phabricator.services.mozilla.com/D141785
2022-04-06 11:57:57 +00:00
criss
0038e58f1f Backed out 3 changesets (bug 1703953) for causing multiple failures. CLOSED TREE
Backed out changeset 871a1fac289e (bug 1703953)
Backed out changeset 8151244bda18 (bug 1703953)
Backed out changeset eaf6d4c353be (bug 1703953)
2022-03-29 17:01:58 +03:00
Kagami Sascha Rosylight
f3126104f8 Bug 1703953 - Part 2: Apply mozilla/use-isInstance rules for .jsm files r=webdriver-reviewers,pip-reviewers,mhowell,Gijs,whimboo
This replaces all `instanceof` uses for DOM interfaces, since the operator in priviliged context works same as .isInstance().

Differential Revision: https://phabricator.services.mozilla.com/D141785
2022-03-29 13:11:00 +00:00
Mark Banner
7ea3c792f2 Bug 1758474 - Implement an ESLint rule to disallow passing {} as the target parameter for ChromeUtils.import. r=Gijs,mossop,perftest-reviewers,preferences-reviewers,sparky
Differential Revision: https://phabricator.services.mozilla.com/D140517
2022-03-11 16:41:29 +00:00
smolnar
19781d16c7 Backed out changeset 5018856d8fee (bug 1758474) for causing node eslint failure. CLOSED TREE 2022-03-10 11:58:45 +02:00
Mark Banner
de1754318e Bug 1758474 - Implement an ESLint rule to disallow passing {} as the target parameter for ChromeUtils.import. r=Gijs,mossop,perftest-reviewers,preferences-reviewers,sparky
Differential Revision: https://phabricator.services.mozilla.com/D140517
2022-03-10 09:25:28 +00:00
Barret Rennie
e7af55cb78 Bug 1745356 - Remove osfile.jsm from browser/components/attribution r=nalexander
Differential Revision: https://phabricator.services.mozilla.com/D133840
2021-12-25 20:59:24 +00:00
Evgenia Kotovich
fad1ce4207 Bug 1733535 - Replace OS.File with IOUtils in the attribution code. r=Standard8
Differential Revision: https://phabricator.services.mozilla.com/D128715
2021-10-25 08:38:52 +00:00
Norisz Fay
85ec9d324c Backed out changeset d63aab4e0688 (bug 1733535) for causing mochitest failures on browser_AttributionCode_telemetry.js CLOSED TREE 2021-10-22 16:32:43 +03:00
Evgenia Kotovich
af2f958525 Bug 1733535 - Replace OS.File with IOUtils in the attribution code. r=Standard8
Differential Revision: https://phabricator.services.mozilla.com/D128715
2021-10-22 11:57:44 +00:00
Kirk Steuber
7718339187 Bug 1677497 - Add support for the attribution parameter: dltoken r=nalexander,Mardak
Differential Revision: https://phabricator.services.mozilla.com/D100969
2021-01-20 01:08:43 +00:00
Nick Alexander
6eadb91247 Bug 1525076 - Part 4: Cache attribution code on macOS. r=mixedpuppy
Querying the macOS system quarantine database is relatively slow, so
we'd like to do it only once.  This is relevant because the Telemetry
subsystem causes the attribution data to be fetched relatively early
during startup.  By caching, we accept additional disk activity but
don't have to invoke an external process, query a database, etc.

We must make BROWSER_ATTRIBUTION_ERRORS histogram apply to macOS as
well as Windows.  We add error codes to capture macOS-specific
detalis.  And we push this out to a later cycle, since there's no
reason to revisit this immediately.

Differential Revision: https://phabricator.services.mozilla.com/D92695
2020-10-11 18:20:11 +00:00
Nick Alexander
11c100d5ee Bug 1525076 - Part 3: Add method to serialize attribution data. r=mixedpuppy
On Windows, we parse the attribution data as if it was the
`URL.search` parameter.  On macOS, we parse a URL slightly
differently: we parse it like `URL.searchParams` and we deduplicate
parameters.

For caching on macOS we need to serialize the parsed data.  We could
use either approach but we elect to write the code serialized as it
would be on Windows, since that is the one that is getting actively
tested in the wild.

Differential Revision: https://phabricator.services.mozilla.com/D92694
2020-10-11 18:19:22 +00:00
Nick Alexander
100c897d26 Bug 1525076 - Part 2: Add AttributionCode.writeAttributionFile. r=mixedpuppy
The next step is to have the cached attribution file location vary
depending on the platform, and to have the tests run on additional
platforms.

Differential Revision: https://phabricator.services.mozilla.com/D84264
2020-10-11 18:19:04 +00:00
Nick Alexander
5e69ce9766 Bug 1525076 - Part 1: Fix macOS native getReferrerUrl by invoking system sqlite binary (from JS). r=mixedpuppy
The issue this is addresses is that
[`CFURLCopyResourcePropertyForKey`](https://searchfox.org/mozilla-central/rev/ea7f70dac1c5fd18400f6d2a92679777d4b21492/xpcom/io/CocoaFileUtils.mm#212)
does not return quarantine data when launched as a GUI App.  What
happens is that launching via the GUI requires the user to override
GateKeeper by going to Security & Privacy > Open Anyway.  Doing that
updates the GateKeeper flags, and then the macOS API denies access:
once the GK flags reach some state, quarantine information is not
returned.  This is not documented (as far as I can see) but moons ago,
[somebody else on the internet witnessed the same
thing](https://cocoa-dev.apple.narkive.com/kkYeAC8o/is-it-possible-to-read-your-own-quarantine-info-after-launch).

To work around, we run the system SQLite binary, to fish the relevant
information out of the per-user quarantine database.  (SQLite is
installed by default on all relevant macOS versions.)

The most significant security concern I see is whether we can trust
this binary (in /usr/bin/sqlite3). Some discussion within the
Install/Update team suggested that an attacker who could corrupt or
modify that binary already had write access to the disk, which is an
attack vector equal to a totally compromised Firefox. If we determine
that we can't use the system SQLite binary, then we could use
Firefox's compiled copy of SQLite, but we might see versioning
issues. The system SQLite binary feels more robust.

This is implemented as a JS component for convenience, mostly: there
is no API for capturing output from `nsIProcess`.  It would be
possible to maintain the existing XPCOM contract by renaming the
existing contract and adding a contract with a JS implementation that
passes through to the renamed implementation, but it doesn't seem
worth the effort.

In the next commits, we will generalize the existing caching mechanism
form Windows to also apply to macOS. This is mostly a performance
optimization, so that we sniff a single well-known location rather
than launching a process at each startup, although there is a
correctness argument here as well, since the quarantine database is
dynamic and the attribution URL could expire.

Differential Revision: https://phabricator.services.mozilla.com/D92693
2020-10-11 18:18:36 +00:00
Nick Alexander
39987ecf12 Bug 1525076 - Part 0: Test more scenarios on macOS; handle URI component encoding. r=mixedpuppy
The existing code handles Windows and macOS quite differently.  On
macOS, the tests were not as comprehensive; this patch brings them
level and makes the form of the attribution data uniform on Windows
and macOS.  In particular, attribution data fields will now be URI
coded, just as they are on Windows.

This will allow us to re-use the parsing machinery we have on Windows
when we cache attribution codes on macOS.

Differential Revision: https://phabricator.services.mozilla.com/D92692
2020-10-11 18:18:11 +00:00
Razvan Maries
a84b6385b0 Backed out 6 changesets (bug 1525076) for perma failures on browser_startup_mainthreadio.js. CLOSED TREE
Backed out changeset a425e504ae7e (bug 1525076)
Backed out changeset 5163df5430c1 (bug 1525076)
Backed out changeset 4313cfb9dfcb (bug 1525076)
Backed out changeset 9178ebc95da2 (bug 1525076)
Backed out changeset 7bde7f6ccbd8 (bug 1525076)
Backed out changeset 6c2c2e07b95f (bug 1525076)
2020-10-09 22:51:36 +03:00
Nick Alexander
ca42485113 Bug 1525076 - Part 4: Cache attribution code on macOS. r=mixedpuppy
Querying the macOS system quarantine database is relatively slow, so
we'd like to do it only once.  This is relevant because the Telemetry
subsystem causes the attribution data to be fetched relatively early
during startup.  By caching, we accept additional disk activity but
don't have to invoke an external process, query a database, etc.

We must make BROWSER_ATTRIBUTION_ERRORS histogram apply to macOS as
well as Windows.  We add error codes to capture macOS-specific
detalis.  And we push this out to a later cycle, since there's no
reason to revisit this immediately.

Differential Revision: https://phabricator.services.mozilla.com/D92695
2020-10-09 17:28:51 +00:00
Nick Alexander
c5125585cf Bug 1525076 - Part 3: Add method to serialize attribution data. r=mixedpuppy
On Windows, we parse the attribution data as if it was the
`URL.search` parameter.  On macOS, we parse a URL slightly
differently: we parse it like `URL.searchParams` and we deduplicate
parameters.

For caching on macOS we need to serialize the parsed data.  We could
use either approach but we elect to write the code serialized as it
would be on Windows, since that is the one that is getting actively
tested in the wild.

Differential Revision: https://phabricator.services.mozilla.com/D92694
2020-10-09 17:28:48 +00:00
Nick Alexander
8d7760c5a7 Bug 1525076 - Part 2: Add AttributionCode.writeAttributionFile. r=mixedpuppy
The next step is to have the cached attribution file location vary
depending on the platform, and to have the tests run on additional
platforms.

Differential Revision: https://phabricator.services.mozilla.com/D84264
2020-10-09 17:28:46 +00:00
Nick Alexander
eb1a1d0a4d Bug 1525076 - Part 1: Fix macOS native getReferrerUrl by invoking system sqlite binary (from JS). r=mixedpuppy
The issue this is addresses is that
[`CFURLCopyResourcePropertyForKey`](https://searchfox.org/mozilla-central/rev/ea7f70dac1c5fd18400f6d2a92679777d4b21492/xpcom/io/CocoaFileUtils.mm#212)
does not return quarantine data when launched as a GUI App.  What
happens is that launching via the GUI requires the user to override
GateKeeper by going to Security & Privacy > Open Anyway.  Doing that
updates the GateKeeper flags, and then the macOS API denies access:
once the GK flags reach some state, quarantine information is not
returned.  This is not documented (as far as I can see) but moons ago,
[somebody else on the internet witnessed the same
thing](https://cocoa-dev.apple.narkive.com/kkYeAC8o/is-it-possible-to-read-your-own-quarantine-info-after-launch).

To work around, we run the system SQLite binary, to fish the relevant
information out of the per-user quarantine database.  (SQLite is
installed by default on all relevant macOS versions.)

The most significant security concern I see is whether we can trust
this binary (in /usr/bin/sqlite3). Some discussion within the
Install/Update team suggested that an attacker who could corrupt or
modify that binary already had write access to the disk, which is an
attack vector equal to a totally compromised Firefox. If we determine
that we can't use the system SQLite binary, then we could use
Firefox's compiled copy of SQLite, but we might see versioning
issues. The system SQLite binary feels more robust.

This is implemented as a JS component for convenience, mostly: there
is no API for capturing output from `nsIProcess`.  It would be
possible to maintain the existing XPCOM contract by renaming the
existing contract and adding a contract with a JS implementation that
passes through to the renamed implementation, but it doesn't seem
worth the effort.

In the next commits, we will generalize the existing caching mechanism
form Windows to also apply to macOS. This is mostly a performance
optimization, so that we sniff a single well-known location rather
than launching a process at each startup, although there is a
correctness argument here as well, since the quarantine database is
dynamic and the attribution URL could expire.

Differential Revision: https://phabricator.services.mozilla.com/D92693
2020-10-09 17:27:13 +00:00
Nick Alexander
2326bf8eb4 Bug 1525076 - Part 0: Test more scenarios on macOS; handle URI component encoding. r=mixedpuppy
The existing code handles Windows and macOS quite differently.  On
macOS, the tests were not as comprehensive; this patch brings them
level and makes the form of the attribution data uniform on Windows
and macOS.  In particular, attribution data fields will now be URI
coded, just as they are on Windows.

This will allow us to re-use the parsing machinery we have on Windows
when we cache attribution codes on macOS.

Differential Revision: https://phabricator.services.mozilla.com/D92692
2020-10-09 17:26:38 +00:00
Razvan Maries
8cb537bb4d Backed out 6 changesets (bug 1525076) for ES Lint failure. CLOSED TREE
Backed out changeset 912fcc3cb274 (bug 1525076)
Backed out changeset 690d730341c6 (bug 1525076)
Backed out changeset 5dd08176812c (bug 1525076)
Backed out changeset 963c8d33d779 (bug 1525076)
Backed out changeset f7d26b270884 (bug 1525076)
Backed out changeset 95fd52531439 (bug 1525076)
2020-10-09 20:17:23 +03:00
Nick Alexander
d639721f77 Bug 1525076 - Part 4: Cache attribution code on macOS. r=mixedpuppy
Querying the macOS system quarantine database is relatively slow, so
we'd like to do it only once.  This is relevant because the Telemetry
subsystem causes the attribution data to be fetched relatively early
during startup.  By caching, we accept additional disk activity but
don't have to invoke an external process, query a database, etc.

We must make BROWSER_ATTRIBUTION_ERRORS histogram apply to macOS as
well as Windows.  We add error codes to capture macOS-specific
detalis.  And we push this out to a later cycle, since there's no
reason to revisit this immediately.

Differential Revision: https://phabricator.services.mozilla.com/D92695
2020-10-09 16:52:26 +00:00
Nick Alexander
72e12d93f7 Bug 1525076 - Part 3: Add method to serialize attribution data. r=mixedpuppy
On Windows, we parse the attribution data as if it was the
`URL.search` parameter.  On macOS, we parse a URL slightly
differently: we parse it like `URL.searchParams` and we deduplicate
parameters.

For caching on macOS we need to serialize the parsed data.  We could
use either approach but we elect to write the code serialized as it
would be on Windows, since that is the one that is getting actively
tested in the wild.

Differential Revision: https://phabricator.services.mozilla.com/D92694
2020-10-09 16:51:48 +00:00
Nick Alexander
f321c177e1 Bug 1525076 - Part 2: Add AttributionCode.writeAttributionFile. r=mixedpuppy
The next step is to have the cached attribution file location vary
depending on the platform, and to have the tests run on additional
platforms.

Differential Revision: https://phabricator.services.mozilla.com/D84264
2020-10-09 16:51:41 +00:00
Nick Alexander
be45f40052 Bug 1525076 - Part 1: Fix macOS native getReferrerUrl by invoking system sqlite binary (from JS). r=mixedpuppy
The issue this is addresses is that
[`CFURLCopyResourcePropertyForKey`](https://searchfox.org/mozilla-central/rev/ea7f70dac1c5fd18400f6d2a92679777d4b21492/xpcom/io/CocoaFileUtils.mm#212)
does not return quarantine data when launched as a GUI App.  What
happens is that launching via the GUI requires the user to override
GateKeeper by going to Security & Privacy > Open Anyway.  Doing that
updates the GateKeeper flags, and then the macOS API denies access:
once the GK flags reach some state, quarantine information is not
returned.  This is not documented (as far as I can see) but moons ago,
[somebody else on the internet witnessed the same
thing](https://cocoa-dev.apple.narkive.com/kkYeAC8o/is-it-possible-to-read-your-own-quarantine-info-after-launch).

To work around, we run the system SQLite binary, to fish the relevant
information out of the per-user quarantine database.  (SQLite is
installed by default on all relevant macOS versions.)

The most significant security concern I see is whether we can trust
this binary (in /usr/bin/sqlite3). Some discussion within the
Install/Update team suggested that an attacker who could corrupt or
modify that binary already had write access to the disk, which is an
attack vector equal to a totally compromised Firefox. If we determine
that we can't use the system SQLite binary, then we could use
Firefox's compiled copy of SQLite, but we might see versioning
issues. The system SQLite binary feels more robust.

This is implemented as a JS component for convenience, mostly: there
is no API for capturing output from `nsIProcess`.  It would be
possible to maintain the existing XPCOM contract by renaming the
existing contract and adding a contract with a JS implementation that
passes through to the renamed implementation, but it doesn't seem
worth the effort.

In the next commits, we will generalize the existing caching mechanism
form Windows to also apply to macOS. This is mostly a performance
optimization, so that we sniff a single well-known location rather
than launching a process at each startup, although there is a
correctness argument here as well, since the quarantine database is
dynamic and the attribution URL could expire.

Differential Revision: https://phabricator.services.mozilla.com/D92693
2020-10-09 16:51:39 +00:00
Nick Alexander
993a6837c1 Bug 1525076 - Part 0: Test more scenarios on macOS; handle URI component encoding. r=mixedpuppy
The existing code handles Windows and macOS quite differently.  On
macOS, the tests were not as comprehensive; this patch brings them
level and makes the form of the attribution data uniform on Windows
and macOS.  In particular, attribution data fields will now be URI
coded, just as they are on Windows.

This will allow us to re-use the parsing machinery we have on Windows
when we cache attribution codes on macOS.

Differential Revision: https://phabricator.services.mozilla.com/D92692
2020-10-09 16:51:36 +00:00
Cosmin Sabou
51278a36a5 Backed out 6 changesets (bug 1525076) for OSX xpcshell failures on test_ASRouterTargeting_attribution. CLOSED TREE
Backed out changeset 25fd5f2e276e (bug 1525076)
Backed out changeset 4f2c3fe9ee78 (bug 1525076)
Backed out changeset e7e05723ceae (bug 1525076)
Backed out changeset d0c81fd932c7 (bug 1525076)
Backed out changeset c1629391c635 (bug 1525076)
Backed out changeset e6b09ef2a26c (bug 1525076)
2020-10-09 07:39:32 +03:00
Nick Alexander
ba34c2e0cd Bug 1525076 - Part 4: Cache attribution code on macOS. r=mixedpuppy
Querying the macOS system quarantine database is relatively slow, so
we'd like to do it only once.  This is relevant because the Telemetry
subsystem causes the attribution data to be fetched relatively early
during startup.  By caching, we accept additional disk activity but
don't have to invoke an external process, query a database, etc.

We must make BROWSER_ATTRIBUTION_ERRORS histogram apply to macOS as
well as Windows.  We add error codes to capture macOS-specific
detalis.  And we push this out to a later cycle, since there's no
reason to revisit this immediately.

Differential Revision: https://phabricator.services.mozilla.com/D92695
2020-10-08 21:54:16 +00:00
Nick Alexander
0ef0aabf15 Bug 1525076 - Part 3: Add method to serialize attribution data. r=mixedpuppy
On Windows, we parse the attribution data as if it was the
`URL.search` parameter.  On macOS, we parse a URL slightly
differently: we parse it like `URL.searchParams` and we deduplicate
parameters.

For caching on macOS we need to serialize the parsed data.  We could
use either approach but we elect to write the code serialized as it
would be on Windows, since that is the one that is getting actively
tested in the wild.

Differential Revision: https://phabricator.services.mozilla.com/D92694
2020-10-08 21:30:09 +00:00
Nick Alexander
ba282a56de Bug 1525076 - Part 2: Add AttributionCode.writeAttributionFile. r=mixedpuppy
The next step is to have the cached attribution file location vary
depending on the platform, and to have the tests run on additional
platforms.

Differential Revision: https://phabricator.services.mozilla.com/D84264
2020-10-08 21:54:49 +00:00
Nick Alexander
87ec636101 Bug 1525076 - Part 1: Fix macOS native getReferrerUrl by invoking system sqlite binary (from JS). r=mixedpuppy
The issue this is addresses is that
[`CFURLCopyResourcePropertyForKey`](https://searchfox.org/mozilla-central/rev/ea7f70dac1c5fd18400f6d2a92679777d4b21492/xpcom/io/CocoaFileUtils.mm#212)
does not return quarantine data when launched as a GUI App.  What
happens is that launching via the GUI requires the user to override
GateKeeper by going to Security & Privacy > Open Anyway.  Doing that
updates the GateKeeper flags, and then the macOS API denies access:
once the GK flags reach some state, quarantine information is not
returned.  This is not documented (as far as I can see) but moons ago,
[somebody else on the internet witnessed the same
thing](https://cocoa-dev.apple.narkive.com/kkYeAC8o/is-it-possible-to-read-your-own-quarantine-info-after-launch).

To work around, we run the system SQLite binary, to fish the relevant
information out of the per-user quarantine database.  (SQLite is
installed by default on all relevant macOS versions.)

The most significant security concern I see is whether we can trust
this binary (in /usr/bin/sqlite3). Some discussion within the
Install/Update team suggested that an attacker who could corrupt or
modify that binary already had write access to the disk, which is an
attack vector equal to a totally compromised Firefox. If we determine
that we can't use the system SQLite binary, then we could use
Firefox's compiled copy of SQLite, but we might see versioning
issues. The system SQLite binary feels more robust.

This is implemented as a JS component for convenience, mostly: there
is no API for capturing output from `nsIProcess`.  It would be
possible to maintain the existing XPCOM contract by renaming the
existing contract and adding a contract with a JS implementation that
passes through to the renamed implementation, but it doesn't seem
worth the effort.

In the next commits, we will generalize the existing caching mechanism
form Windows to also apply to macOS. This is mostly a performance
optimization, so that we sniff a single well-known location rather
than launching a process at each startup, although there is a
correctness argument here as well, since the quarantine database is
dynamic and the attribution URL could expire.

Differential Revision: https://phabricator.services.mozilla.com/D92693
2020-10-08 21:29:35 +00:00
Nick Alexander
2be713287f Bug 1525076 - Part 0: Test more scenarios on macOS; handle URI component encoding. r=mixedpuppy
The existing code handles Windows and macOS quite differently.  On
macOS, the tests were not as comprehensive; this patch brings them
level and makes the form of the attribution data uniform on Windows
and macOS.  In particular, attribution data fields will now be URI
coded, just as they are on Windows.

This will allow us to re-use the parsing machinery we have on Windows
when we cache attribution codes on macOS.

Differential Revision: https://phabricator.services.mozilla.com/D92692
2020-10-08 21:29:12 +00:00
Andrei Oprea
e2f56d0415 Bug 1621402 - Add telemetry to capture failures to read attribution data r=janerik,mixedpuppy
Differential Revision: https://phabricator.services.mozilla.com/D67163
2020-03-24 18:32:57 +00:00
Ed Lee
d663021fa5 Bug 1595063 - Support user agent attribution codes r=mixedpuppy,nanj
Add "ua" and increase allowed length for attribution codes. Allows asrouter devtools to set all keys.

Differential Revision: https://phabricator.services.mozilla.com/D57906
2019-12-24 17:04:10 +00:00