Cannot jetify library with invalid commented out proguard rules [134100420]

Fixed

Bug

Status Update

No update yet.

Description

ca...@gmail.com

created issue #1

May 31, 2019 11:37AM

OS: macOS 10.14.5
Java version "1.8.0_121"
Jetifier version: 1.0.0-beta04

Let there be a Library Foo with the following 'proguard-rules.txt' file:

# DBFlow
-keep class * extends com.raizlabs.android.dbflow.config.DatabaseHolder { *; }
-dontwarn com.raizlabs.android.dbflow.**
# Note: the configuration keeps the entry point 'com.raizlabs.android.dbflow.config.FlowManager$GlobalDatabaseHolder { FlowManager$GlobalDatabaseHolder(com.raizlabs.android.dbflow.config.FlowManager$1); }', but not the descriptor class 'com.raizlabs.android.dbflow.config.FlowManager$1'
-keep class com.raizlabs.android.dbflow.config.FlowManager$*

When trying to jetify this library:

./jetifier-standalone -i /Users/someuser/Downloads/foo/foo.aar -o newfoo.aar

The Jetifier fails with the following output:

Exception in thread "main" java.util.regex.PatternSyntaxException: Illegal repetition near index 147
com/raizlabs/android/dbflow/.*
# Note: the configuration keeps the entry point 'com/raizlabs/android/dbflow/config/FlowManager$GlobalDatabaseHolder { FlowManager$GlobalDatabaseHolder(com/raizlabs/android/dbflow/config/FlowManager$1); }'
^
at java.util.regex.Pattern.error(Pattern.java:1955)
at java.util.regex.Pattern.closure(Pattern.java:3157)
at java.util.regex.Pattern.sequence(Pattern.java:2134)
at java.util.regex.Pattern.expr(Pattern.java:1996)
at java.util.regex.Pattern.compile(Pattern.java:1696)
at java.util.regex.Pattern.<init>(Pattern.java:1351)
at java.util.regex.Pattern.compile(Pattern.java:1028)
at com.android.tools.build.jetifier.core.type.TypesMap.matchOldProguardForNewTypes(TypesMap.kt:112)
[...]

If i remove the notes, the new 'proguard-rules.txt' looks like this:

# DBFlow
-keep class * extends com.raizlabs.android.dbflow.config.DatabaseHolder { *; }
-dontwarn com.raizlabs.android.dbflow.**
-keep class com.raizlabs.android.dbflow.config.FlowManager$*

And this runs through the jetifier just fine. Please provide a fix for this issue, since i have to patch my dependencies currently in order to make AndroidX work for me.

Comments

do...@traveloka.com <do...@traveloka.com> #2Jul 9, 2019 09:39AM

Thank you for this feedback. The team may reach out for more information on triaging or reproducing this issue.

pa...@gmail.com <pa...@gmail.com> #3Oct 17, 2019 09:40AM

pallavi@ - Have we done any testing on Win7? Looks like this issue should be easily reproducible.

pa...@gmail.com <pa...@gmail.com> #4Oct 17, 2019 09:44AM

a.sworowska: could you click the "view problem details" button and send us what that shows?

zyy: is there any other useful information about a crash on Windows that's easily available?

ga...@google.com <ga...@google.com> #5Oct 17, 2019 11:41AM

Assigned to pa...@google.com.

Problem signature:
Problem Event Name: APPCRASH
Application Name: adb.exe
Application Version: 0.0.0.0
Application Timestamp: 34f534d5
Fault Module Name: adb.exe
Fault Module Version: 0.0.0.0
Fault Module Timestamp: 34f534d5
Exception Code: 40000015
Exception Offset: 00016717
OS Version: 6.1.7601.2.1.0.256.48
Locale ID: 1033
Additional Information 1: 7763
Additional Information 2: 7763cd91f17ba1394e15130ad28436a8
Additional Information 3: 622c
Additional Information 4: 622c3aed538ac58f09882fe423edc57a

Also there's a console output: (for adb devices)
List of devices attached
* daemon not running. starting it now at tcp:5037 *
error: cannot open transport registration socketpair: Input/output error

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.

pa...@google.com <pa...@google.com> #6Oct 25, 2019 03:40PM

Accepted by pa...@google.com.

We have the same problem at my workplace (computers running with the same configuration).
Changing the port (5037) didn't solve.
We were thinking about a conflict with the company's proxy but we gave up to find a solution as soon as we downgrade to 25.0.3 and understood that the update was responsible of the crash.

ap...@google.com <ap...@google.com> #7Oct 31, 2019 06:24PM

>zyy: is there any other useful information about a crash on Windows that's easily available?

One can generate a minidump file on crash, here's instructions (WER part):

https://blogs.msdn.microsoft.com/chaun/2013/11/12/steps-to-catch-a-simple-crash-dump-of-a-crashing-process/

Other option would be to integrate Breakpad (

https://chromium.googlesource.com/breakpad/breakpad/); it will take much longer but would allow for automatic / user-initiated crash information submission

pa...@google.com <pa...@google.com> #8Nov 26, 2019 02:34PM

I did the WER method and generate a dmp file (2 in fact because of the DumpCount set to 2... Should I set it higher ?)
But it is a 40MB file which I can't upload. Any suggestion ?

pa...@gmail.com <pa...@gmail.com> #9Nov 26, 2019 02:40PM

Thank you so much! You can either archive the dump with 7zip or plain Zip - .dmp files pack really well. Or you could upload it to some file share, e.g. Google Drive, and give a link here. Anything would work

hu...@google.com <hu...@google.com> #10Nov 27, 2019 05:34PM

Marked as fixed.

the stack overflow match for "Exception Code: 40000015" says it's what happens when you call abort(3).

that combined with the error message points to this:

if (adb_socketpair(s)) {
fatal_errno("cannot open transport registration socketpair");

unfortunately adb_socketpair only writes its failures to the debug log.

i don't think we need a dmp file, but can someone run adb with the ADB_TRACE environment variable set to "all" so we can see the underlying error? we're looking for an error that starts "adb_socketpair:". (and our QA folks did test on Windows 7 but didn't hit this.)

(i'm being deliberately vague about how to set an environment variable on Windows because i assume everyone reading this knows better than i do!)