C#: Sign analysis by tamasvajk · Pull Request #4193 · github/codeql · GitHub

tamasvajk · 2020-09-02T14:52:55Z

Sign analysis ported from Java. It is going to be used by the ported range analysis. The public API contains four predicates: positive(Expr e), negative(Expr e), strictlyPositive(Expr e), strictlyNegative(Expr e). Range analysis doesn't need an isZero(Expr e), so it's not added for the time being.

The code is a 1:1 migration from the Java implementation. Java had no direct tests on it, I added a single one for it, because I was anyways debugging through the Java code. I added a lot more test for C#.

csharp/ql/src/semmle/code/csharp/dataflow/SignAnalysis.qll

tamasvajk · 2020-09-02T15:12:44Z

csharp/ql/src/semmle/code/csharp/dataflow/SignAnalysis.qll

+}
+
+private predicate containerSizeAccess(PropertyAccess pa) {
+  propertyOverrides(pa.getTarget(), "System.Collections.Generic.ICollection<>", "Count") or


This should probably go to a helper file, because we use it in ContainerSizeCmpZero.ql too. But I couldn't find a good location for it. (Also, the check is using a slightly different variant.

csharp/ql/src/semmle/code/csharp/dataflow/SignAnalysis.qll

tamasvajk · 2020-09-02T15:34:59Z

csharp/ql/src/semmle/code/csharp/dataflow/SignAnalysis.qll

+  exists(AssignOperation a | a.getLValue() = f.getAnAccess() | result = exprSign(a))
+  or
+  f.fromSource() and not exists(f.getInitializer()) and result = TZero()
+}


Java also handles access to the field through reflection.

csharp/ql/test/library-tests/dataflow/signanalysis/SignAnalysis.cs

csharp/ql/src/semmle/code/csharp/dataflow/SignAnalysis.qll

csharp/ql/src/semmle/code/csharp/dataflow/internal/rangeanalysis/RangeUtils.qll

hvitved

Awesome work, @tamasvajk . Great to have this logic shared with Java.

csharp/ql/src/semmle/code/csharp/dataflow/internal/rangeanalysis/SsaReadPositionSpecific.qll

csharp/ql/src/semmle/code/csharp/dataflow/internal/rangeanalysis/SsaReadPositionCommon.qll

hvitved · 2020-09-16T07:41:59Z

csharp/ql/src/semmle/code/csharp/dataflow/internal/rangeanalysis/SignAnalysisCommon.qll

+      not exists(SsaReadPositionBlock bb | getAnExpression(bb) = e) and
+      result = ssaDefSign(v)


I'm not quite sure why this disjunct is needed. If it is, could we have a test case?

@aschackmull can you help me out with this question? This is the disjunct before the refactoring: https://github.com/github/codeql/blob/main/java/ql/src/semmle/code/java/dataflow/SignAnalysis.qll#L482.

If you look at the disjuncts surrounding this case then you can see some hints. We split into cases based on Expr e and these first couple of cases all cover the VarAccess cases. The most precise case is if this is a well-behaved read of an SSA variable, then there are also cases for non-SSA variable accesses, which are split into fields and non-fields. However, the SSA case would like to assume the existence of a BasicBlock containing the VarAccess in order to look at the SsaReadPosition, and there are a few cases where the Java SSA library can produce an (SSA-variable, SSA-read) pair outside the CFG, for instance in compile-time evaluated expressions in case constants. So in order for all VarAccess expressions to receive a sign we have to include this case.

OK, I thought it was weird to have both getARead(v) = e and not exists(SsaReadPositionBlock bb | getAnExpression(bb) = e) at the same time, i.e. an SSA read without a basic block. It sounds like a corner case in Java then.

csharp/ql/src/semmle/code/csharp/dataflow/internal/rangeanalysis/SignAnalysisCommon.qll

csharp/ql/test/library-tests/dataflow/signanalysis/SignAnalysis.cs

csharp/ql/src/semmle/code/csharp/dataflow/internal/rangeanalysis/SignAnalysisSpecific.qll

tamasvajk · 2020-09-17T10:48:00Z

@hvitved I pushed some extra commits to address your review. The unsigned integer improvements are not done yet. I think I covered the rest.

hvitved

Some of the private imports are potentially breaking changes, so should be reverted.

csharp/ql/src/semmle/code/csharp/dataflow/SSA.qll

java/ql/src/semmle/code/java/Collections.qll

java/ql/src/semmle/code/java/Maps.qll

java/ql/src/semmle/code/java/Reflection.qll

java/ql/src/semmle/code/java/controlflow/Guards.qll

java/ql/src/semmle/code/java/dataflow/RangeUtils.qll

java/ql/src/semmle/code/java/dataflow/SSA.qll

csharp/ql/src/semmle/code/csharp/commons/ComparisonTest.qll

csharp/ql/src/semmle/code/csharp/dataflow/internal/rangeanalysis/SignAnalysisSpecific.qll

tamasvajk · 2020-09-17T12:26:22Z

@hvitved I reverted the last commit, and only made the imports private in the new code, so that I don't introduce breaking changes.

hvitved · 2020-09-17T13:02:34Z

The unsigned integer improvements are not done yet

If you prefer, we could do that in a follow-up PR.

tamasvajk · 2020-09-17T14:48:33Z

I don't mind handling the unsigned integers in this PR. I'll work on that tomorrow.

Also, I pushed some extra test cases that are not covered. I'll fix them, and try to find more such cases.

csharp/ql/src/semmle/code/csharp/dataflow/internal/rangeanalysis/SignAnalysisSpecific.qll

hvitved · 2020-09-18T06:49:44Z

csharp/ql/src/semmle/code/csharp/dataflow/internal/rangeanalysis/SignAnalysisCommon.qll

-    exists(string charlit | charlit = e.(CharacterLiteral).getValue() |
-      if charlit = "\\0" or charlit = "\\u0000" then result = TZero() else result = TPos()
+    exists(string charlit | charlit = getCharValue(e) |
+      if charlit = "\\0" or charlit = "\\u0000" or charlit.regexpMatch("\\u0000")


Why do we need both disjuncts charlit = "\\u0000" or charlit.regexpMatch("\\u0000"?

This is shared between Java and C#. And I don't know what gets extracted for Java. @aschackmull Can you tell me what gets extracted for (char)0 in Java? See this ticket and the slack discussion for context: https://github.com/github/codeql-coreql-team/issues/539

(char) 0 is just a cast of a 32 bit literal to an unsigned 16 bit value as far as I know. In Java you can put NUL characters in the source code. You can write (where ^@ represents a real unescaped NUL symbol):

class X { String normal = "\0"; String unicode = "\u0000"; String nullChar = "^@"; }

I had expected unicode and nullChar to be extracted similarly because as far as I know unicode escape sequences are replaced before lexical analysis. However, in the extractor output the normal and unicode lines are extracted similarly, while the last is different.

The generated TRAP output is

#10000=@"/home/arthur/Devel/code/ql/java/ql/test/query-tests/Nullness/X.java;sourcefile" files(#10000,"/home/arthur/Devel/code/ql/java/ql/test/query-tests/Nullness/X.java","X","java",0) #10001=@"/home/arthur/Devel/code/ql/java/ql/test/query-tests/Nullness;folder" folders(#10001,"/home/arthur/Devel/code/ql/java/ql/test/query-tests/Nullness","Nullness") #10002=@"/home/arthur/Devel/code/ql/java/ql/test/query-tests;folder" folders(#10002,"/home/arthur/Devel/code/ql/java/ql/test/query-tests","query-tests") #10003=@"/home/arthur/Devel/code/ql/java/ql/test;folder" folders(#10003,"/home/arthur/Devel/code/ql/java/ql/test","test") #10004=@"/home/arthur/Devel/code/ql/java/ql;folder" folders(#10004,"/home/arthur/Devel/code/ql/java/ql","ql") #10005=@"/home/arthur/Devel/code/ql/java;folder" folders(#10005,"/home/arthur/Devel/code/ql/java","java") #10006=@"/home/arthur/Devel/code/ql;folder" folders(#10006,"/home/arthur/Devel/code/ql","ql") #10007=@"/home/arthur/Devel/code;folder" folders(#10007,"/home/arthur/Devel/code","code") #10008=@"/home/arthur/Devel;folder" folders(#10008,"/home/arthur/Devel","Devel") #10009=@"/home/arthur;folder" folders(#10009,"/home/arthur","arthur") #10010=@"/home;folder" folders(#10010,"/home","home") #10011=@"/;folder" folders(#10011,"/","") containerparent(#10011,#10010) containerparent(#10010,#10009) containerparent(#10009,#10008) containerparent(#10008,#10007) containerparent(#10007,#10006) containerparent(#10006,#10005) containerparent(#10005,#10004) containerparent(#10004,#10003) containerparent(#10003,#10002) containerparent(#10002,#10001) containerparent(#10001,#10000) #10012=@"loc,{#10000},0,0,0,0" locations_default(#10012,#10000,0,0,0,0) hasLocation(#10000,#10012) numlines(#10000,6,5,0) #10013=@"package;" packages(#10013,"") cupackage(#10000,#10013) #10014=@"class;X" #10015=@"field;{#10014};nullChar" #10016=@"class;java.lang.String" fields(#10015,"nullChar",#10016,#10014,#10015) #10017=@"/home/arthur/Devel/code/ql/java/ql/test/query-tests/Nullness/X.class;sourcefile" files(#10017,"/home/arthur/Devel/code/ql/java/ql/test/query-tests/Nullness/X.class","X","class",0) containerparent(#10001,#10017) packages(#10013,"") cupackage(#10017,#10013) #10018=@"loc,{#10017},0,0,0,0" locations_default(#10018,#10017,0,0,0,0) hasLocation(#10017,#10018) hasLocation(#10015,#10018) #10019=@"field;{#10014};unicode" fields(#10019,"unicode",#10016,#10014,#10019) hasLocation(#10019,#10018) #10020=@"field;{#10014};normal" fields(#10020,"normal",#10016,#10014,#10020) hasLocation(#10020,#10018) #10021=@"type;void" primitives(#10021,"void") #10022=@"unknown;sourcefile" files(#10022,"","","",0) #10023=@"loc,{#10022},0,0,0,0" locations_default(#10023,#10022,0,0,0,0) hasLocation(#10021,#10023) #10024=@"callable;{#10014}.<init>(){#10021}" constrs(#10024,"X","X()",#10021,#10014,#10024) hasLocation(#10024,#10018) isDefConstr(#10024) #10025=@"loc,{#10000},1,7,1,7" locations_default(#10025,#10000,1,7,1,7) hasLocation(#10014,#10025) numlines(#10014,5,4,0) #10026=* methods(#10026,"<obinit>","<obinit>()",#10021,#10014,#10026) #10027=@"modifier;private" modifiers(#10027,"private") hasModifier(#10026,#10027) locations_default(#10025,#10000,1,7,1,7) hasLocation(#10026,#10025) #10028=* stmts(#10028,0,#10026,0,#10026) locations_default(#10025,#10000,1,7,1,7) hasLocation(#10028,#10025) #10029=* #10030=* #10031=* stmts(#10029,14,#10028,0,#10026) #10032=* locations_default(#10032,#10000,3,1,3,21) hasLocation(#10029,#10032) exprs(#10030,4,#10016,#10029,0) callableEnclosingExpr(#10030,#10026) statementEnclosingExpr(#10030,#10029) #10033=* locations_default(#10033,#10000,3,1,3,21) hasLocation(#10030,#10033) exprs(#10031,60,#10016,#10030,0) callableEnclosingExpr(#10031,#10026) statementEnclosingExpr(#10031,#10029) #10034=* locations_default(#10034,#10000,3,1,3,21) hasLocation(#10031,#10034) variableBinding(#10031,#10020) #10035=* exprs(#10035,22,#10016,#10030,1) callableEnclosingExpr(#10035,#10026) statementEnclosingExpr(#10035,#10029) #10036=* locations_default(#10036,#10000,3,17,3,20) hasLocation(#10035,#10036) numlines(#10035,1,1,0) namestrings("""\0""","^@",#10035) #10037=* #10038=* #10039=* stmts(#10037,14,#10028,1,#10026) #10040=* locations_default(#10040,#10000,4,1,4,26) hasLocation(#10037,#10040) exprs(#10038,4,#10016,#10037,0) callableEnclosingExpr(#10038,#10026) statementEnclosingExpr(#10038,#10037) #10041=* locations_default(#10041,#10000,4,1,4,26) hasLocation(#10038,#10041) exprs(#10039,60,#10016,#10038,0) callableEnclosingExpr(#10039,#10026) statementEnclosingExpr(#10039,#10037) #10042=* locations_default(#10042,#10000,4,1,4,26) hasLocation(#10039,#10042) variableBinding(#10039,#10019) #10043=* exprs(#10043,22,#10016,#10038,1) callableEnclosingExpr(#10043,#10026) statementEnclosingExpr(#10043,#10037) #10044=* locations_default(#10044,#10000,4,18,4,25) hasLocation(#10043,#10044) numlines(#10043,1,1,0) namestrings("""\u0000""","^@",#10043) #10045=* #10046=* #10047=* stmts(#10045,14,#10028,2,#10026) #10048=* locations_default(#10048,#10000,5,1,5,22) hasLocation(#10045,#10048) exprs(#10046,4,#10016,#10045,0) callableEnclosingExpr(#10046,#10026) statementEnclosingExpr(#10046,#10045) #10049=* locations_default(#10049,#10000,5,1,5,22) hasLocation(#10046,#10049) exprs(#10047,60,#10016,#10046,0) callableEnclosingExpr(#10047,#10026) statementEnclosingExpr(#10047,#10045) #10050=* locations_default(#10050,#10000,5,1,5,22) hasLocation(#10047,#10050) variableBinding(#10047,#10015) #10051=* exprs(#10051,22,#10016,#10046,1) callableEnclosingExpr(#10051,#10026) statementEnclosingExpr(#10051,#10045) #10052=* locations_default(#10052,#10000,5,19,5,21) hasLocation(#10051,#10052) numlines(#10051,1,1,0) namestrings("""^@""","^@",#10051) locations_default(#10025,#10000,1,7,1,7) hasLocation(#10024,#10025) numlines(#10024,1,1,0) #10053=* stmts(#10053,0,#10024,0,#10024) #10054=* locations_default(#10054,#10000,1,7,1,7) hasLocation(#10053,#10054) numlines(#10053,1,1,0) #10055=* stmts(#10055,20,#10053,0,#10024) #10056=* locations_default(#10056,#10000,1,7,1,7) hasLocation(#10055,#10056) numlines(#10055,1,1,0) #10057=* stmts(#10057,14,#10053,1,#10024) #10058=* locations_default(#10058,#10000,1,7,1,7) hasLocation(#10057,#10058) numlines(#10057,1,1,0) #10059=* exprs(#10059,61,#10021,#10057,0) callableEnclosingExpr(#10059,#10024) statementEnclosingExpr(#10059,#10057) #10060=* locations_default(#10060,#10000,1,7,1,7) hasLocation(#10059,#10060) numlines(#10059,1,1,0) callableBinding(#10059,#10026) #10061=@"class;java.lang.Object" #10062=@"callable;{#10061}.<init>(){#10021}" callableBinding(#10055,#10062) #10063=* fielddecls(#10063,#10014) #10064=* locations_default(#10064,#10000,3,1,3,21) hasLocation(#10063,#10064) #10065=* exprs(#10065,62,#10016,#10063,0) #10066=* locations_default(#10066,#10000,3,1,3,6) hasLocation(#10065,#10066) numlines(#10065,1,1,0) #10067=@"loc,{#10000},3,8,3,13" locations_default(#10067,#10000,3,8,3,13) hasLocation(#10020,#10067) fieldDeclaredIn(#10020,#10063,0) #10068=* fielddecls(#10068,#10014) #10069=* locations_default(#10069,#10000,4,1,4,26) hasLocation(#10068,#10069) #10070=* exprs(#10070,62,#10016,#10068,0) #10071=* locations_default(#10071,#10000,4,1,4,6) hasLocation(#10070,#10071) numlines(#10070,1,1,0) #10072=@"loc,{#10000},4,8,4,14" locations_default(#10072,#10000,4,8,4,14) hasLocation(#10019,#10072) fieldDeclaredIn(#10019,#10068,0) #10073=* fielddecls(#10073,#10014) #10074=* locations_default(#10074,#10000,5,1,5,22) hasLocation(#10073,#10074) #10075=* exprs(#10075,62,#10016,#10073,0) #10076=* locations_default(#10076,#10000,5,1,5,6) hasLocation(#10075,#10076) numlines(#10075,1,1,0) #10077=@"loc,{#10000},5,8,5,15" locations_default(#10077,#10000,5,8,5,15) hasLocation(#10015,#10077) fieldDeclaredIn(#10015,#10073,0)

had expected unicode and nullChar to be extracted similarly because as far as I know unicode escape sequences are replaced before lexical analysis. However, in the extractor output the normal and unicode lines are extracted similarly, while the last is different.

Really? The TRAP you quote shows the three cases with identical values, right?

namestrings("""\0""","^@",#10035) namestrings("""\u0000""","^@",#10043) namestrings("""^@""","^@",#10051)

I think perhaps some of the Java QL code here predates an improvement that I made a long time ago to the handling of the value of string literals, which precisely gives this uniform representation.

I believe the conclusion in slack was that we'd only need the charlit.regexpMatch("\\u0000") case and that the others currently don't work?

@aschackmull Yes, you're right. The things get extracted the same way in all three cases. I misread things in my terminal before I replaced the NUL character with something more easy to read ;-)

csharp/ql/src/semmle/code/csharp/dataflow/internal/rangeanalysis/SignAnalysisSpecific.qll

tamasvajk · 2020-09-18T10:19:45Z

Differences job: https://jenkins.internal.semmle.com/job/Changes/job/CSharp-Differences/439/

hvitved · 2020-09-20T09:29:55Z

Differences job: https://jenkins.internal.semmle.com/job/Changes/job/CSharp-Differences/439/

This confirmed the performance issue caused by the guards logic change, as expected. I have started new jobs based on a rebased branch:

csharp/ql/test/library-tests/dataflow/signanalysis/MissingSign.expected

csharp/ql/src/semmle/code/csharp/dataflow/internal/rangeanalysis/SignAnalysisSpecific.qll

Synced between Java and C# through `identical-files.json`.

tamasvajk · 2020-09-21T14:18:36Z

Differences job after rebase, squash: https://jenkins.internal.semmle.com/job/Changes/job/CSharp-Differences/446/

hvitved · 2020-09-21T17:54:14Z

Differences job after rebase, squash: https://jenkins.internal.semmle.com/job/Changes/job/CSharp-Differences/446/

And here is one for Java as well: https://jenkins.internal.semmle.com/job/Changes/job/Java-Differences/929/

tamasvajk · 2020-09-22T09:35:13Z

Differences job after rebase, squash: https://jenkins.internal.semmle.com/job/Changes/job/CSharp-Differences/446/

And here is one for Java as well: https://jenkins.internal.semmle.com/job/Changes/job/Java-Differences/929/

I restarted the C# one: https://jenkins.internal.semmle.com/job/Changes/job/CSharp-Differences/448/
The Java one finished okay with no changes.

hvitved

Let's get this merged once the CSharp-Differences job completes (without regressions).

Thanks for you patience @tamasvajk ; this is excellent work.

java/ql/src/semmle/code/java/dataflow/internal/rangeanalysis/SignAnalysisCommon.qll

aschackmull · 2020-09-24T11:54:50Z

java/ql/src/semmle/code/java/dataflow/internal/rangeanalysis/SignAnalysisSpecific.qll

+    result = exprSign(e.(CastExpr).getExpr())
+  }
+
+  private Sign binaryOpLhsSign(Expr e) {


move to common

binaryOpLhsSign and binaryOpRhsSign have different implementation for C# and Java.

java/ql/src/semmle/code/java/dataflow/internal/rangeanalysis/SignAnalysisSpecific.qll

aschackmull · 2020-09-24T11:55:05Z

java/ql/src/semmle/code/java/dataflow/internal/rangeanalysis/SignAnalysisSpecific.qll

+  }
+
+  pragma[noinline]
+  private predicate binaryOpSigns(Expr e, Sign lhs, Sign rhs) {


move to common

This could be moved to the common bit, but the gain would be minimal. I think it's better to keep it together with binaryOpLhsSign and binaryOpRhsSign.

java/ql/src/semmle/code/java/dataflow/internal/rangeanalysis/SignAnalysisCommon.qll

tamasvajk · 2020-09-28T08:48:10Z

Thanks @aschackmull for your findings. I'll go through them during the week...

tamasvajk added the C# label Sep 2, 2020

tamasvajk commented Sep 2, 2020

View reviewed changes

tamasvajk requested a review from a team September 2, 2020 15:38

tamasvajk marked this pull request as ready for review September 2, 2020 15:39

tamasvajk commented Sep 2, 2020

View reviewed changes

csharp/ql/test/library-tests/dataflow/signanalysis/SignAnalysis.cs Outdated Show resolved Hide resolved

intrigus-lgtm reviewed Sep 2, 2020

View reviewed changes

csharp/ql/src/semmle/code/csharp/dataflow/SignAnalysis.qll Outdated Show resolved Hide resolved

tamasvajk marked this pull request as draft September 2, 2020 21:00

tamasvajk force-pushed the feature/sign-analysis branch from 381d906 to 1267db9 Compare September 3, 2020 08:54

calumgrant reviewed Sep 3, 2020

View reviewed changes

csharp/ql/src/semmle/code/csharp/dataflow/SignAnalysis.qll Outdated Show resolved Hide resolved

tamasvajk force-pushed the feature/sign-analysis branch 4 times, most recently from 61bf4b5 to 45db4d4 Compare September 3, 2020 14:11

tamasvajk commented Sep 3, 2020

View reviewed changes

csharp/ql/src/semmle/code/csharp/dataflow/SignAnalysis.qll Outdated Show resolved Hide resolved

tamasvajk force-pushed the feature/sign-analysis branch from d3270c8 to eedaab7 Compare September 4, 2020 08:16

tamasvajk requested a review from a team September 4, 2020 08:40

tamasvajk marked this pull request as ready for review September 4, 2020 08:48

tamasvajk requested a review from a team as a code owner September 4, 2020 08:48

tamasvajk commented Sep 8, 2020

View reviewed changes

csharp/ql/src/semmle/code/csharp/dataflow/internal/rangeanalysis/RangeUtils.qll Outdated Show resolved Hide resolved

tamasvajk added the Java label Sep 8, 2020

tamasvajk force-pushed the feature/sign-analysis branch from 6183a69 to d195bb7 Compare September 8, 2020 20:11

tamasvajk commented Sep 9, 2020

View reviewed changes

csharp/ql/src/semmle/code/csharp/dataflow/internal/rangeanalysis/RangeUtils.qll Outdated Show resolved Hide resolved

hvitved requested changes Sep 16, 2020

View reviewed changes

tamasvajk force-pushed the feature/sign-analysis branch 3 times, most recently from 4fbca9f to a021769 Compare September 17, 2020 09:49

hvitved requested changes Sep 17, 2020

View reviewed changes

tamasvajk force-pushed the feature/sign-analysis branch from 67b76a2 to 92f9bdc Compare September 17, 2020 14:46

tamasvajk commented Sep 17, 2020

View reviewed changes

csharp/ql/src/semmle/code/csharp/dataflow/internal/rangeanalysis/SignAnalysisSpecific.qll Outdated Show resolved Hide resolved

hvitved reviewed Sep 18, 2020

View reviewed changes

tamasvajk commented Sep 21, 2020

View reviewed changes

csharp/ql/test/library-tests/dataflow/signanalysis/MissingSign.expected Show resolved Hide resolved

csharp/ql/test/library-tests/dataflow/signanalysis/MissingSign.expected Show resolved Hide resolved

hvitved reviewed Sep 21, 2020

View reviewed changes

csharp/ql/src/semmle/code/csharp/dataflow/internal/rangeanalysis/SignAnalysisSpecific.qll Show resolved Hide resolved

tamasvajk force-pushed the feature/sign-analysis branch 2 times, most recently from 8a52f61 to 5bc8c46 Compare September 21, 2020 13:01

Add Java test file for sign analysis

441fbe3

tamasvajk force-pushed the feature/sign-analysis branch 2 times, most recently from 688cd92 to fdc9e01 Compare September 21, 2020 13:20

tamasvajk and others added 2 commits September 21, 2020 16:15

C#: Sign analysis

8bf4a42

Synced between Java and C# through `identical-files.json`.

C#: Add Guard::controlsBasicBlock() and simplify Guard::isEquality()

06dbec7

tamasvajk force-pushed the feature/sign-analysis branch from fdc9e01 to 06dbec7 Compare September 21, 2020 14:15

hvitved approved these changes Sep 22, 2020

View reviewed changes

tamasvajk merged commit 54c3574 into github:main Sep 22, 2020

aschackmull reviewed Sep 24, 2020

View reviewed changes

tamasvajk mentioned this pull request Sep 29, 2020

Sign analysis cleanup #4362

Merged

		not exists(SsaReadPositionBlock bb \| getAnExpression(bb) = e) and
		result = ssaDefSign(v)

Conversation

tamasvajk commented Sep 2, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

hvitved left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

tamasvajk commented Sep 17, 2020

Uh oh!

hvitved left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

tamasvajk commented Sep 17, 2020

Uh oh!

hvitved commented Sep 17, 2020

Uh oh!

tamasvajk commented Sep 17, 2020

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

tamasvajk commented Sep 18, 2020

Uh oh!

hvitved commented Sep 20, 2020

Uh oh!

Uh oh!

Uh oh!

Uh oh!

tamasvajk commented Sep 21, 2020

tamasvajk commented Sep 2, 2020 •

edited

Loading