Repository navigation
StringIndexOutOfBoundsException exception from org.spdx.utility.compare.TemplateRegexMatcher.findTemplateWithinText() #242
Description
When calling org.spdx.utility.compare.LicenseCompareHelper.matchingStandardLicenseIdsWithinText(), I'm getting a StringIndexOutOfBoundsException thrown from within org.spdx.utility.compare.TemplateRegexMatcher.findTemplateWithinText() because the start index is larger than the end index. At this time I'm not sure exactly what the values of the arguments are to that method that are causing the issue, but I'll start trying to track that down next. This used to work with SPDX License List v3.23, so it's possible this is related to something that changed in SPDX License List v3.24.0.
The full exception stack trace is as follows, and note that the formatting may look unfamiliar as Clojure has its own way of printing exception stack traces:
{:clojure.main/message
"Execution error (StringIndexOutOfBoundsException) at jdk.internal.util.Preconditions$1/apply (Preconditions.java:55).\nRange [18134, 17912) out of bounds for length 29880\n",
:clojure.main/triage
{:clojure.error/class java.lang.StringIndexOutOfBoundsException,
:clojure.error/line 55,
:clojure.error/cause
"Range [18134, 17912) out of bounds for length 29880",
:clojure.error/symbol jdk.internal.util.Preconditions$1/apply,
:clojure.error/source "Preconditions.java",
:clojure.error/phase :execution},
:clojure.main/trace
{:via
[{:type java.util.concurrent.ExecutionException,
:message "java.util.concurrent.ExecutionException: java.lang.StringIndexOutOfBoundsException: Range [18134, 17912) out of bounds for length 29880",
:at [java.util.concurrent.FutureTask report "FutureTask.java" 122]}
{:type java.util.concurrent.ExecutionException,
:message "java.lang.StringIndexOutOfBoundsException: Range [18134, 17912) out of bounds for length 29880",
:at [java.util.concurrent.FutureTask report "FutureTask.java" 122]}
{:type java.lang.StringIndexOutOfBoundsException,
:message "Range [18134, 17912) out of bounds for length 29880",
:at [jdk.internal.util.Preconditions$1 apply "Preconditions.java" 55]}],
:trace
[[jdk.internal.util.Preconditions$1 apply "Preconditions.java" 55]
[jdk.internal.util.Preconditions$1 apply "Preconditions.java" 52]
[jdk.internal.util.Preconditions$4 apply "Preconditions.java" 213]
[jdk.internal.util.Preconditions$4 apply "Preconditions.java" 210]
[jdk.internal.util.Preconditions outOfBounds "Preconditions.java" 98]
[jdk.internal.util.Preconditions outOfBoundsCheckFromToIndex "Preconditions.java" 112]
[jdk.internal.util.Preconditions checkFromToIndex "Preconditions.java" 349]
[java.lang.String checkBoundsBeginEnd "String.java" 4865]
[java.lang.String substring "String.java" 2834]
[org.spdx.utility.compare.TemplateRegexMatcher findTemplateWithinText "TemplateRegexMatcher.java" 322]
[org.spdx.utility.compare.TemplateRegexMatcher isTemplateMatchWithinText "TemplateRegexMatcher.java" 278]
[org.spdx.utility.compare.LicenseCompareHelper isStandardLicenseWithinText "LicenseCompareHelper.java" 890]
[org.spdx.utility.compare.LicenseCompareHelper matchingStandardLicenseIdsWithinText "LicenseCompareHelper.java" 959]
[spdx.matching$licenses_within_text invokeStatic "matching.clj" 84]
[spdx.matching$licenses_within_text invoke "matching.clj" 71]
[lice_comb.impl.matching$eval23925$fn__23926$fn__23927 invoke "matching.clj" 136]
[embroidery.api$pmap_STAR_$fn$reify__22819 call "vthreads.clj" 36]
[java.util.concurrent.FutureTask run "FutureTask.java" 317]
[java.lang.VirtualThread run "VirtualThread.java" 309]],
:cause "Range [18134, 17912) out of bounds for length 29880"}}
Execution error (StringIndexOutOfBoundsException) at jdk.internal.util.Preconditions$1/apply (Preconditions.java:55).
Range [18134, 17912) out of bounds for length 29880
This was reproduced with Java v21.0.3+9, Spdx-Java-Library v1.1.11, and SPDX License List v3.24.0.
Activity
goneall commentedon Jun 5, 2024
I did a quick look at the code and the out of bound is likely caused by an a substring range where the end is before the start.
I can see a corner case where this can happen if the pattern searching for the end of the license happens to also match the pattern before the start of the license.
We could add a check for this and look for a following match if this occurs.
The code to be updated is in the findTemplateInText method.
goneall commentedon Feb 26, 2025
@pmonks - Do you have some text that reproduces this?
Check for end index being greater then start in TemplateRegexMatcher
pmonks commentedon Mar 7, 2025
Hmmmm I can't find a test case for it in my usual set, and it's long enough ago now that I don't recall what it was that triggered it. I have a very vague memory that perhaps the problem had nothing to do with the input text though, and instead occurred when a license list v3.24.0 template was being "compiled" for the purposes of matching, that had previously worked in license list v3.23.0. Perhaps looking at what changed between those two versions of the license list would narrow down the potential culprits?
goneall commentedon Mar 7, 2025
@pmonks - no worries - if you could review #299 - I'm about 90% sure this will fix the issue.
Check for end index being greater then start in TemplateRegexMatcher