JRuby

Incorrect behavior of RubyMatchData updateCharOffset() and Strings with special characters

Details

  • Type: Bug Bug
  • Status: Closed Closed
  • Priority: Major Major
  • Resolution: Fixed
  • Affects Version/s: JRuby 1.6.4
  • Fix Version/s: JRuby 1.6.6
  • Component/s: Core Classes/Modules
  • Labels:
  • Environment:
    jruby 1.6.4 (ruby-1.9.2-p136) (2011-08-23 17ea768) (OpenJDK 64-Bit Server VM 1.7.0-internal) [darwin-x86_64-java]
  • Number of attachments :
    0

Description

In Ruby 1.9 mode, an ArrayIndexOutOfBoundsException is thrown when special characters are used in Rails ERB templates.

This bug relates to a similar resolved issue, JRUBY-5924. This bug report also uses the same gist to demonstrate the issue:

https://gist.github.com/1080407

mbp-scott:~ scott$ ruby -v
jruby 1.6.4 (ruby-1.9.2-p136) (2011-08-23 17ea768) (OpenJDK 64-Bit Server VM 1.7.0-internal) [darwin-x86_64-java]

mbp-scott:~ scott$ ruby test.rb 
RubyMatchData.java:137:in `updateCharOffset': java.lang.ArrayIndexOutOfBoundsException: 6
	from RubyMatchData.java:479:in `begin19'
	from RubyMatchData$i$1$0$begin19.gen:65535:in `call'
	from CachingCallSite.java:312:in `cacheAndCall'
	from CachingCallSite.java:169:in `call'
	from test.rb:12:in `__file__'
	from test.rb:-1:in `load'
	from Ruby.java:679:in `runScript'
	from Ruby.java:672:in `runScript'
	from Ruby.java:579:in `runNormally'
	from Ruby.java:428:in `runFromMain'
	from Main.java:278:in `doRunFromMain'
	from Main.java:198:in `internalRun'
	from Main.java:164:in `run'
	from Main.java:148:in `run'
	from Main.java:128:in `main'

mbp-scott:~ scott$ export JRUBY_OPTS=--1.8
mbp-scott:~ scott$ ruby test.rb 
21
34

Activity

Ascending order - Click to sort in descending order
Hide
Permalink
Wei Jen Lu added a comment -

I had same problem, when erb included chinese characters.

Show
Wei Jen Lu added a comment - I had same problem, when erb included chinese characters.
Hide
Permalink
Tobias H. Michaelsen added a comment -

There seems to be some problems with ERB and encoding in general with JRuby.
In this tiny test, the ERB template seems to be treated as ASCII https://gist.github.com/1287425

Show
Tobias H. Michaelsen added a comment - There seems to be some problems with ERB and encoding in general with JRuby. In this tiny test, the ERB template seems to be treated as ASCII https://gist.github.com/1287425
Hide
Permalink
Craig Waterman added a comment -

I can confirm this bug still exists in JRuby 1.6.5; any unicode character in an ERB template throws the error: https://gist.github.com/1325372

Show
Craig Waterman added a comment - I can confirm this bug still exists in JRuby 1.6.5; any unicode character in an ERB template throws the error: https://gist.github.com/1325372
Hide
Permalink
Jarbas Brito added a comment -

I'm with the same problem because my erb template is in portuguese, so today I forked jruby project from github and will go to try resolve the problem. I hope to get and send "pull request".

Show
Jarbas Brito added a comment - I'm with the same problem because my erb template is in portuguese, so today I forked jruby project from github and will go to try resolve the problem. I hope to get and send "pull request".
Hide
Permalink
Brandon Hays added a comment -

I'm having the same issue. Here's a simple ERB example:

https://gist.github.com/1355431

Show
Brandon Hays added a comment - I'm having the same issue. Here's a simple ERB example: https://gist.github.com/1355431
Hide
Permalink
Carlo Scarioni added a comment -

This also happens when using cucumber with JRuby and there are special characters in the feature file

Show
Carlo Scarioni added a comment - This also happens when using cucumber with JRuby and there are special characters in the feature file
Hide
Permalink
Carlo Scarioni added a comment -

Hi,
Looking at the source code for a while, I managed to fix/workaround this for my particular case with cucumber.
But I'm not sure if I have broken anything. I just changed one line in the class org.jruby.RubyMatchData. In the method updateCharOffset there is an "if" that looks like this:

if (regs != null) {
for (;regs.beg[numPairs] >= 0; numPairs++) {}
}

I don't know exactly what would it mean for regs.beg[numPairs] to be < 0, but in my case it never is. the variable "regs" contains the matching groups from the MatchData RegEx.

By replacing that "if" with the following:

if (regs != null)

{ numPairs=numRegs; }

numRegs is how many groups there are in the match data.

In my case it seems to work with the cucumber with special characters,

My problem now is that I can't see to run the tests from the project or to build it properly, from maven and/or ant.

both ant test, and mvn test are failing, I will update on any progress I can make

Show
Carlo Scarioni added a comment - Hi, Looking at the source code for a while, I managed to fix/workaround this for my particular case with cucumber. But I'm not sure if I have broken anything. I just changed one line in the class org.jruby.RubyMatchData. In the method updateCharOffset there is an "if" that looks like this: if (regs != null) { for (;regs.beg [numPairs] >= 0; numPairs++) {} } I don't know exactly what would it mean for regs.beg [numPairs] to be < 0, but in my case it never is. the variable "regs" contains the matching groups from the MatchData RegEx. By replacing that "if" with the following: if (regs != null) { numPairs=numRegs; } numRegs is how many groups there are in the match data. In my case it seems to work with the cucumber with special characters, My problem now is that I can't see to run the tests from the project or to build it properly, from maven and/or ant. both ant test, and mvn test are failing, I will update on any progress I can make
Hide
Permalink
Carlo Scarioni added a comment -

Yesterday I made a mvn install without a problem, and this fix is working for me, but as I said, I don't know in what cases it might not work as there is apparently no cases in tests for this (As I said I did mvn install without a problem), so for the time being I will deploy this to my company's nexus (as a patched version) and use it from there,
As I don't know if the fix broke anything else (as if not all edge cases are tested, or simply the fix didn't broke anything) I won't make at this moment a pull request for this fix. I'll try it in my environment for a while, although I'm only using cucumber at the moment (not ERB).

If I'm missing in the way to run a full suite of tests for JRuby please let me know.

Show
Carlo Scarioni added a comment - Yesterday I made a mvn install without a problem, and this fix is working for me, but as I said, I don't know in what cases it might not work as there is apparently no cases in tests for this (As I said I did mvn install without a problem), so for the time being I will deploy this to my company's nexus (as a patched version) and use it from there, As I don't know if the fix broke anything else (as if not all edge cases are tested, or simply the fix didn't broke anything) I won't make at this moment a pull request for this fix. I'll try it in my environment for a while, although I'm only using cucumber at the moment (not ERB). If I'm missing in the way to run a full suite of tests for JRuby please let me know.
Hide
Permalink
Brian Johnson added a comment -

It looks like this error happens in 1.9 mode with erubis. See: https://gist.github.com/1355431

That's probably a show stopper for using jruby in 1.9 mode with erubis with any unicode characters. Maybe this is a higher priority bug?

Show
Brian Johnson added a comment - It looks like this error happens in 1.9 mode with erubis. See: https://gist.github.com/1355431 That's probably a show stopper for using jruby in 1.9 mode with erubis with any unicode characters. Maybe this is a higher priority bug?
Hide
Permalink
Thomas E Enebo added a comment -

Fixed on master in b4a5469 and on jruby-1_6 4c01cf4

Show
Thomas E Enebo added a comment - Fixed on master in b4a5469 and on jruby-1_6 4c01cf4

People

  • Assignee:
    Thomas E Enebo
    Reporter:
    Scott Lowe
Vote (9)
Watch (7)

Dates

  • Created:
    Updated:
    Resolved: