Eclipse Community Forums: TMF (Xtext) » inconsistency of case sensitivity in xtext suffix

Home » Modeling » TMF (Xtext) » inconsistency of case sensitivity in xtext suffix(xtext is only sometimes sensitive to case of suffix)

Show: Today's Messages :: Show Polls :: Message Navigator

inconsistency of case sensitivity in xtext suffix [message #1787405]

Thu, 24 May 2018 20:47

Jay Norwood

Messages: 155
Registered: July 2009

Senior Member

I pulled down the tracing_codegen example from github and substituted 2.14 xtext on the oxygen DSL eclipse, on Windows 10.

Things were looking good.

I created a file with .MyDsl suffix and eclipse recognized it and converted my project to an xtext project.

I edited the model, using the example dsl code from github, and the editor accepted all that, and the outline view updated.

However, the .c and .h suffix files weren't generated, and no clue given why.

After debugging, I found that the code that does the generation was more picky about the suffix, and was accepting only "mydsl", all lower case, which is in fact the suffix in the genmodel file.

So, I don't know, but seems like it should consistently accept or reject if the case doesn't match.

I believe it came down to registry.getExtensionToFactoryMap.containsKey(name.substring(index+1)), which apparently doesn't ignore case (in the UriValidator isPossiblyManaged()

Report message to a moderator