- removed comment
parameter file parser aborts when findeing first error
The attached parfile contains multiple errors (one per line).
However the parameter file parser only reports the first one, then stops. This makes verifying parfiles for correctness hard. It might be good to defer aborting until the end of the file or until a larger number of parsing errors were encountered.
Keyword:
Comments (11)
-
reporter -
- removed comment
I don't know a generic way to recover from a syntax error. My experience is that when compilers try to do this, usually only the first error message is meaningful. I can make specific syntax errors recoverable, however. I could make piraha recover from the specific errors in this file.
-
- removed comment
Take a look at the multiple_syntax_errors branch of the flesh and see if it does what you want.
-
- changed status to open
- assigned issue to
- removed comment
Can I put this in review?
-
reporter - removed comment
You can certainly put it up for review. The review report may or may not be "please apply" in the end (by whoever ends up reviewing it which does not have to be me :-)). For review though please either link the pull request here or provide a diff to look at.
-
- changed status to open
- removed comment
-
reporter - changed status to open
- removed comment
I added comments on the code to a pull request for the branch that you can find here: https://bitbucket.org/cactuscode/cactus/pull-requests/37/multiple-syntax-errors-enabled-through-a/diff
Testing functionality I find that it still aborts after the first error, even for the fail.par file attached in this ticket. In fact I have been unable to make it produce more than one syntax error. Would you be able to provide a parameter file where the change in the code makes a difference, please?
-
- changed status to open
- removed comment
Roland, I replied to your comments and issues and pushed a new revision to the branch.
-
- changed status to open
- removed comment
-
- changed status to resolved
- removed comment
-
reporter - edited description
- changed status to closed
- Log in to comment
This still applies even with the piraha based parser. It is particularly annoying on clusters where one cannot run
cactus_sim -S fail.par
on the login nodes (eg Cray machines).