I get the following error running a ruby program in Textmate, but it works fine running it from the command line. The program creates a lot of strings, and slightly reducing this just slightly makes the error go away.
I get this in Textmate ... 2 DATE 11 Jan 2001 at top level in 3 TIME 15 at line 56 1 RIN 13
but this on the command line (which is correct) ... 2 DATE 11 Jan 2001 3 TIME 15:56:49 1 RIN 13
Oddly, if I "puts s.reverse" in textmate, I get the right output too (reversed of course:). I also don't get the error if I print to a file, rather than printing to the Textmate output window (and I'm only printing out a very short segment of text from the program).
s = f.to_gedcom puts s #output window copy gives the error (before File.open). File.open('/tmp/xx.txt','w') { |fd| fd.print s } #file copy is correct puts s #output window copy gives the error (and also after File.open)
So is there a heap or stack size issue running ruby from textmate? Any other guesses at what is happening? It does it under Ruby 1.8.6 and Ruby 1.9.1.
On 08 Apr 2009, at 00:03, rburrowes wrote:
I get the following error running a ruby program in Textmate [...]
Looks like normal ruby error (stack dump) intermixed with regular output. Are you using fork()?
So is there a heap or stack size issue running ruby from textmate? Any other guesses at what is happening?
Maybe your program has a race condition and the different timings in the TM environment triggers it.
Can you include code which cause the error?
I could email you a copy if you wanted, but it is about 500K. It is online at Rubyforge, both through their web interfaces and via git. I haven't packaged the version that causes the problems, but could. I really need to sit down and isolate what causes the problem.
I have tried just creating lots of strings and tried doing this with a lot of recursion, both of which the code base does. I couldn't reproduce it. The code does produces a lot of objects, so I will try experimenting with this to see if I can reproduce it.
The code is not threaded, so it shouldn't have race conditions and the error printing the string is consistent. I can print the string, getting the error. Output the string to a file and not get the error, and print the same string again and get the error again. I can also print the string reversed, and I don't get the error.
The code reads a GEDCOM file (GEDCOM is a text base DB file), parsing the records into an object hierarchy. I get the error when I am testing the result of this code, by walking a small subset of the object hierarchy and producing GEDCOM text records as output to the Textmate output window.
The error is printed while outputting the string, but the Textmate window doesn't stop printing the rest of the same string. That seems odd to me.
The code is on Rubyforge in
http://gedcom.rubyforge.org/git?p=gedcom.git%3Ba=tree
Or via git using
git clone git://rubyforge.org/gedcom.git
Running test/gedcom, after removing the comments from around lib/gedcom/transmission.rb's summary method, will print the string, print the string to /tmp/xx.txt and print the string again. Both prints give the error, but the file copy is correct.
The string is produced calling f.to_gedcom, which is a method in lib/gedcom/gedcom_base.rb. This actually walks through all the other objects that were created from the classes in lib/gedcom/*.rb. It isn't that easy to follow.
The error will go away if @@tabs = false, rather than true. This stops the leading spaces being added to the lines in the string (in method GedcomBase.tabstop in gedcom_base.rb). The line " 3 Time ..." becoming "3 TIME ...". The error is there for Ruby 1.8.6 and Ruby 1.9.1.
Allan Odgaard-4 wrote:
On 08 Apr 2009, at 00:03, rburrowes wrote:
I get the following error running a ruby program in Textmate [...]
Looks like normal ruby error (stack dump) intermixed with regular output. Are you using fork()?
So is there a heap or stack size issue running ruby from textmate? Any other guesses at what is happening?
Maybe your program has a race condition and the different timings in the TM environment triggers it.
Can you include code which cause the error?
textmate mailing list textmate@lists.macromates.com http://lists.macromates.com/listinfo/textmate
-
Allan Odgaard -
rburrowes