Variables get corrupted with read from text file

[PaavoN]

Hi

I have a quite large project with tens or hundreds of variables defined. When creating the database,
I use 'Read from text file' to read a SQL script into a variable and do some replacements there
before serving it to MS SQL plugin.

Now when one of the SQL script files has grown over 1.6Mb and I read the file (using #READALL) into
a variable, few of my previously defined variables get corrupted. Is there a memory leak or what
could be causing this behaviour ? Or isn't that command meant to be used with such a big files ...

I solved the problem temporarily by splitting the file into a smaller pieces, but this problem will
come back some day and it might not be noticed until some weird problems when our customer
is installing our product.

We are using version 9 R2 sp2 and the problems has occurred on various platforms.

Paavo Nieminen

[mills]

What do you mean that your variables get corrupted?

[PaavoN]

I mean that the variable's value get lost. After reading a large script when
I write the variable to a log file I get e.g HTTPPORT =$HTTPPORT$

This happens for some of the variables but not all which makes me think of memory leak ...

PaavoN

[mills]

Is the 1.6MB the most recognizable threshold you've found?

[PaavoN]

Threshold was actually somewhere between 1300Kb and 1400Kb.

PaavoN

[mills]

I'll have to do some testing around this. It's on my list and should get some attention some time next week if I don't get to it tomorrow.