Performance issue with code page feature

[RossEvans]

I have been using Text Pipe Pro to convert some files from their native IBM code page (37 IBM EBCDIC US/Canada) to my standard Windows code page (1252 ANSI Latin I).

Prior to the introduction of your new code page feature, which I installed with version 7.4.2, I routinely have been using a custom character map that I constructed myself. I was glad to see the new code page feature in v. 7.4; processes using it would be more maintainable since they would no longer rely on my custom map.

However, I hit a hard stop with performance problems with this feature.

Using my old, custom filter to convert a 695 Mb file on my desktop (1.8 Ghz, 2 Gb RAM), Text Pipe Pro runs fine and completes in 58 seconds.

But trying to use Text Pipe Pro's code page filter, the program churns away for 17.5 minutes, having processed only about 250 Mb of the file, then aborts with an "Out of Memory" error.

[DataMystic Support]

Hi Ross,

You're right, the new filter doesn't yet take advantage of TextPipe's architecture for these conversions.

We're working on this for the next version.