Re: [EXTERNAL] Source parsing, for outline, performance

Hey Steve,



I was afraid you would ask...  This is the copy book that defines all of our constants, common data structures, and general prototypes.  It is currently 3,280 lines long.





Inside this copy book:

7 other copy books:  these are about 2,000 lines combined

Lots of constants

Several C prototypes

Many data structures, mostly templates, some with extname(), so they would reference external files
Many of the fields are defined like definitions, so they would reference external files



Otherwise, nothing stands out as too out of the ordinary or complex, it is just big.









---- On Fri, 25 Jul 2025 12:26:56 -0700 Steve Ferrell via WDSCI-L <wdsci-l@xxxxxxxxxxxxxxxxxx> wrote ---



Hi David,

Define very large copy books? How many lines?

And do they all reference other copy members?

And do they include DS or DS with arrays? Multi level DS?

I agree caching those would help.

Steve
Fortra

-----Original Message-----
From: WDSCI-L < mailto:wdsci-l-bounces@xxxxxxxxxxxxxxxxxx > On Behalf Of David Wright
Sent: Friday, July 25, 2025 2:13 PM
To: Rational Developer for IBM i < mailto:wdsci-l@xxxxxxxxxxxxxxxxxx >
Subject: Re: [WDSCI-L] [EXTERNAL] Source parsing, for outline, performance

External: Pause and review the sender's email address, any URLs before clicking links, opening attachments, or following requests. When in doubt, contact the Service Desk.


I have mostly seen the extreme slowness in programs with very large copy books. These copy books rarely change, so if copy books could be parsed/cached locally I think it might dramatically improve performance. A copy book should only be reprocessed if it changes. A copy book cache in our case would provide system-wide benefit because every program has the same copy book I am thinking of...









---- On Fri, 25 Jul 2025 12:05:40 -0700 Greg Wilburn < mailto:gwilburn@xxxxxxxxxxxxxxxxxxxxxxx > wrote ---



Yes, I have... mine sometimes never finishes parsing. I experienced this in 9.6 as well.

I have also noticed that the feature that allows you to hover over a procedure to see the "comments" from the prototype source doesn't always work. I end up closing and restarting RDi.

Like others, I chalked this up to Java as well.

-----Original Message-----
From: WDSCI-L < mailto: mailto:wdsci-l-bounces@xxxxxxxxxxxxxxxxxx > On Behalf Of David Gibbs via WDSCI-L
Sent: Friday, July 25, 2025 11:12 AM
To: Rational Developer for IBM i / Websphere Development Studio Client for System i & iSeries < mailto: mailto:wdsci-l@xxxxxxxxxxxxxxxxxx >
Cc: David Gibbs < mailto: mailto:david@xxxxxxxxxxxx >
Subject: [EXTERNAL] [WDSCI-L] Source parsing, for outline, performance

Folks:

Does anyone know why RDi 9.8 takes such a long time to parse the source to update the outline?

Frequently I've observed the 'live parsing' takes multiple minutes ...
most of the time the progress is shown as 0%.

Granted, the source I'm editing has a lot of copy members, but (according to the log) the parser is using a cached copy of the copy member.

Is anyone else seeing this behavior?

Thanks!

David