This is the mail archive of the gdb-patches@sources.redhat.com mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: RFA/types: Clean up use of field bitsize


On Wed, Nov 06, 2002 at 03:41:10PM -0500, Andrew Cagney wrote:

>Does anyone have a comment on this patch?  If not, I'll commit it in a
>couple of days, after I'm added to the global write list.
>
>(The type code has no specific maintainer, the debug reader and
>language parts I consider obvious, and the patch is over a month old
>now.)


I'm mainly wondering if we're that desperate for memory space.

I thought a data structure was added to GDB so that it could spot duplicate type info and, hence, keep its memory size down.

If so, I don't see it.  The debug readers will create a new copy when
they hit a new definition.
Sigh, looks depressingly like a proposal that fell flat :-(
There are bcache's for macro and psymbol stuff but not types.

Besides, wasting memory is still bad.  And that's not the reason I did
it, anyway:
True, the real problem is (true?) the type duplication - gdb is wasting memory by duplicating type information - fixing that eliminates the problem removing the need for a micro optomization?

Can't find the discussion :-(

>>The goal is to allow more kinds of fields to be marked artificial -
>>particularly data members.  After this patch I'll submit the followup to
>>mark DW_AT_artificial members as artificial types.

... in other words, moving artificial out of loc without wasting an
additional 32 bits.
I'm definitly not questioning this.

Andrew



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]