Re: strings

2001-07-25 Thread Glenn F. Maynard
On Wed, Jul 25, 2001 at 09:51:57PM +0200, Nicolas Noble wrote: > > Alright. I'll revisit this once the current queue stuff is settled and tested. > > If you need some dummy code like this, I can handle this. Well, I've > already started a String class for myself. So if you want it, just yell, >

Re: strings

2001-07-25 Thread Nicolas Noble
> On Wed, Jul 25, 2001 at 04:25:31PM +0400, Alexander V. Lukyanov wrote: > > A simple string class can be useful. But passing it by value can be costly. > > I think much of c++ bloat goes from that. > > Well, most of the time it's passed by reference (const string &); more often, the > cost is te

Re: strings

2001-07-25 Thread Glenn F. Maynard
On Wed, Jul 25, 2001 at 04:25:31PM +0400, Alexander V. Lukyanov wrote: > A simple string class can be useful. But passing it by value can be costly. > I think much of c++ bloat goes from that. Well, most of the time it's passed by reference (const string &); more often, the cost is temporaries du

Re: strings

2001-07-25 Thread Dmitry A. Fedorov
On Wed, 25 Jul 2001, Alexander V. Lukyanov wrote: > On Tue, Jul 24, 2001 at 06:30:00PM -0400, Glenn F. Maynard wrote: > > Why not introduce some sort of string class into lftp? It's understandable > A simple string class can be useful. But passing it by value can be costly. > I think much of c+

Re: queue feeder

2001-07-25 Thread Alexander V. Lukyanov
On Tue, Jul 24, 2001 at 03:42:29PM -0400, Glenn F. Maynard wrote: > Anyhow, this isn't really an issue worth another day of deliberation. :) > Here's a better idea: change is_queue to has_queue, of type QueueFeeder. > If there's a queue, it points to it; if not, it's NULL. No extra code, > no ass

Re: strings

2001-07-25 Thread Alexander V. Lukyanov
On Tue, Jul 24, 2001 at 06:30:00PM -0400, Glenn F. Maynard wrote: > Why not introduce some sort of string class into lftp? It's understandable > to avoid system STL and C++ strings (they're taking absurdly long to actually > become implemented well outside of g++), but it'd probably be worthwhile

Re: eep! memleak

2001-07-25 Thread Glenn F. Maynard
On Wed, Jul 25, 2001 at 01:12:27PM +0400, Alexander V. Lukyanov wrote: > On Tue, Jul 24, 2001 at 08:29:48PM -0400, Glenn F. Maynard wrote: +0400. -0400. Exact opposite timezones ... which explains why you're always replying right as I'm going to bed. :) > > Try this: create a file, containing a

Re: eep! memleak

2001-07-25 Thread Alexander V. Lukyanov
On Tue, Jul 24, 2001 at 08:29:48PM -0400, Glenn F. Maynard wrote: > Try this: create a file, containing a simple command (ie. just an echo); and > issue "repeat 0 que source filename". It'll leak like crazy. I think it queues commands faster than executes them. So it is not a leak. -- Alexan

Re: eep! memleak

2001-07-25 Thread Alexander V. Lukyanov
On Tue, Jul 24, 2001 at 05:39:35PM -0400, Glenn F. Maynard wrote: > glenn 8389 0.0 51.2 84056 6 pts/30 SJul16 3:40 lftp > > That's from: > repeat "find ; cache f" > repeated a few hundred times. This is Debian/unstable's 2.3.11, not CVS. It > appears to be trivially reproducable,