Re: [darcs-devel] a measurement of using 7z instead of gz to compress patches

2008-01-18 Thread zooko
On Jan 18, 2008, at 3:25 PM, John Meacham wrote: > Yes. I was thinking of zip. sorry. how does 'bzip2' compare actually? In my experience, 7z provides tighter compression and similar speed so I've stopped looking at bz2. But here you go: 81592 trunk-nocompress-individual-patches 48980 tru

Re: [darcs-devel] a measurement of using 7z instead of gz to compress patches

2008-01-18 Thread John Meacham
On Fri, Jan 18, 2008 at 01:03:03PM -0700, zooko wrote: > > Well, the tar version compresses much better than zip because it is > > streaming and doesn't have an index. since you tar then gzip, it > > doesn't > > compress each patch independently but rather compresses a single > > stream > > con

Re: [darcs-devel] a measurement of using 7z instead of gz to compress patches

2008-01-18 Thread zooko
> Well, the tar version compresses much better than zip because it is > streaming and doesn't have an index. since you tar then gzip, it > doesn't > compress each patch independently but rather compresses a single > stream > containing all the patches. since patches are extremely similar, this

Re: [darcs-devel] a measurement of using 7z instead of gz to compress patches

2008-01-17 Thread John Meacham
On Thu, Jan 17, 2008 at 10:02:40AM +0900, Stephen J. Turnbull wrote: > zooko writes: > > > In short, it probably isn't worth it, unless the new repository > > formats (which I haven't looked at) tend to bundle more patches > > together into one file for the compression tool to find redundan

Re: [darcs-devel] a measurement of using 7z instead of gz to compress patches

2008-01-16 Thread Stephen J. Turnbull
zooko writes: > On Jan 16, 2008, at 6:02 PM, Stephen J. Turnbull wrote: > > > Better to use zip format than tar, as zip has an index (that's why > > it's used by jar files and Python eggs, IIRC). Doesn't 7zip have a > > full archive mode like zip? > > Yes, it does. It didn't occur to me

Re: [darcs-devel] a measurement of using 7z instead of gz to compress patches

2008-01-16 Thread zooko
On Jan 16, 2008, at 6:02 PM, Stephen J. Turnbull wrote: > Better to use zip format than tar, as zip has an index (that's why > it's used by jar files and Python eggs, IIRC). Doesn't 7zip have a > full archive mode like zip? Yes, it does. It didn't occur to me to use it for these measurements,

[darcs-devel] a measurement of using 7z instead of gz to compress patches

2008-01-16 Thread Stephen J. Turnbull
zooko writes: > In short, it probably isn't worth it, unless the new repository > formats (which I haven't looked at) tend to bundle more patches > together into one file for the compression tool to find redundancy in: Better to use zip format than tar, as zip has an index (that's why it's

[darcs-devel] a measurement of using 7z instead of gz to compress patches

2008-01-16 Thread zooko
In short, it probably isn't worth it, unless the new repository formats (which I haven't looked at) tend to bundle more patches together into one file for the compression tool to find redundancy in: Current: 37188 trunk-get/_darcs Each patch un-gzipped and 7zipped: 32684 trunk-7z/_darcs