FWIW, I just tried Racket 6.1 installing PLaneT "neil/csv:1:7", and two
observations:
* I don't immediately know why, but my Racket 6.1 (command-line
"racket", and DrRacket) doesn't build the documentation when installing
the PLaneT package. I can't look into this right now.
* The performance looks fine, for not building the documentation. (This
is on a 6+ year-old 32-bit x86 GNU/Linux laptop with slow spinning hard
drive, running a Racket I built from source.)
I haven't played with Racket 6.x much, since the main project I'm
working on right now is still using Racket 5.3.4 (with Ryan's essential
PostgreSQL SSL patch).
(Aside, so I don't give the wrong impression... That project doesn't
have anything against 6.x. When 6.0 came out, preliminary tests with it
were fine, but we decided to take a wait&see, since the Racket
architecture was being shaken up, and the timing wasn't great for us to
evaluate stability and security. Since then, I haven't heard of any big
issues with 6.x, and I ran a development tree OK under 6.1 incidentally
recently, but there hasn't yet been a need or good opportunity to switch
production to 6.x. For my personal and open source work, I use the
latest released version of Racket.)
If anyone finds a problem with my open source packages, please CC my
personal email address, so that I'm sure to see your message.
[~] rm -rf ~/.racket/6.1 ~/.racket/planet/300/6.1
~/.racket/planet/300/packages/
[~] /usr/local/racket-6.1/bin/racket --version
Welcome to Racket v6.1.
[~] time /usr/local/racket-6.1/bin/racket -e '(begin (require (planet
neil/csv:1:=7)) (csv->sxml "a,b"))'
'(*TOP* (row (col-0 "a") (col-1 "b")))
real 0m21.526s
user 0m18.365s
sys 0m2.744s
[~] time /usr/local/racket-6.1/bin/racket -e '(begin (require (planet
neil/csv:1:=7)) (csv->sxml "a,b"))'
'(*TOP* (row (col-0 "a") (col-1 "b")))
real 0m0.660s
user 0m0.568s
sys 0m0.088s
[~]
Neil V.
George Neuner wrote at 08/25/2014 09:42 PM:
Hi all,
I'm running Racket under 64-bit Win7 Home on an i7 with 12GB of RAM.
Just upgraded to 6.0.1 from 5.3.4. I know it's a big leap and 6.1 is
out already, but I tried 6.1 and quickly experienced several of those
non-reproducible crashes I've been reading about in the lists.
6.0.1 is behaving itself, but installing new packages from PLaneT or
migrating packages from the older version is taking HOURS ...
virtually all of it spent processing .scrbl files.
I have both 32 and 64 bit versions: doesn't matter which version I
use, how much memory I give to Racket (in racket-prefs.rktd), or
whether I use raco from the command line or the GUI-based package
manager.
Example: trying to install *just* Neil's CSV:2:0 took nearly an hour
(yes, it has several dependencies). With 64-bit Racket configured for
3GB, memory use spiked over 1.9GB according to Process Explorer. And
that was the *fastest* installation ... reducing the memory limit
below 2GB just slows it down.
[I didn't try running the 32-bit version with more than 1.5GB, but it
behaved the same way ... reducing the memory limit just slowed it
down.]
For giggles I tried migrating CSV:1:7 (which has a single dependency
not shared with 2:0) from 5.3.4. That still took almost 16 minutes,
and again nearly all of it processing some .scrbl file. For more
giggles, I nuked the PLaneT cache so as to start over: same 16 minutes
for fresh install of 1:7.
For comparison, 5.3.4 installs CSV:1:7 fresh in under 3 minutes.
??? Is anyone else running 6.x on Windows? I can't figure out what,
if anything, I'm doing wrong.
Subjectively, 6.0.1 feels sluggish compared to 5.3.4 (with same memory
limits), but I haven't gotten to the point where I can objectively
test much because a lot of my programs depend on external packages.
Based on the few I have tried to install/migrate so far, Halloween may
be past before I am able to do any real work with 6.0.1.
George
____________________
Racket Users list:
http://lists.racket-lang.org/users
____________________
Racket Users list:
http://lists.racket-lang.org/users
