ddfreyne changed the topic of #nanoc to: 3.8.0 (May 4th) - 4.0.0rc2 (July 11th) | web http://nanoc.ws/ | repo http://bit.ly/XE6e3G | issues http://bit.ly/VfXaSV | forum http://ho.io/n-discuss | irclog http://irclog.whitequark.org/nanoc
TobiasFar has quit [Excess Flood]
TobiasFar has joined #nanoc
countsickness has quit [Ping timeout: 252 seconds]
countsickness has joined #nanoc
suy has joined #nanoc
pkunc has joined #nanoc
yaloki has joined #nanoc
<yaloki> howdy
<yaloki> writing a filter that renders PNGs from plain text files using plantuml
<yaloki> a bit at loss because the input file is text but the output is binary and nanoc assumes that a filter keeps the "type"
<yaloki> what works: just leave .pu (text) files as binary and produce the PNG, but then I can't use other text filters to preprocess the input file (in this case, I'd like to also use :erb)
<yaloki> an alternative would be to create new Item objects programatically for the resulting PNG files, and route the source .pu files into oblivion
yaloki has quit [Quit: leaving]
<guardian> hello
<guardian> just read the ML, what's wrong about system() or ``? how is popen3 (much) better?
VitamineD has quit [Quit: mi dormas]
<suy> guardian: I assume because that reads all stdout/err into memory
pkunc has quit [Read error: Connection reset by peer]
pkunc has joined #nanoc
pkunc has left #nanoc [#nanoc]
cDlm has joined #nanoc
<guardian> suy: I don't think I'm following you
<suy> guardian: neither do I without more context, I think. :)
<guardian> I'll read about system vs popen3 :)
<suy> guardian: I think that `cat large-file` will keep reading the stdout of the launched process, potentially blocking or exhausting memory. That's what I used to read.
cDlm has quit [Ping timeout: 250 seconds]
pkunc has joined #nanoc
pkunc has quit [Client Quit]
ndim has quit [Ping timeout: 244 seconds]
ndim has joined #nanoc
cDlm has joined #nanoc
pkunc has joined #nanoc
pkunc has quit [Quit: Leaving.]