IRC channel logs
2014-05-16.log
back to list of logs
<civodul>for the purposes of line-buffering, shouldn't \\r be considered a line ending? *civodul prepares a patch <taylanub>civodul: IIRC only legacy MacOS used sole \\r for line-endings. <taylanub>but I guess it doesn't have any other semantics attached so it sure wouldn't hurt to support that <civodul>on Unices \\r goes to the beginning of the line <taylanub>hrm, terminal emulators still support that, it might perhaps cause problems when one does (read-line) and only gets the part up to \\r from a line that really meant to control a terminal? <nalaginrut>the current read-line doesn't eat \\r, which caused a bug in my multiform-data parser, then I add \\r to trimming list <taylanub>nalaginrut: WDYM "doesn't eat"? on 2.0.11: <nalaginrut>but anyway I trimmed it, because the string has to be trimmed for other chars later <taylanub>I think both of these behaviors are correct, and might even be explicitly desired or relied upon when one deals with a terminal. I'm not sure how common it is for a sole '\\r' to be desired to delimit a line <nalaginrut>this bug caused my uploaded filename has a redundant #\\return <civodul>taylanub: that's because it's coarse grain: as soon as there's a line ending in the string, it flushes the port, including what's after the line ending <civodul>also, line buffering is for file ports only (surprisingly) <taylanub>civodul: what I wanted to demonstrate is that it doesn't stop at the \\r, it includes it and what follows, resulting in "foo" being overwritten by "bar" <civodul>but there's no real line buffering for non-file ports anyway <civodul>so string ports cannot be used to test for that <civodul>taylanub: i've posted a patch, comments welcome! <taylanub>civodul: I kind of confused buffering with actual line-delimitation. should be fine I guess <madsy>Woo.. can hardly wait for my Xtion PRO LIVE <add^_>I have a question, I just compiled guile from the 2.0.11 branch on git, but when I start guile it says it's version 2.0.9 <madsy>add^_: Probably an oversight <add^_>nothing to bother worrying about them :-) <daviid>add^_: you need to make _version <add^_>but why would I need that? It's the first time I install guile on this system <add^_>I haven't needed that the first times the other times <daviid>did you remove the other version? <daviid>that does not make sence you probably stil run another version imo <add^_>It's a newly installed system <add^_>Unless it's preinstalled with it <daviid>well you're not new to this chanel, so I assumed you were using guile already <add^_>I did say that it was the first time on *this* system <add^_>I'm still trying what you said :-) <add^_>Need to redo something.. You'll have to wait for the result a while longer :-/ <lloda>add^: running ./autogen.sh before make & make install will set the version tag. <add^_>I've been doing that from the start <add^_>Now it's correct, thanks guys :-) <add^_>It might have been that I was in the wrong branch, should have been in origin/stable2.0 but I was in v2.0.11 <add^_>So I changed that, and ran the make _version besides the normal stuff <daviid>ahh! yes, now it's a nice friday :) <add^_>Finally, dsmith-work has approved! <add^_>Ah, feels good to be back in scheme/lisp world. <3 <add^_>I was in Haskell world for a short visit, was nice, but I consider the scheme/lisp world to be home. <add^_>And like people say in Sweden "Borta bra men hemma bäst" which translates to "Gone is good, but home is best" roughly.