Perl 6 - the future is here, just unevenly distributed

IRC log for #rosettacode, 2011-01-25

| Channels | #rosettacode index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
01:02 BenBE2 Anyone using ProFTPd? http://benbe.home.omorphia.de:43815/geshi-misc/profiling/run_tests.php?show=proftpd
01:02 fedaykin "GeSHi v.1.0.8.9 Test page" http://rldn.net/E5I
01:43 Mathnerd314 left #rosettacode
01:44 Mathnerd314_ joined #rosettacode
01:48 Mathnerd314_ left #rosettacode
03:51 Dgamey[RCo] joined #rosettacode
03:52 Dgamey[RCo] Has anyone noticed that some pages aren't being serverd.  They get internal server 500 errors?
04:07 SoniaKeys joined #rosettacode
04:12 Dgamey[RCo] left #rosettacode
04:13 Dgamey[RCo] joined #rosettacode
04:18 Dgamey[RCo]_ joined #rosettacode
04:22 Dgamey[RCo] left #rosettacode
04:24 Mathnerd314 joined #rosettacode
04:30 Dgamey[RCo]_ left #rosettacode
04:36 parsleyfirefly joined #rosettacode
05:26 shortcircuit Updated packages on the server. Doesn't seem to have fixed the pages which are persistently returning HTTP 500.
05:26 shortcircuit Rebuilding SMW data, on the possibility that it's DB corruption. Can't do more than that tonight.
05:26 shortcircuit I'll poke at it again tomorrow evening.
05:42 parsleyfirefly left #rosettacode
05:58 Coderjoe_ what pages are giving 500?
06:00 SoniaKeys http://rosettacode.org/wiki/Arrays is one I ran across
06:07 Coderjoe_ oh nice
06:18 Coderjoe_ did geshi get upgraded?
06:18 Coderjoe_ is now known as Coderjoe
06:20 SoniaKeys didn't hear.  but that reminds me, i saw some bad coloring...i should go see if it's fixed and report it if not...
06:23 SoniaKeys mm, the inconsistency is gone in the new version, so, i guess it's not updated on rc.
06:27 BenBE2 Good morning ^^
06:27 SoniaKeys speak of the devil
06:28 BenBE2 Coderjoe Yes. It got. 1.0.8.9 with most latest fixes for 1.0.8.10; see my blog for some details.
06:29 BenBE2 Python will receive some fixes soon; those will also introduce a Python Console mode ;-)
06:29 SoniaKeys http://rosettacode.org/wiki/Non-decimal_radices/Output#Go is an example of what I saw.  Only one Printf is highlighted.
06:29 fedaykin "Non-decimal radices/Output - Rosetta Code" http://rldn.net/XpeP
06:29 Coderjoe BenBE2: um... it seems to be hitting the max execution time on some RC pages
06:30 Coderjoe BenBE2: I'm going to guess you are the go-to for geshi, then?
06:35 BenBE2 CodeBlock Jep. Project Maintainer of GeSHi and RC is a welcome guineapig ;-)
06:36 SoniaKeys lol.  but wikipedia is a big beneficiary, right?
06:36 Coderjoe works (but somewhat slow): http://rosettacode.org/mw/index.php?title=Arrays&direction=next&oldid=63411
06:36 fedaykin "Arrays - Rosetta Code" http://rldn.net/9bkv
06:36 Coderjoe the next newer rev works
06:37 Coderjoe er
06:37 Coderjoe I mean the next newer rev breaks
06:38 BenBE2 shortcircuit Can you try something regarding the GeSHi integration?
06:38 Coderjoe is there a way to get MW to show a diff of the wikisource, without also trying to render the source to html?
06:38 BenBE2 Since large pages with lot of code seem to sometimes break: What about caching code snippets and avoid rehighlighting them, thus continuing the highlighting where it left of the previous run?
06:39 BenBE2 CodeBlock just ask it for the rraw page.
06:39 BenBE2 Don't know the params for it though.
06:40 Coderjoe action=raw it seems
06:41 Coderjoe or view with diffonly
06:41 BenBE2 okay.
06:42 * CodeBlock looks around
06:42 * CodeBlock has been pinged.
06:42 Coderjoe uh...
06:42 Coderjoe http://rosettacode.org/mw/index.php?title=Arrays&diffonly=1&diff=next&oldid=63412
06:42 fedaykin "Arrays - Rosetta Code" http://rldn.net/6YA
06:42 Coderjoe the only differences between the working and first broken revisions would be changing lang from sh to bash
06:45 BenBE2 hmmm.
06:46 BenBE2 Can you try if the bash scripts highlight here? http://benbe.home.omorphia.de:43815/geshi-misc/profiling/geshi-trunk/contrib/example.php
06:46 fedaykin "GeSHi examples" http://rldn.net/7uK
06:48 Coderjoe testing
06:49 Coderjoe I seem to have found a problem case
06:49 Coderjoe alist[${#alist[*]}]=new_item
06:50 Coderjoe yep. finally got a error page with max execution time as the error and a call stack
06:52 BenBE2 That's bash???
06:53 Coderjoe yeah
06:54 Coderjoe ${#alist[*]} is the number of items in the $alist array
06:54 BenBE2 Let me debug this for a moment ;-)
06:54 Coderjoe and alist[...]=new_item is assigning a value at that point
06:54 Coderjoe at least I am pretty sure it is valid bash
06:56 Coderjoe granted, I don't use arrays in bash terribly often, so I could be incorrect
06:58 BenBE2 hmmm, that's weird ;-)
07:10 BenBE2 Ursache grob lakalisiert ...
07:11 BenBE2 Reason for the crash found ...
07:11 Coderjoe that's good
07:12 Coderjoe i hate when a bug evades attemts to track them down
07:12 BenBE2 I'll just upload this intermediate language file I have right now ...
07:12 BenBE2 Well, I have a prreliminary fix in place on my dev system right now ...
07:13 BenBE2 That's why it isn't happening there right now.
07:14 BenBE2 Reason: SOMETHING with the bash prompt detection causes a REALLY slow execution with PCRE (how I HATE that library!!!)
07:14 BenBE2 Bug should be fixed for now on RC.
07:15 BenBE2 shortcircuit Please clear the cache for the highlighting.
07:15 BenBE2 Let's get some debugging going:
07:15 BenBE2 4 => "/\A(?!#!)\s*(?>[\w:@\\/\\-\\._~]*[$#]\s?)?(?=[^\n]+\n?\Z)|^(?!#!)(\w+@)?[\w\\-\\.]+(:~?)[\w\\/\\-\\._]*?[$#]\s?/ms"
07:16 Coderjoe it appears to have allowed the page to load
07:16 Coderjoe uhhh
07:16 Coderjoe *head exploade*
07:20 BenBE2 Well, I also kinda get what was causing the problem ...
07:20 BenBE2 Currently reworking the regexp.
07:21 BenBE2 Short version: The regexp seems to allow empty matches ...
07:22 Coderjoe that would happen if there is a ?
07:22 Coderjoe iirc
07:22 Coderjoe though I am amused by the *?
07:23 Coderjoe what does that even do?
07:24 Coderjoe * is zero or more, and ? is 0 or 1
07:24 BenBE2 In this case the ? means ungreedy.
07:25 Coderjoe ah
07:25 BenBE2 I always match greedy by default and with a suffix-? after a quantifier I make that ungreedy.
07:25 Coderjoe yeah. looking at the perlre docs now
07:29 BenBE2 Okay, let's try again ...
07:29 BenBE2 \A(?!#!)(?:(?>[\w:@\\/\\-\\._~]*)[$#]\s?)(?=(?>[^\n]+)\n?\Z)|^(?!#!)(?:\w+@)?(?>[\w\\-\\.]+)(?>:~?[\w\\/\\-\\._]*?)?[$#]\s?
07:29 BenBE2 GeSHi uses PCRE (Perl /Compatible) Regular Expressions ;-)
07:31 Coderjoe yes. I went to the PCRE site to look for docs on the re metachars it handles, and the link went to perldoc's perlre page
07:33 BenBE2 Well, the fixed line in the Bash language file now is the following:
07:33 BenBE2 4 => "/\A(?!#!)(?:(?>[\w:@\\/\\-\\._~]*)[$#]\s?)(?=(?>[^\n]+)\n?\Z)|^(?!#!)(?:\w+@)?(?>[\w\\-\\.]+)(?>:~?[\w\\/\\-\\._]*?)?[$#]\s?/sm"
07:34 BenBE2 Please note two things:
07:34 BenBE2 1. This line contains ?> markers which will trip PHP off if you comment out this line with //
07:35 BenBE2 2. This line switched the modifiers m and s to indicate the bruteness of this PCRE ;-)
07:36 BenBE2 (3. if this was in a released version I now would have to allocate a CVE for a DoS)
07:37 Coderjoe while it makes the re harder to read, you could split the string up between the ? and >
07:38 Coderjoe which would alleviate #1
07:38 BenBE2 Well, no ;-) The ?> doesn't matter with double quotes OR if you comment out this PCRE with /* */.
07:39 Coderjoe no... "blabla?" . ">blahblah" or such
07:40 Coderjoe which would allow // to work
07:40 BenBE2 It's always (?> ...
07:40 BenBE2 I usually don't want people to mess around with the regexp in a file.
07:42 BenBE2 Well, just updating my live site at qbnz.com/highlighter to eliminate the problem there.
07:42 fedaykin "GeSHi - Generic Syntax Highlighter :: Home"
07:44 BenBE2 AND fixed in my blog at blog.benny-baumann.de/?cat=3
07:44 fedaykin "GeSHi << BenBE's humble thoughts"
07:48 BenBE2 Any other issues you found?
07:49 Coderjoe not yet
07:49 BenBE2 Well, send me a mail or report them in #GeSHi.
07:49 Coderjoe though [[Arrays]] loaded last time i checked
07:49 fedaykin http://rosettacode.org/wiki/Arrays
07:50 Coderjoe and now it is faster than it was after the first fix
07:53 BenBE2 Well, that can't be objectively true, because the first fix was to comment out the offending Regexp.
07:53 BenBE2 The impression it is faster now is due to RC's caching of the rendered page.
07:53 Coderjoe ah
07:53 Coderjoe well that would do it
07:54 BenBE2 That's why I asked shortcircuit to clear the cache of the highlighting, because the fix will have some minor effects on the display of some of the bash snippets.
07:55 Coderjoe i just used the mw refresh tab to purge the cached version of the page and it went about the same as the first successful render
07:56 Coderjoe i had forgotten that MW caches the rendered page
07:58 BenBE2 See?
07:58 BenBE2 BTW: Also just uploaded a small fix for asm.php (x86 assembly)
07:58 BenBE2 There was equ missing in the list of directives.
08:12 MigoMipo joined #rosettacode
08:47 SoniaKeys left #rosettacode
09:43 MigoMipo_ joined #rosettacode
09:46 MigoMipo left #rosettacode
10:00 FireFly joined #rosettacode
10:43 MigoMipo_ left #rosettacode
12:06 FireFly left #rosettacode
13:38 Mathnerd314 left #rosettacode
14:10 SoniaKeys joined #rosettacode
15:08 parsleyfirefly joined #rosettacode
15:32 FireFly joined #rosettacode
16:55 BenBE2 left #rosettacode
16:55 BenBE2 joined #rosettacode
17:29 BenBE2 left #rosettacode
17:36 BenBE joined #rosettacode
17:38 BenBE left #rosettacode
17:47 BenBE joined #rosettacode
17:49 BenBE left #rosettacode
17:57 BenBE joined #rosettacode
18:00 BenBE left #rosettacode
18:08 BenBE joined #rosettacode
18:58 mwn3d_phone left #rosettacode
19:00 mwn3d_phone joined #rosettacode
19:31 tylercurtis joined #rosettacode
19:48 MigoMipo joined #rosettacode
21:33 BenBE2 joined #rosettacode
21:33 BenBE left #rosettacode
22:00 parsleyfirefly left #rosettacode
22:03 Mathnerd314 joined #rosettacode
22:07 tylercurtis left #rosettacode
22:51 SoniaKeys left #rosettacode
22:59 Mathnerd314 left #rosettacode
23:06 MigoMipo left #rosettacode
23:06 Mathnerd314 joined #rosettacode
23:29 * shortcircuit has a wtf moment.
23:30 shortcircuit As in, wtf doesn't mysqldup dumps to XML default to encoding blobs to 7-bit safe strings?
23:32 Coderjoe because?
23:32 Coderjoe utf8'd?
23:32 shortcircuit xmlstarlet puked it, complaining that it wasn't UTF-8.
23:32 Coderjoe there's also an option to hex blobs
23:33 shortcircuit I see that. Now I need to go in and verify I can pull these things back in.
23:33 Coderjoe dunno if it works with xml output. I'm using normal sql output
23:40 FireFly left #rosettacode
23:47 shortcircuit Ok, that's more like it.
23:51 * shortcircuit moved some folders around so BenBE2 and kodiarfer can still access and mod them, but they get swept up in a broader backup sequence.
23:53 kpreid joined #rosettacode

| Channels | #rosettacode index | Today | | Search | Google Search | Plain-Text | summary