Pike-devel September 2004

pike-devel@lists.lysator.liu.se

8 participants
36 discussions

MySQL big_query in 7.7
by Martin Stjernholm, Roxen IS ＠ Pike developers forum 10 Apr '05

10 Apr '05

I think the new streaming behavior in Sql.mysql.big_query is nice and as it should be, but it's causing too much headaches in existing applications - the old de-facto behavior is too ingrained. So a proposal is to instead do like this: query: Like today - slow and memory consuming. big_query: A variant of query() that still is nonstreaming. I.e. it only avoids the very bulky response format of query(). streaming_query: Like big_query but guaranteed to stream if it exists. It's database dependent whether other queries can be issued while a streaming_query response object exists. It's of course also database dependent if there are locks on the tables in the server while the response object exists. While at it, we could also consider generalizing the big_typed_query interface that the oracle module provides: big_typed_query: Like big_query but doesn't convert everything to strings. Integers and floats are kept that way, Date/timestamps are Calendar objects, there are objects for representing the NULL value for each type. streaming_typed_query: Streaming variant of big_typed_query.

9 20

memory leak in oracle module
by Michael J. Stenitzer ＠ Pike developers forum 14 Oct '04

14 Oct '04

[i moved this discussion to the developers' forum as i think it is more suitable for my problem.] thanks to martin and marcus for your hints for debugging. unfortunately it did not bring any results. what i did so far: * i discovered a memory-leak when i use roxen 4+ or chilimoon running with pike 7.4.*, 7.6.* (i have tested many different releases and fresh from cvs). this happens with a website connecting to an oracle db. memory demand grew to several hundred MBs within an hour or two of rather heavy requests. * this has been reproduced on three different hosts (all running linux). * roxen / debug information / Pike memory usage information shows more or less constant memory demand (total: 12000 kB). * the same website on roxen Pike v7.2 release 543 shows no memory leaks (constant memory demand ~38 MB). i used the same oracle libraries so i guess they are not leaking. * compiling pike 7.6 with-dmalloc and running roxen with: twice i managed to run and exit roxen cleanly but i got no memory leaks reported (last message was "counting bytes") * running hilfe with pike 7.6 with-dmalloc and doing some queries on oracle did also not report any memory leaks after exiting normally. * tried to run roxen or pike with valgrind but it failed. # valgrind --leak-check=yes --trace-children=yes --tool=memcheck (see error messages below) i have never used valgrind before and could not find any hints with google. anybody else knows how to use valgrind with pike? any further advice? /michael =====8<=====8<=====8<=====8<===== pike error messages when running with valgrind: /usr/local/pike/7.6.25/lib/master.pike:396: Failed to index module 'Files' with 'Stat' (module doesn't exist?) /usr/local/pike/7.6.25/lib/master.pike:1322: Failed to index module 'Files' with 'Stat' (module doesn't exist?) /usr/local/pike/7.6.25/lib/master.pike:2212: Failed to index module 'Files' with 'Stat' (module doesn't exist?) /usr/local/pike/7.6.25/lib/master.pike:2225: Cannot inherit program which is not fully compiled yet. /usr/local/pike/7.6.25/lib/master.pike:2230: Cannot inherit program which is not fully compiled yet. /usr/local/pike/7.6.25/lib/master.pike:3863: Failed to index module 'Builtin' with 'array_iterator' (module doesn't exist?) /usr/local/pike/7.6.25/lib/master.pike:4231: Cannot inherit program which is not fully compiled yet. /usr/local/pike/7.6.25/lib/master.pike:4238: No inherit or surrounding class Encoder. /usr/local/pike/7.6.25/lib/master.pike:4243: Cannot inherit program which is not fully compiled yet. / Michael J. Stenitzer Previous text: >2004-09-24 09:16: >Subject: memory leak in oracle module >-------------------------------------------------------------------- >First of all, if I recall correctly, dmalloc takes quite a bit of time >to finish up. If there is a leak, it is possible that your pike is >acctually doing something useful. > >Another option, might be to attempt to run roxen in Valgrind. I don't >know if this can be done or if it is recommenden at all, but it could >be worth a try... > > / Marcus Agehall (PacketFront) >

8 33

Bug in replace
by David Gourdelier 07 Oct '04

07 Oct '04

Hi, Replace doesn't seem to like \. : string s = Pike v7.6 release 13 running Hilfe v3.5 (Incremental Pike Frontend) > string s = "foo.bar.com"; > replace(s, ".", "\."); (1) Result: "foo.bar.com" > replace(s, "a", "\."); (2) Result: "foo.b.r.com" > replace(s, "a", "\b"); (3) Result: "foo.b\br.com" It does the same with 7.2, 7.4 / David

11 24

is there a `||()
by Martin Bähr 02 Oct '04

02 Oct '04

hi, i am wondering why there seem to be no functions for || or &&. what is the reason for that? greetings, martin.

6 7

^C
by Mirar ＠ Pike developers forum 01 Oct '04

01 Oct '04

Recently, when I exit hilfe with ^C, it doesn't reset the terminal correctly. Is this a recent change in hilfe, or is it my system? If I exit with ^D, it resets the terminal, so I suspect Pike.

2 2

Stdio.File() bug?
by Martin Bähr 29 Sep '04

29 Sep '04

hi, http://pike.ida.liu.se/generated/manual/modref/ex/predef_3A_3A/Stdio/File/c… claims that the default open mode is "rw", yet hilfe says different: Pike v7.6 release 7 running Hilfe v3.5 (Incremental Pike Frontend) > Stdio.File("/tmp/ANNOUNCE")->mode(); (1) Result: 4099 > Stdio.File("/tmp/ANNOUNCE", "rw")->mode(); (2) Result: 12291 is this a bug in the docs or in the implementation? greetings, martin.

5 4

Odd failure mode whith stray precompiled .o-files
by Mikael Brandstr�m (ogg! ogg! ogg!) ＠ Pike (-) developers forum 29 Sep '04

29 Sep '04

I just compiled and installed a pike 7.6 from CVS, to realise that I had to remove a gtk-package which was not available on another computer. After redoing a make install and running my GTK-program, I got this error message: /opt/pike/pike/7.6.25/lib/modules/GTK.pmod.o:-: Warning: Decode failed: Error while decoding program(/opt/pike/pike/7.6.25/lib/modules/GTK.pmod): Bad number of identifiers in inherit: 965 != 691 /opt/pike/pike/7.6.25/lib/modules/GTKSupport.pmod/Alert.pike.o:-: Warning: Decode failed: Error while decoding program(/opt/pike/pike/7.6.25/lib/modules/GTKSupport.pmod/Alert.pike): Bad number of identifiers in inherit: 130 != 113 /opt/pike/pike/7.6.25/lib/modules/GTKSupport.pmod/pDrawingArea.pike.o:-: Warning: Decode failed: Failed to get program from ID 1116 After removing all stray .o-files and remaking install it worked fine. I think it would be good if the make install target could remove old installed .o files if the same version of pike is reinstalled, or if there was a make reinstall which did it.

1 0

memory leak in oracle module
by Per Hedbor () ＠ Pike (-) developers forum 27 Sep '04

27 Sep '04

It's a roxen 4.0.158, as distributed from download.roxen.com, btw.

4 6

foreach(x, y) vs foreach(x; y; z)
by Martin Bähr 27 Sep '04

27 Sep '04

hi, what was the reason for foreach(x; y; z)? why is it not possible to have the parser distinguish foreach(x, y) from foreach(x, y, z)? i remember reading that there is a backwards compatibility problem, but i don't see where this would be. greetings, martin.

4 4

duplicates in a|b
by Martin Bähr 25 Sep '04

25 Sep '04

hi, when doing a union (|) on arrays it appears as if the first array remains as is, and then any missing values from the second array are appended, with duplicates in the first array remaining. a = ({ 1, 7, 2, 5, 4 }); b = ({ 2, 1, 3, 5, 5 }); a|b is ({ 1, 7, 2, 5, 4, 3, 5 }) b|a is ({ 2, 1, 3, 5, 5, 7, 4 }) the reason for these questions is that the tutorial claims: the elements in the result can come in any order and duplicates may or may not be removed. should pike be (and remain) that unpredictable? greetings, martin.

6 11

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

Pike-devel September 2004