Pike-devel November 2003

pike-devel@lists.lysator.liu.se

7 participants
35 discussions

Start a nNew thread

-DNOT_INSTALLED
by Henrik Grubbstr�m (Lysator) ＠ Pike (-) developers forum 13 Nov '03

13 Nov '03

If you are able to trigg the bug; please add some debug to lib/modules/Getopt.pmod, and check where it goes wrong.

2 2

Can I get detailed documents about Pike languages design
by chen david 13 Nov '03

13 Nov '03

hi, I am a Pike user and very interested in Pike languages design. I downloaded the source code of pike-v7.4.28, but did't find any design documents, such as the theory of Pike interpreter, the architecture of Pike construction, etc. BTW, I have some experiences of simple compiler design and compiler optimization. How can I begin to understand the Pike design and furthermore do some development to it? Any suggestion or help will be highly appreciated. Thanks. Regards. David _________________________________________________________________ Add photos to your messages with MSN 8. Get 2 months FREE*. http://join.msn.com/?page=features/featuredemail

3 2

Re: module modref
by Johan Sundstr�m (f�ngar heffaklumpar) ＠ Pike (-) developers forum 12 Nov '03

12 Nov '03

Or, better still, cook your own refdoc/structure file, to only list documentation entities you have and want in the result? That file was at least intended to be pluggable and to make it easier to do lots of different views of the refdocs. / Johan Sundström (fångar heffaklumpar) Previous text: >2003-11-12 01:54: >Subject: Re: module modref >-------------------------------------------------------------------- >ah, ok. I'll try ignoring the return on that step then. thanks! > >On Tuesday, November 11, 2003, at 07:30 PM, Martin Nilsson (saturator) >@ Pike (-) developers forum wrote: > >> I think that failure is nothing more than a warning. It is anyway >> caused >> by the <insert-move entity="_Roxen"/> and the likes in the xml files in >> refdoc/structure. >> >> / Martin Nilsson (saturator) >> >> Previous text: >>> 2003-11-12 00:44: >>> Subject: module modref >>> -------------------------------------------------------------------- >>> I'm working on creating a module only modref, but am running into the >>> following problem: >>> >>> I've created the autodoc.xml containing only the markup for my module. >>> Then I run the following command: >>> >>> pike -x assemble_autodoc >>> /usr/local/pike/7.5.12/doc//src/structure/modref.xml \ >>> plib/autodoc.xml >plib/modref.xml >>> >>> And get the following output: >>> >>> Parsing structure file >>> "/usr/local/pike/7.5.12/doc//src/structure/modref.xml". >>> Executing reference expansion and queueing node insertions. >>> Parsing autodoc file "plib/autodoc.xml". >>> Executing node insertions. >>> Failed to move predef::._Roxen >>> Failed to move predef::.__builtin >>> >>> Can anyone with more knowledge of the autodoc system help me out with >>> this >>> one? It looks like there must be some references to these "modules" >>> but i >>> don't know how to remove those dependencies. >>> >>> Any thoughts? >>> >>> Bill >>> >>> >>> >>> / Brevbäraren >>> > > > > / Brevbäraren >

1 0

module modref
by Bill Welliver 11 Nov '03

11 Nov '03

I'm working on creating a module only modref, but am running into the following problem: I've created the autodoc.xml containing only the markup for my module. Then I run the following command: pike -x assemble_autodoc /usr/local/pike/7.5.12/doc//src/structure/modref.xml \ plib/autodoc.xml >plib/modref.xml And get the following output: Parsing structure file "/usr/local/pike/7.5.12/doc//src/structure/modref.xml". Executing reference expansion and queueing node insertions. Parsing autodoc file "plib/autodoc.xml". Executing node insertions. Failed to move predef::._Roxen Failed to move predef::.__builtin Can anyone with more knowledge of the autodoc system help me out with this one? It looks like there must be some references to these "modules" but i don't know how to remove those dependencies. Any thoughts? Bill

2 1

pike external module documentation
by Bill Welliver 11 Nov '03

11 Nov '03

I've started the first draft of a document describing how to build modules outside of the pike source tree. Keeping in mind it's only about 45 minutes old, please take a look at it and let me know if there are any corrections or additions to it at this point. I've not gotten into any of the goodies about the module repository, but that will come soon enough. I think the only requriements that should happen from that standpoint would be a recommendation that a constant __version should be placed in the class of the module, and that the module distribution directory be saved as FullModuleName-ver. https://steam.iaeste.at/home/Pike%20Community/tutorials/External%20Modules/ Note that there is both a HTML version and a StarWriter version. If anyone wants to make changes, at this point, please make them to the SW version (or just send me your updates). As always, comments, suggestions and donations of time from technical writing enthusiasts are welcome! Bill

2 2

Problem with MIME module
by David Gourdelier 10 Nov '03

10 Nov '03

Hello, I am using the MIME module of Pike 7.2 and I think it is broken when the mail contains the Content-Transfer-Encoding header. Indeed the RFC says the header is of type: encoding := "Content-Transfer-Encoding" ":" mechanism mechanism := "7bit" / "8bit" / "binary" / "quoted-printable" / "base64" / ietf-token / x-token But in the Pike code of MIME.pmod we have: if (headers["content-transfer-encoding"]) { array(string) arr=tokenize(headers["content-transfer-encoding"]); if (sizeof(arr)!=1 || !stringp(arr[0])) throw (({ "invalid Content-Transfer-Encoding in message\n", backtrace() })); transfer_encoding = lower_case(arr[0]); } And MIME.tokenize for a valid value such as Content-Transfer-Encoding: 8bit returns: Result: ({ /* 3 elements */ "Content-Transfer-Encoding", 58, "8bit" }) So in the Pike code the size of arr is neither 1 unless no encoding is specified (sizeof(arr)!=1 in the code). I think the offending code must be replaced by: if(!stringp(arr[2])) throw... string transfer_encoding = lower_case(arr[2] || ""); multiset possible_values = (< "7bit", "8bit", "binary", "quoted-printable", "base64" >); if(!possible_values[transfer_encoding]) throw... Can I change it this way ? PS: It seems to break especially with some Lotus Notes servers. / David Gourdelier

4 4

A note on .so files in 7.5
by Johan Sundstr�m (f�ngar heffaklumpar) ＠ Pike (-) developers forum 08 Nov '03

08 Nov '03

It seems .so files since recently (?) *need* to have a .pmod file that inherits them. A heads-up about such changes here is never wrong; it's not trivial to track down bugs related to such sudden incompatibility.

2 2

Nettle cipher.pmod
by Niels M�ller (igelkottsr�ddare) ＠ Pike (-) developers forum 08 Nov '03

08 Nov '03

I've made a slight reorganization of cipher.cmod: : Use a custom struct pike_cipher to keep track of information about : ciphers, instead of Nettle's primitive struct nettle_cipher. : : The only significant difference is that the pike version uses : set_key-functions that deal with invalid or weak keys, and throws pike : errors. : : The check_key_size method is gone, but that's the only user-visible : change. To test if a key is valid, just try to use it, and catch the : error. It passes the testsuite, but I haven't done any testing beyond that.

1 0

Hohum
by Per Hedbor () ＠ Pike (-) developers forum 08 Nov '03

08 Nov '03

aido/pike-modules/GLUE.pmod/module.pmod:1242:Wrong return type. aido/pike-modules/GLUE.pmod/module.pmod:1242:Expected: { BaseTexture = object(implements 65701) } aido/pike-modules/GLUE.pmod/module.pmod:1242:Got : object(is 65735) aido/pike-modules/GLUE.pmod/module.pmod:1242:Type of identifier "texture_type" does not match. aido/pike-modules/GLUE.pmod/module.pmod:1242:Expected: int(34382..34382) aido/pike-modules/GLUE.pmod/module.pmod:1242:Got : int(3553..3553) aido/pike-modules/GLUE.pmod/module.pmod:1310:Class definition failed. Eh, yes? I want to have different constant values in my different classes. That was the whole point of that specific constant.

1 0

| on arrays
by Martin Stjernholm, Roxen IS ＠ Pike developers forum 08 Nov '03

08 Nov '03

I have a feeling this probably has been discussed before, but please refresh my memory in that case. When | is used on arrays, all elements that occurs in both follows the order in the second operand: > ({1,2}) | ({2,1}); (1) Result: ({ /* 2 elements */ 2, 1 }) This has surprised users on several occasions, and it's also at odds with how & handles it: > ({1,2}) & ({2,1}); (2) Result: ({ /* 2 elements */ 1, 2 }) The old tutorial says that the order is taken from the left argument, but afaics it has never been that way for |. So I think the behavior in | is a bug and would like to fix it (maybe with compat provisions). But I see there's a test in the testsuite that appears to test this very consciously: test_equal( ({1,2,3,4,4}) | ({3,5,6}), ({1,2,4,4,3,5,6})) So before I go ahead and break something, I'd like to if there's a thought behind it and what it is in that case.

2 1

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 November 2003