Pike-devel

pike-devel@lists.lysator.liu.se

3044 discussions

Re: Inherited files being compiled
by Martin Stjernholm, Roxen IS ＠ Pike developers forum 22 Aug '04

22 Aug '04

It's intentional that the master register programs that fail to compile. The reason is to avoid cascade errors when the same module is resolved repeatedly. / Martin Stjernholm, Roxen IS Previous text: >2004-07-08 10:04: >Subject: Re: Inherited files being compiled >-------------------------------------------------------------------- >I'll just send this e-mail to you directly. I'm not sure if I found the problem or not but it seems as if the unregister function inside of the master object is being called like it should. I did end up fixing the problem by calling the unregister function by my own means which I will show you below: > >//! This function is called when we want to clone an object >object|int clone_object(string sFile) >{ > /* Remove unneccesary parts */ > if(sFile[0] == '/') > { > sFile = sFile[1 ..]; > } > > /* Prepend the filename to include the path to Emage */ > sFile = "C:/EMAGE/" + sFile; > > /* Make sure the path is relevant to pike standards */ > sFile = normalize_path(sFile); > > if(programs[sFile]) > { > return ((program)sFile)(); > } > > /* Check to make sure the file exists */ > if(file_stat(sFile)) > { > program p = 0; > > mixed err = catch > { > p = compile_file(sFile); > }; > > if(err) > { > unregister(p); > > sscanf(sFile, "C:/EMAGE%s", sFile); > > Send("Error compiling " + sFile + ". Please fix any/all errors.\n"); > return 0; > } > else > { > /* Return the cloned object */ > return (p()); > } > } > > /* Throw an error if file does not exist. */ > error("File " + sFile + " not found.\n"); > > /* No file exists return 0 */ > return 0; >} > >I hope this helps in one way or another. It seems to me as if the calling program to be compiled is not the one being unregistered by the compiler but yet the inherited file is the one being called. Just an off the wall idea. I am in no way a professional programmer. Thanks for your suggestion anyways though. > >"Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum" <10353(a)lyskom.lysator.liu.se> wrote: >Try running pike with --autoreload. > >/ Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) > >Previous text: >>2004-07-08 02:15: >>Subject: Inherited files being compiled >>-------------------------------------------------------------------- >>I have what seems to be a simple problem but is causing me a huge headache. The problem is that if an inherited file does not compile it still adds that file to master->programs. Now I am not sure if that has anything to do with it but doesn't this cause the file not to be recompiled when called again? I am just wondering if there was a fix I could plug in so all i would have to do is make the fix, upload the file and carry on instead of having to restart the master. >> >> >> >>--------------------------------- >>Do you Yahoo!? >>New and Improved Yahoo! Mail - 100MB free storage! >> >> >> / Brevbäraren >> > > >--------------------------------- >Do you Yahoo!? >Yahoo! Mail - Helps protect you from nasty viruses. > > > / Brevbäraren >

1 0

out of memory in allocate
by Martin Nilsson (DivX Networks) ＠ Pike (-) developers forum 19 Aug '04

19 Aug '04

I would like to have the following > allocate(1617191936); Program received signal SIGSEGV, Segmentation fault. 0x08092695 in low_allocate_array (size=1617191936, extra_space=0) at /home/nilsson/Pike/7.7/src/array.c:118 118 ITEM(v)[e].u.integer=0; generate an out of memory message instead (if indeed that is the problem). But the malloc call do returns a pointer, so I don't know what is wrong.

7 18

managing comport settings in linux
by Marcus Agehall (PacketFront) ＠ Pike (-) developers forum 17 Aug '04

17 Aug '04

Can you ever with these things? It all boils down to the fact that ioctls are very lowlevel, system specific and you shouldn't poke around with them unless you know what you are doing. By implementing my suggestion, you can implement default ioctls for "standard" functions and the user can easily implement her own ioctls as needed. But, since this is dangerous stuff for the average user-joe, I would suggest making it a compiletime feature which has to be enabled with --enable-user-ioctls or something similar. / Marcus Agehall (PacketFront) Previous text: >2004-08-17 13:35: >Subject: managing comport settings in linux >-------------------------------------------------------------------- >Can you guarantee that the process wont crash if you don't feed it the >wrong information? > > / Mirar >

5 5

managing comport settings in linux
by Marc Dirix ＠ Pike developers forum 17 Aug '04

17 Aug '04

Ok I'll have go on it. /Marc / Marc Dirix Previous text: >2004-08-17 14:36: >Subject: managing comport settings in linux >-------------------------------------------------------------------- >Sorry. Lost my CVS-repository to a fsck a few months ago. However, it >is not very hard to fix up a cmod that accomplishes what you need. > >Look at some (fairly simple) cmod in the pike source and I'm sure you >can figure it out. I usually start with the Bz-module, rip out almost >everything and then add my stuff. Just copy the post_modules/Bz dir to >post_modules/MyIOCTL and start changing... ;) > > / Marcus Agehall (PacketFront) >

1 0

managing comport settings in linux
by Marcus Agehall (PacketFront) ＠ Pike (-) developers forum 17 Aug '04

17 Aug '04

I had kind of the same problem about 1½ ago. I needed to interface with my soundcard and thus needed ioctl. However, there was some argument against adding a general ioctl call, but I'm not sure what it was. I ended up using a tiny cmod for the purpose of setting my parameter. It might be that there is no real way of knowing what parameters are passed and how they should be interperted on the C-level. But that could be fixed using some smart heuristics, IMHO. / Marcus Agehall (PacketFront) Previous text: >2004-08-17 08:47: >Subject: managing comport settings in linux >-------------------------------------------------------------------- >It seems like that ioctl is missing from Pike. :/ > >I wonder if ioctl should be added on a general basis, or if special >calls should be added for all ioctls... > >(crossposting developers list) > >Anyhow, it should be added somehow to src/modules/files/termios.c or >nearby. It shouldn't be too tricky. If I have time, I'll do it today >(so it lands in 7.7-cvs), but if not I'm leaving for austria tomorrow >morning, so can't do anything then for some weeks :/ > > / Mirar >

3 4

managing comport settings in linux
by Henrik Grubbstr�m (Lysator) ＠ Pike (-) developers forum 17 Aug '04

17 Aug '04

Not a good idea; the design philosophy of Pike is that Pike-level code should never be able to cause memory corruption at C-level. With ioctls there's no way to know what data a specific ioctl wants without reading the documentation for that specific ioctl for that specific OS. ioctls and fcntls are a way for device handlers to hook in APIs that aren't standard across all filedescriptors. / Henrik Grubbström (Lysator) Previous text: >2004-08-17 13:32: >Subject: managing comport settings in linux >-------------------------------------------------------------------- >What would be nice, is a way to define ioctl methods in >Pike. Perhaps something like > >object i = Stdio.IOCTL(4242 /* request_id */ ); >i->add_int_parameter(); >i->add_string_parameter(); >i->set_return_type(STRING); > >string ret = i(fd, 1, "somestring"); > > >That ought to be possible, or..? > > / Marcus Agehall (PacketFront) >

2 1

managing comport settings in linux
by Mirar ＠ Pike developers forum 17 Aug '04

17 Aug '04

Can you guarantee that the process wont crash if you don't feed it the wrong information? / Mirar Previous text: >2004-08-17 13:32: >Subject: managing comport settings in linux >-------------------------------------------------------------------- >What would be nice, is a way to define ioctl methods in >Pike. Perhaps something like > >object i = Stdio.IOCTL(4242 /* request_id */ ); >i->add_int_parameter(); >i->add_string_parameter(); >i->set_return_type(STRING); > >string ret = i(fd, 1, "somestring"); > > >That ought to be possible, or..? > > / Marcus Agehall (PacketFront) >

1 0

managing comport settings in linux
by Mirar ＠ Pike developers forum 17 Aug '04

17 Aug '04

It seems like that ioctl is missing from Pike. :/ I wonder if ioctl should be added on a general basis, or if special calls should be added for all ioctls... (crossposting developers list) Anyhow, it should be added somehow to src/modules/files/termios.c or nearby. It shouldn't be too tricky. If I have time, I'll do it today (so it lands in 7.7-cvs), but if not I'm leaving for austria tomorrow morning, so can't do anything then for some weeks :/ / Mirar Previous text: >2004-08-16 23:47: >Subject: managing comport settings in linux >-------------------------------------------------------------------- >I've narrowed it down to this please advise if this is possible in >pike. >They do this: >mdlns &= ~TIOCM_RTS; >ioctl(dmm,TIOCMSET,&mdlns) > >Which from "man 4 tty_ioctl" means: > >TIOCMSET const int *argp > set the status of modem bits. > >So they activly set the RTS (hardware control) bit high on the port. > >/Marc > > / Marc Dirix >

1 0

pike 7.6 for windows?
by Martin Bähr 14 Aug '04

14 Aug '04

hi, this is a report from the irc front. more and more chatters are are firing questions about a windows build for 7.6 in our direction and we have no urls to fire back at them. what is the status of a windows build? is there one? will there be one? what is the holdup? (explaining people why it's not there is better than just telling them "we don't know") ah, the last attacker just left, hope it remains peacefull for a while... greetings, martin.

6 5

CMOD questions
by Marcus Agehall (PacketFront) ＠ Pike (-) developers forum 13 Aug '04

13 Aug '04

It's been a while since I did any cmods, but now I need to write one for work related stuff. All questions are in the context of late CVS 7.7. Q1: How should I define my classes to get the following function: > indices(Foo); ({ get_bar }) > object x = Foo(); > indices(x); ({ get_bar }) > get_bar(); Foo.bar(); Ie, I want a class within Foo that can only be instantiated using the Foo.get_bar() method. Q2: Is there a way that a Foo.bar() object (as described above) can access the variables defined in Foo? (The variables which are accessed thru the THIS-pointer that is.) References to where things like this is already done within the Pike-source is appreciated.

2 9

Jump to page:

2026

2025

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