From: UnixOS2 Archive To: "UnixOS2 Archive" Date: Fri, 2 Aug 2002 04:35:05 EST-10EDT,10,-1,0,7200,3,-1,0,7200,3600 Subject: [UnixOS2_Archive] No. 291 ************************************************** Thursday 01 August 2002 Number 291 ************************************************** Subjects for today 1 Re: Lindows : John Drabik" 2 Re: New Lynx : Patrick Ash 3 Re: PIPE & TEE : Maynard" 4 Re: New Lynx : Patrick Ash 5 Re: Lindows : tstevic at attglobal.net 6 Re: PIPE & TEE : Maynard" 7 RE: Lindows : Dave Webster 8 Re: New Lynx : Patrick Ash 9 Re: New Lynx : John Poltorak 10 Re: Building gettext v0.11.4 : John Poltorak 11 Re: New Lynx : John Poltorak 12 Re: New Lynx : John Poltorak 13 Re: PIPE & TEE : DWParsons at t-online.de (Dave Parsons) 14 Lindows : John Poltorak 15 Re: PIPE & TEE : John Poltorak 16 GNU Hello : John Poltorak 17 Re: Lindows : Stefan Neis 18 Re: New Lynx : John Poltorak 19 Re: New Lynx : Michel SUCH" 20 Re: PIPE & TEE : DWParsons at t-online.de (Dave Parsons) 21 openssl-0.9.7-beta3 : Michel SUCH" 22 Re: PIPE & TEE : Stefan Neis 23 Re: GNU Hello : Stefan Neis 24 Re: New Lynx : Stefan Neis **= Email 1 ==========================** Date: Fri, 02 Aug 2002 08:30:51 -0600 (MDT) From: "John Drabik" Subject: Re: Lindows On Fri, 2 Aug 2002 15:16:00 +0100, John Poltorak wrote: >Has anyone come across Lindows? It is basically just Linux, with the WINE emulator. Oh, and some marketing hype. >Apparently it is something which enables Linux to run Windows apps. The WINE part is. You could also try CodeWeavers, VMWare, and probably a few others. >I guess it must be something like Odin but it does suggest that it can be >used for running MS apps, which I don't think Odin can. Well, "most" applications anyway. YMMV. >Is source code available for Lindows? Wonder how well it would compile on >OS/2... Maybe it uses wxWindows to convert between the Win GUI and X... You should be able to find the source for all of the underpinnings (Linux, WINE, etc.). However, I'm guessing that they have at least added some form of configurator or ??? You might not be able to get source to that. John **= Email 2 ==========================** Date: Fri, 02 Aug 2002 08:53:47 +0000 From: Patrick Ash Subject: Re: New Lynx I really cannot give an answer. I am using this version now, and have not seen this problem. I can go to the url you specified using both the runlynx.cmd file and calling lynx directly. Can you generate a trace log to see what is happening? runlynx -trace will generate a file called lynx.trace and place it in $home. Pat On Fri, 02 Aug 2002 11:59:16 +0100, John Poltorak wrote: >On Tue, Jul 30, 2002 at 08:05:10AM +0000, Patrick Ash wrote: >> I applied the patches to the lynx 2.8.4 rel.1 source. I also added >> the recent fix for the gridtext.c problem reported by Walter Ian >> Kaye. > >I've just tried this new version but it crashes as soon as a page is >displayed. When running:- > >runlynx ftp://ftp.gnu.org/pub/gnu > >I get:- > > >A Fatal error has occurred in Lynx/2 Ver. 2.8.4rel.1b > >Please notify your system administrator to confirm a bug, and >if confirmed, to notify the lynx-dev list. Bug reports should >have concise descriptions of the command and/or URL which causes >the problem, the operating system name with version number, the >TCPIP implementation, and any other relevant information. > >Do NOT mail the core file if one was generated. > >Lynx now exiting with signal: 11 > > >Abnormal program termination >core dumped > >I've been using Lynx Version 2.8.1dev.13 for around four years and thought >I'd give a more recent version a try. -- Patrick Ash patash at comcast.net This OS/2 system uptime is 0 days, 23:38 hours and 05 seconds **= Email 3 ==========================** Date: Fri, 02 Aug 2002 09:44:20 -0500 (CDT) From: "Maynard" Subject: Re: PIPE & TEE >> My example for this is 'tar --help | more' >> which doesn't do the 'more' part. >> > >Try:- > >tar --help 2>&1 | more Exactly; but why is this good? why should this output be to stderr anyway, or is it? I'm not sure what it is, but something in this tar compilation or source is not good; this behavior is anomolous and not consistent across other applications or platforms. So, the best thing to do is to fix the program, not kludge the user, eh. -- Maynard **= Email 4 ==========================** Date: Fri, 02 Aug 2002 09:57:18 +0000 From: Patrick Ash Subject: Re: New Lynx I tried to build lynx without requiring any special dll's. [D:\lynx2-8-4]chk4dlls lynx.exe Loading DLL 'emx' --> D:\EMX\DLL\EMX.DLL. Loading DLL 'EMXLIBCM' --> D:\EMX\DLL\EMXLIBCM.DLL. Loading DLL 'doscalls' --> loaded. Loading DLL 'pmwin' --> C:\OS2\DLL\PMWIN.DLL. Loading DLL 'Z' --> D:\EMX\DLL\Z.DLL. Loading DLL 'viocalls' --> C:\OS2\DLL\VIOCALLS.DLL. Loading DLL 'moucalls' --> C:\OS2\DLL\MOUCALLS.DLL. All DLL's used by 'lynx.exe' could be loaded. Perhaps there is something in your environment that is causing a conflict. I don't see anything obviously wrong in the trace, but Thomas is much more experienced in analyzing these than I. Pat On Fri, 02 Aug 2002 14:27:44 +0100, John Poltorak wrote: >On Fri, Aug 02, 2002 at 08:53:47AM +0000, Patrick Ash wrote: >> I really cannot give an answer. I am using this version now, and have >> not seen this problem. I can go to the url you specified using both >> the runlynx.cmd file and calling lynx directly. >> >> Can you generate a trace log to see what is happening? >> >> runlynx -trace will generate a file called lynx.trace and place it in >> $home. > >It doesn't show anything out of the ordinary. At the end it has:- > >HTParse: result:http://www.ibm.com >HTParse: aName:`http://www.ibm.com/us/' > relatedName:`' >HTParse: (ABS) >HTParse: result:/us/ >Starting realm is 'http://www.ibm.com/us/' > >GridText: HText_pageDisplay at line 1 started >_Switch_Display_Charset(cp=45, really=1). >Display font set to 'auto-cp437'. >UCChangeTerminalCodepage: Switch_Display_Charset(45) returned 0 > >GridText: Not showing link, hightext=__ >not found, assume . >found 97, x_offset=-16. > >When I run it with the -trace option I get an access violation in >EMXLIBCM. > >I suspect the problem is due to something like a DLL conflict or a wrong >environment setting. -- Patrick Ash patash at comcast.net This OS/2 system uptime is 1 day, 00:41 hours and 36 seconds **= Email 5 ==========================** Date: Fri, 02 Aug 2002 10:29:20 -0400 From: tstevic at attglobal.net Subject: Re: Lindows I heard about Lindows on NPR. My understanding is that it is a ultra cheep ( $300 - $400 ) computer system that runs a custom flavor of Linux, with a bunch of apps already set up so it 'look' like a windows box to the user. They said it was going to be sold at Wall Mart. The guy they interviewed said he planned to make money offering additional software on a subscription basis. In <20020802151600.F11551 at eyup.org>, on 08/02/02 at 03:16 PM, John Poltorak said: >Has anyone come across Lindows? >Apparently it is something which enables Linux to run Windows apps. I >guess it must be something like Odin but it does suggest that it can be >used for running MS apps, which I don't think Odin can. >Is source code available for Lindows? Wonder how well it would compile on > OS/2... Maybe it uses wxWindows to convert between the Win GUI and X... -------------------------------------------- tstevic at attglobal.net V.O.I.C.E. RexxLA TeamOS/2 ' I'm just to lazy to use windows.................... ' **= Email 6 ==========================** Date: Fri, 02 Aug 2002 10:41:57 -0500 (CDT) From: "Maynard" Subject: Re: PIPE & TEE On Fri, 2 Aug 2002 15:51:25 +0100, John Poltorak wrote: >This version of TAR is unfixable since no one can get hold of the source. I suppose that the extended option of 'fix' then is to 'replace' ;-} But, back to proper business at hand ... -- Maynard **= Email 7 ==========================** Date: Fri, 2 Aug 2002 11:21:16 -0500 From: Dave Webster Subject: RE: Lindows Yes Odin/WINE and wxWindows are entirely different techniques for arriving at the same destination. One can also add Java into that mix. Odin tries, with only very marginal results, to take code written under the WIN32 API and provide a low level translation layer map WIN32 calls to PM calls (or a series of PM calls) to get a result. Java depends on a runtime to hide the OS differences. wx is neither. Instead of a one for one WIN32 call in wxMSW, wxOS2 impalements the various class methods, often using a very different approach than used in the same method under wxMSW. The result is a common class public interface that in reality, in many cases, does things very differently under the covers to arrive at the same result. -----Original Message----- From: Stefan Neis [mailto:neis at cdc.informatik.tu-darmstadt.de] Sent: Friday, August 02, 2002 9:57 AM To: os2-unix at eyup.org Subject: Re: Lindows On Fri, 2 Aug 2002, John Drabik wrote: > >I guess it must be something like Odin but it does suggest that it can be > >used for running MS apps, which I don't think Odin can. > > Well, "most" applications anyway. YMMV. And IIRC, Odin is also using WINE, so the difference shouldn't be all that big (except for man power when it comes to "fine-tuning" the code to the base OS needs). > > (wxWindows) Certainly not. wxWindows is fine for writing code which can compiled to either use Win GUI or plain X or GTK or Motif, but it doesn't help at all if your code wasn't written to use wxWindows (instead of directly using e.g. MFC or Motif) - and even less if you have only a binary, > You should be able to find the source for all of the underpinnings > (Linux, WINE, etc.). I seem to remember some heated debates about Lindows not publishing all their changes - at least not yet ... Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'. **= Email 8 ==========================** Date: Fri, 02 Aug 2002 11:34:09 +0000 From: Patrick Ash Subject: Re: New Lynx I have looked at the various version that I have compiled in the past. I do not actively link to this dll, but I find that anything I have compiled with gcc 2.9.5 or later requires this dll. 2.8.1 and 2.9.1 do not seem to require it. My build script has remained pretty much the same for all versions. Pat On Fri, 02 Aug 2002 15:08:14 +0100, John Poltorak wrote: >> [D:\lynx2-8-4]chk4dlls lynx.exe >> >> Loading DLL 'pmwin' --> C:\OS2\DLL\PMWIN.DLL. > >Do you have to have this one? It means that it will only work under PM, >but since it's a text mode browser, I would expect that it ought to be >capable of working under TSHELL... > -- Patrick Ash patash at comcast.net This OS/2 system uptime is 1 day, 02:18 hours and 28 seconds **= Email 9 ==========================** Date: Fri, 2 Aug 2002 11:59:16 +0100 From: John Poltorak Subject: Re: New Lynx On Tue, Jul 30, 2002 at 08:05:10AM +0000, Patrick Ash wrote: > I applied the patches to the lynx 2.8.4 rel.1 source. I also added > the recent fix for the gridtext.c problem reported by Walter Ian > Kaye. I've just tried this new version but it crashes as soon as a page is displayed. When running:- runlynx ftp://ftp.gnu.org/pub/gnu I get:- A Fatal error has occurred in Lynx/2 Ver. 2.8.4rel.1b Please notify your system administrator to confirm a bug, and if confirmed, to notify the lynx-dev list. Bug reports should have concise descriptions of the command and/or URL which causes the problem, the operating system name with version number, the TCPIP implementation, and any other relevant information. Do NOT mail the core file if one was generated. Lynx now exiting with signal: 11 Abnormal program termination core dumped I've been using Lynx Version 2.8.1dev.13 for around four years and thought I'd give a more recent version a try. > Pat -- John **= Email 10 ==========================** Date: Fri, 2 Aug 2002 14:15:19 +0100 From: John Poltorak Subject: Re: Building gettext v0.11.4 On Thu, Aug 01, 2002 at 11:48:08AM +0200, Franz Bakan wrote: > On Thu, 1 Aug 2002 10:30:53 +0100, John Poltorak wrote: > > >This actually does build using the standard autoconf, configure, make > >routine, and I do manage to create an INTL.DLL, but it doesn't work with > >anything I've tried it with. > > If you _read_ \os2\README.OS2 you will know why it does not work It mentions something about backward compatibility with 0.10.35. Is this something that we necessarily must have? When starting to put together a set of apps from scratch, why would I need backward compatibility. Presumably that is for any existing apps, so if I built those apps themselves would any need for backward compatibility disappear. I don't want to include any legacy stuff which isn't required so if I rebuild the apps which have failed, is there any prospect of those apps running correctly with this new INTL library? > Franz -- John **= Email 11 ==========================** Date: Fri, 2 Aug 2002 14:27:44 +0100 From: John Poltorak Subject: Re: New Lynx On Fri, Aug 02, 2002 at 08:53:47AM +0000, Patrick Ash wrote: > I really cannot give an answer. I am using this version now, and have > not seen this problem. I can go to the url you specified using both > the runlynx.cmd file and calling lynx directly. > > Can you generate a trace log to see what is happening? > > runlynx -trace will generate a file called lynx.trace and place it in > $home. It doesn't show anything out of the ordinary. At the end it has:- HTParse: result:http://www.ibm.com HTParse: aName:`http://www.ibm.com/us/' relatedName:`' HTParse: (ABS) HTParse: result:/us/ Starting realm is 'http://www.ibm.com/us/' GridText: HText_pageDisplay at line 1 started _Switch_Display_Charset(cp=45, really=1). Display font set to 'auto-cp437'. UCChangeTerminalCodepage: Switch_Display_Charset(45) returned 0 GridText: Not showing link, hightext=__ not found, assume . found 97, x_offset=-16. When I run it with the -trace option I get an access violation in EMXLIBCM. I suspect the problem is due to something like a DLL conflict or a wrong environment setting. > Pat > > -- > Patrick Ash > patash at comcast.net -- John **= Email 12 ==========================** Date: Fri, 2 Aug 2002 15:08:14 +0100 From: John Poltorak Subject: Re: New Lynx On Fri, Aug 02, 2002 at 09:57:18AM +0000, Patrick Ash wrote: > I tried to build lynx without requiring any special dll's. > > [D:\lynx2-8-4]chk4dlls lynx.exe > > Loading DLL 'pmwin' --> C:\OS2\DLL\PMWIN.DLL. Do you have to have this one? It means that it will only work under PM, but since it's a text mode browser, I would expect that it ought to be capable of working under TSHELL... > Perhaps there is something in your environment that is causing a > conflict. I don't see anything obviously wrong in the trace, but > Thomas is much more experienced in analyzing these than I. I guess it needs some analysis of the core dump, but I have no idea what to do with these files. Does anyone know of any core dump analyzer programs which might be able to generate some sort of report? > Pat -- John **= Email 13 ==========================** Date: Fri, 02 Aug 2002 15:09:22 +0200 (CES) From: DWParsons at t-online.de (Dave Parsons) Subject: Re: PIPE & TEE On Sat, 20 Jul 2002 23:00:17 -0500 (CDT), Maynard wrote: > On Sat, 20 Jul 2002 21:55:35 -0300 (ADT), Lord Spigol wrote: > > >Yes, some utils does not output like normal things. > > > >I have problems with the pipe from some unix utils. > > > >There is a way to solve it only by recompile? > > My example for this is 'tar --help | more' > which doesn't do the 'more' part. > Try:- tar --help 2>&1 | more -- Dave **= Email 14 ==========================** Date: Fri, 2 Aug 2002 15:16:00 +0100 From: John Poltorak Subject: Lindows Has anyone come across Lindows? Apparently it is something which enables Linux to run Windows apps. I guess it must be something like Odin but it does suggest that it can be used for running MS apps, which I don't think Odin can. Is source code available for Lindows? Wonder how well it would compile on OS/2... Maybe it uses wxWindows to convert between the Win GUI and X... -- John **= Email 15 ==========================** Date: Fri, 2 Aug 2002 15:51:25 +0100 From: John Poltorak Subject: Re: PIPE & TEE On Fri, Aug 02, 2002 at 09:44:20AM -0500, Maynard wrote: > > >> My example for this is 'tar --help | more' > >> which doesn't do the 'more' part. > >> > > > >Try:- > > > >tar --help 2>&1 | more > > Exactly; but why is this good? why should this output be to stderr > anyway, or is it? I'm not sure what it is, but something in this tar > compilation or source is not good; this behavior is anomolous and not > consistent across other applications or platforms. > > So, the best thing to do is to fix the program, not kludge the user, > eh. This version of TAR is unfixable since no one can get hold of the source. > -- > Maynard -- John **= Email 16 ==========================** Date: Fri, 2 Aug 2002 16:22:49 +0100 From: John Poltorak Subject: GNU Hello If you have never tried bilding a GNU app from scratch, I'd suggest giving GNU Hello a try:- ftp://ftp.gnu.org/pub/gnu/hello/hello-2.1.1.tar.gz If you have a basic toolkit installed, including autoconf, you ought to be able to build it without any trouble and maybe even work out what autoconf actually does. I'm hoping so at least. I'd like to see if I can get it to build using the external intl.lib which I made with GETTEXT 0.11.4... Can anyone explain why I end up with a binary which needs 116kB to say? :- Hello, world! -- John **= Email 17 ==========================** Date: Fri, 2 Aug 2002 16:56:55 +0200 (CEST) From: Stefan Neis Subject: Re: Lindows On Fri, 2 Aug 2002, John Drabik wrote: > >I guess it must be something like Odin but it does suggest that it can be > >used for running MS apps, which I don't think Odin can. > > Well, "most" applications anyway. YMMV. And IIRC, Odin is also using WINE, so the difference shouldn't be all that big (except for man power when it comes to "fine-tuning" the code to the base OS needs). > > (wxWindows) Certainly not. wxWindows is fine for writing code which can compiled to either use Win GUI or plain X or GTK or Motif, but it doesn't help at all if your code wasn't written to use wxWindows (instead of directly using e.g. MFC or Motif) - and even less if you have only a binary, > You should be able to find the source for all of the underpinnings > (Linux, WINE, etc.). I seem to remember some heated debates about Lindows not publishing all their changes - at least not yet ... Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'. **= Email 18 ==========================** Date: Fri, 2 Aug 2002 16:58:06 +0100 From: John Poltorak Subject: Re: New Lynx On Fri, Aug 02, 2002 at 05:40:49PM +0100, Michel SUCH wrote: > > I missed beginning of this thread, but could it be that lynx cannot access > terminfo? > Check out the terminfo environment variable and see if it points to the > right location. If terminfo can't be found I get this error msg:- Error opening terminal: ansi-color-2. Lynx does actually start up OK as well as retrieving and displaying the requested URL. Actually, I've just got it working... I needed to change the screen size to mode 80,50. I normally run it in mode 80,70. I've just found that it dumps under that mode. > ---------------------------- > Michel SUCH TEAM OS/2 FRANCE > ICQ # 51654489 -- John **= Email 19 ==========================** Date: Fri, 02 Aug 2002 17:40:49 +0100 (CET) From: "Michel SUCH" Subject: Re: New Lynx I missed beginning of this thread, but could it be that lynx cannot access terminfo? Check out the terminfo environment variable and see if it points to the right location. On Fri, 2 Aug 2002 11:59:16 +0100, John Poltorak wrote: >On Tue, Jul 30, 2002 at 08:05:10AM +0000, Patrick Ash wrote: >> I applied the patches to the lynx 2.8.4 rel.1 source. I also added >> the recent fix for the gridtext.c problem reported by Walter Ian >> Kaye. > >I've just tried this new version but it crashes as soon as a page is >displayed. When running:- > >runlynx ftp://ftp.gnu.org/pub/gnu > >I get:- > > >A Fatal error has occurred in Lynx/2 Ver. 2.8.4rel.1b > >Please notify your system administrator to confirm a bug, and >if confirmed, to notify the lynx-dev list. Bug reports should >have concise descriptions of the command and/or URL which causes >the problem, the operating system name with version number, the >TCPIP implementation, and any other relevant information. > >Do NOT mail the core file if one was generated. > >Lynx now exiting with signal: 11 > > >Abnormal program termination >core dumped > >I've been using Lynx Version 2.8.1dev.13 for around four years and thought >I'd give a more recent version a try. > >> Pat > > >-- >John > > > ---------------------------- Michel SUCH TEAM OS/2 FRANCE ICQ # 51654489 **= Email 20 ==========================** Date: Fri, 02 Aug 2002 18:02:34 +0200 (CES) From: DWParsons at t-online.de (Dave Parsons) Subject: Re: PIPE & TEE On Fri, 02 Aug 2002 09:44:20 -0500 (CDT), Maynard wrote: > > >> My example for this is 'tar --help | more' > >> which doesn't do the 'more' part. > >> > > > >Try:- > > > >tar --help 2>&1 | more > > Exactly; but why is this good? why should this output be to stderr > anyway, or is it? I'm not sure what it is, but something in this tar > compilation or source is not good; this behavior is anomolous and not > consistent across other applications or platforms. It isn't good in my opinion. I can't see why they would consider help information to be error output but many UNIX originated programs behave this way. Some won't even give you any help without using man. > > So, the best thing to do is to fix the program, not kludge the user, > eh. > If you can get the sources and have got the spare time, then rewrite them. Otherwise you will have to use the redirection operators. -- Dave **= Email 21 ==========================** Date: Fri, 02 Aug 2002 20:50:32 +0100 (CET) From: "Michel SUCH" Subject: openssl-0.9.7-beta3 I would like to especially thank Brian Havard for what he has done for openssl. The 0.97 really builds from out of the box after running his os2-emx.cmd. thanks again ---------------------------- Michel SUCH TEAM OS/2 FRANCE ICQ # 51654489 **= Email 22 ==========================** Date: Fri, 2 Aug 2002 21:08:54 +0200 (CEST) From: Stefan Neis Subject: Re: PIPE & TEE On Fri, 2 Aug 2002, Maynard wrote: > Exactly; but why is this good? why should this output be to stderr > anyway, or is it? I think that's where it was supposed to be going back at that time ... > this behavior is anomolous and not > consistent across other applications or platforms. I've seen lots of this behaviour, also on other platforms - I think it's in fact rather common for _older_ versions of GNU tools. On most other platforms it's never been a problem thanks to xterms ability of scrolling back, but with the standard CMD window, it's really quite annoying ... > So, the best thing to do is to fix the program, not kludge the user, > eh. If just somebody could donate a few days to porting a current tar version ... ;-) Honestly, that's something that would be _way_ down on my todo list, almost everything else is more interesting to me ... Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'. **= Email 23 ==========================** Date: Fri, 2 Aug 2002 21:10:48 +0200 (CEST) From: Stefan Neis Subject: Re: GNU Hello On Fri, 2 Aug 2002, John Poltorak wrote: > Can anyone explain why I end up with a binary which needs 116kB to say? :- > > Hello, world! Well, it's hard to say without knowing your linker flags, but not using -Zcrtdll and building a debug version could be a sufficient explanation. Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'. **= Email 24 ==========================** Date: Fri, 2 Aug 2002 21:17:42 +0200 (CEST) From: Stefan Neis Subject: Re: New Lynx On Fri, 2 Aug 2002, Patrick Ash wrote: > I find that anything I > have compiled with gcc 2.9.5 or later requires this dll. BTW, that's gcc 2.95.something, not 2.9.5. ;-) Anyway, wouldn't that behaviour suggest to stay with gcc-2.8.1?+ C++-handling on Intel isn't really _that_ much better with those gcc-2.95 versions, so unless you fall on something which really doesn't compile with gcc-2.8.1, what's the point? Gcc-3 is a different beast, but with the announcement of the upcoming gcc-3.2 being binary incompatible to all earlier versions _again_, existing gcc-3 versions on OS/2 aren't really a tempting choice for more than pure gcc-testing purposes either. :-( Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'.