S.E.O. for cygwin.com
Dear cygwin.com Team, I went to your Site cygwin.com and wanted to shoot you a quick note. If you want I can make a few changes (aesthetically and/or SEO-wise) to make your site convert more visitors into leads and to get it placed higher in the organic search for a few of the select terms. This is NOT like one of those foreign emails you probably get in your inbox every day. Just to be upfront we have 50+ agents that work for development /SEO. I would just need to know which (if not both) services you're open to checking out information about, either web design or SEO. Would you be open to seeing more brief info / quote for what I would like to accomplish? Internet Marketing Experts - We know how to make Google Happy. If you are not on Google page 1, your website is a waste. Getting your site on Google 1st page GUARANTEED. All the techniques used are ethical and proprietary. Please reply back to get the *Website Audit Report* we have done. Best Regards, Shane Bell : DM Expert Caution: This is our marketing strategy that we use the Gmail. If you want to receive detailed plan of action, please feel free to reply Yes, We will get back to you with Details. [image: beacon] -- Problem reports: https://cygwin.com/problems.html FAQ: https://cygwin.com/faq/ Documentation:https://cygwin.com/docs.html Unsubscribe info: https://cygwin.com/ml/#unsubscribe-simple
Cygwin 1.7 Windows 7/2008 Public key intermittent problem
Hi, I have about 8 machines experiencing problems while trying to use passwordless public key authentication, via passwd -R. Happening on W7, w2k8, x86 x64 sshd running on all as a domain user, with the correct local security policy changes made, and /var/empty owner by sshd user etc. pub key auth always works for the user sshd is been run as, but not any other user - connection closed by... error - windows event log does show sshd pid xxx fatal initgroups permission denied error If I reboot any of the above, generally passwordless pub key auth doesn't work - if I then rdp to windows machine, login as another user, log out - I can then ssh using keys to that machine as the user I RDP'd as - until it gets rebooted again. And then - after several more reboots - without any changes been made on the machines - I can use keys to login to the machine as any user successfully - until it gets rebooted and reverts to the usual problem I have tried installing cyglsa - but after installing and rebooting - I don't notice any difference - although I'm unsure if there is something else needed for cyglsa to function or is adding the users to the passwd/groups sufficient? Thanks for any suggestions - the fact that it intermittently works is baffling me Cheers Shane -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: 1.7 Public Key Authentication problem
Thanks for the info - I wasn't aware of passwd -R - just tried it and it works which is a good relief. It's a dev lab - anyone with access to the keys is allowed full rights to the machines - so security not a major concern. BTW - I had installed cyglsa-config and rebooted and gave the users the Act as part of OS right - but it doesn't work for me. I must be missing something . Thanks again - you've saved me considerable problems! On 2010/02/03 10:07 PM, shane fenton wrote: Hi, First time poster - so hopefully will get it right :) Cygwin 1.7 installed on approx 10 machines - XP /2008 domain cyg_server user created Added above user to Quotas/create token/replace token log on as service local admins on pc's added cyg_server to passwd file ssh-host-config (found above user and used it and did the right perms on /var/empty /var/log/sshd.log ) added domain user accounts to passwd domain users group group You didn't mention whether you set up the LSA authentication package (with /usr/bin/cyglsa-config), or used 'passwd -R' for each user. Did you try either of those? The Cygwin User Guide goes into great detail about the methods of changing user context, in this chapter: http://cygwin.com/cygwin-ug-net/ntsec.html The gist of that chapter is this: If you want to be able to login via ssh as a user that is not running the sshd daemon, you have basically two options: (1) Provide a valid Windows password to the sshd daemon, either interactively (which you obviously don't want to do, since you're attempting public key auth), or stored statically in the registry via 'passwd -R'. (2) Use the LSA authentication package. Bear in mind that if you use this option to avoid giving sshd your password entirely, I believe that certain privileges are withheld from the logged in user. [I don't remember exactly what privs are missing in this case... access to network resources maybe?] Hope this helps, -SM -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
1.7 Public Key Authentication problem
Hi, First time poster - so hopefully will get it right :) Cygwin 1.7 installed on approx 10 machines - XP /2008 domain cyg_server user created Added above user to Quotas/create token/replace token log on as service local admins on pc's added cyg_server to passwd file ssh-host-config (found above user and used it and did the right perms on /var/empty /var/log/sshd.log ) added domain user accounts to passwd domain users group group I can ssh in the machines as any user using password logon. But I can only successfully login using keys as the cyg_server user itself. Using keys for any user logs me in successfully, but I get the below error and most things don't work 4 [main] -bash 1368 fork: child -1 - CreateProcessW failed, errno 13 -bash: fork: Permission denied If I change the ssh daemon to run as a different domain user, with the required privliges set in the local security policy - I get the same result. ie. the daemon user can log in passwordless with keys, but all other users generate the same error - yet password logons always work. Please help! I've been trying for a few days and have not been able to make any progress - I've been reading the faq's/mailing lists but I couldn't find the same issue. Thanks in advance Shane -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
esniper complains cURL 7.1.1 needed...
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 I have cURL 7.16.3 installed (checked via curl -V at the prompt). I have downloaded esniper, tried to run ./configure but it complains that configure: error: cURL 7.1.1 or newer required. Before the stop, it also flags checking for curl_easy_setopt... no. Any help in resolving this would be greatly appreciated. -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.7 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFITHq+RZgP6dNgieoRAtvPAJwO5er72ARKXID0DOlCLjRQR1qu1gCfWn3N XjLrP53x0JUiVxQM+6y+mVY= =St5u -END PGP SIGNATURE- -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Weird Dialog Problem
Sorry that I don't have a solution to your problem, but I have the same problem if it makes you feel any better. I also recently switched from exceed to cygwin in order to run Gambit and I have the same annoying problem with the dialogs. If you ever do find anything out about it, please let me know too. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://x.cygwin.com/docs/ FAQ: http://x.cygwin.com/docs/faq/
Checking XCOPY Exit Value in Cygwin Bash
Hi all, I am writing a automated build script for my project that will be run under cygwin. I will copy my updated source files to the build directory and if there are updated files, the executables will be built. To copy the source files, I had to use XCOPY since the directory structure should be preserved in the destination directory also. To copy only the updated files, I used the /D switch for XCOPY. Now since I want to execute the source compile only if files in the build directory have been updated, I have to use the exit codes of XCOPY inside the script. I tried checking the value of $! after executing XCOPY but it didnt work. I couldn't find a solution in the internet too. Currently I am piping the standard output to a file and checking if the number of files copied is 0 or not. But I think this is not an elegant solution. This is what I am doing now. [script] copied=false # Helper Function copy_files() { echo copying *.$1 files in $2 to $3\\$2 xcopy /DSYI $2\\*.$1 $3\\$2 | tee copy.log while read amount ; do if [ ${amount::1} != 0 ]; then copied=true; fi done copy.log } cd ../source copy_files h. ..\\build copy_files c. ..\\build copy_files cpp . ..\\build rm -f copy.log ! $copied echo Files up-to-date. Skipping build exit 0 cd ../build # Start the Build Process [/script] Can you please provide me a way of checking the XCOPY exit code: reference [http://www.microsoft.com/resources/documentation/windows/xp/all/proddocs/en-us/xcopy.mspx?mfr=true] within Bash? Thank you for your time. Shane -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Checking XCOPY Exit Value in Cygwin Bash
Igor Peshansky wrote: Nope, you didn't have to. Something like (cd $2/.. find $2 -name *.$1 | tar cfT - -) | tar xfC - $3 would do the job of XCOPY /S using POSIX means. If you go POSIX, you can use the --keep-newer-files tar option. Of course it didn't. Please read a good bash tutorial, or the Special Parameters section of the bash manpage. Hi Igor and Mark, Thank you very much for the quick reply. I was initially using tar -cf - `find $source_dir -name *.$file_ext -print` | ( cd $dest_dir tar xBf - ) but it had a problem with path names with spaces. Obviously being not that good in bash scripting, I couldn't get over that issue. So that was why I decided to use the XCOPY command. I will use your method and see. Thanks again. I made a silly mistake in my former email. I was actually checking $? (not $!) for the exit code, but it didn't work. But I saw in a later reply from Mark that it worked for him. I will check it again. Maybe I was doing something silly. thanks again Regards Shane -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Checking XCOPY Exit Value in Cygwin Bash
David Christensen wrote: There are standard software development tools that solve the problems you are facing -- CVS and Make: http://ximbiot.com/cvs/wiki/index.php?title=Main_Page http://ximbiot.com/cvs/wiki/index.php?title=Main_Page Both are included in Cygwin. In the long run, you'd be better off investing in a basic to intermediate understanding of both rather than hacking together custom scripts to implement a subset of their functionality. To David, Thank you for the tip. Actually I am using Visual Source Safe as the Source Management tool. I was considering the use of CVS, but decided against at the last moment because most of the fellow developers including me, had been using VSS for a considerable amount of time, and felt that the migration from a VSS to CVS would take a some time. Similarly for Make. We are primarily a group of developers who are conversent with MS Windows than the Unix environment. Cygwin basically gives us the power of bash scripting and the ease of Windows at the same time. :) What I am trying to do is, checkout the source to the build directory and if there are any local changes in my working directory copy them to the build directory, build and do a test run from there. This is so that I can test my code before I do the actual check in. To Igor, Your method worked perfectly for paths with spaces too. :) Now if only I had a way of detecting if files were updated or not. To Mark I tried it again. Unfortunately echo $? gives 0 for both the cases of, number of files copied = 0 and, greater than 0. The link I posted from MSDN http://www.microsoft.com/resources/documentation/windows/xp/all/proddocs/en-us/xcopy.mspx?mfr=true says that XCOPY returns 1 when there were no files to be copied. So I guess I am back to square one. :( Thanks and best regards Shane GET FREE 5GB ONLINE STORAGE - Safely store your documents, photos and music online! Visit http://www.inbox.com/storage to find out more! -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Checking XCOPY Exit Value in Cygwin Bash
Igor Peshansky wrote: As David said, cvs has an easy way of doing this (using cvs diff and patch), which will also deal with local and checked in changes to the same file (while your method won't). Point taken. I will certainly look into it. Did you happen to notice the mention of the --keep-newer-files tar option in my original reply to you? Just add that to the last tar, and you will only copy the files that were changed in your copy (presumably by you) after the checked in version. Yeah I saw that reply and I had tried it. There were two problems. 1. New files will not be added to the build directory. It will say something like tar: ./Test/res/Test.manifest: Warning: Cannot stat: No such file or directory tar: Current `./Test/res/Test.manifest' is newer and the required manifest file is not copied into the build folder. So for the initial copy I have to use it without the --keep-newer-files, and for the subsequent copies I will have to use the --keep-newer-files. 2. This is the real problem. That is getting an indication whether none of the files were updated or not. I want to proceed with the rest of the building script only if more than one files have been copied. I do not know how to get that using the tar command. I tried echoing the $? value but it gives 0 all the time. The source compiler can detect if the sources were updated or not, on it's own, but there are a lot of projects in one Visual Studio Solution (about 60), that I can't wait until all those projects have been parsed. I am using while read amount ; do if [ ${amount::1} != 0 ]; then copied=true; fi done copy.log for that purpose. MSDN apparently lies. XCOPY for me returns non-zero on error, and 0 on normal execution (no matter how many files were copied). If that is the case, then there is no point in trying to check for the xcopy return value. As a short term solution I will stick with my original XCOPY solution. But I will try to find out what CVS, Make and the other tools have to offer. If there is a way of getting if files have been replaced using the tar command, I will try to implement that into my solution. Although I am fairly competent at programming in C/C++, this is my first attempt in writing a serious bash script, and I must admit that I am both impressed and overwhelmed by it's power. :) Thank you all for the help offered so far. Regards Shane GET FREE 5GB ONLINE STORAGE - Safely store your documents, photos and music online! Visit http://www.inbox.com/storage to find out more! -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Checking XCOPY Exit Value in Cygwin Bash
Christopher Faylor wrote: Is there some reason why you are not using cp to accomplish your task? cp --help should provide you with all sorts of options for copying files. You should be able to press cp into service for this. Using DOS utilities and DOS paths for this type of thing is putting you on the fringes of support for Cygwin. I really wouldn't recommend it. Clearly this is not such a Windows-specific problem that it outside of the capabilities of a UNIX solution. My initial attempt was with cp. But I didn't see a way of preserving the original directory structure of the source dir, inside the destination directory. XCOPY just seemed easier. Will have a go at 'cp' again. Thanks and regards Shane -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Checking XCOPY Exit Value in Cygwin Bash
What's wrong with cp -a or cp -r? It only copied files that were directly under the source directory. It didn't traverse the directories inside source recursively. I did some searching and, I came up with a similar thread. Finally the tar method was recommended in it too. Please refer : http://lists.samba.org/archive/samba/1999-December/016328.html and it's follow-ups. Thanks and Regards Shane -Original Message- From: [EMAIL PROTECTED] Sent: Sun, 6 Aug 2006 21:04:27 -0400 To: cygwin@cygwin.com Subject: Re: Checking XCOPY Exit Value in Cygwin Bash On Mon, Aug 07, 2006 at 06:52:40AM +0900, Shane wrote: Christopher Faylor wrote: Is there some reason why you are not using cp to accomplish your task? cp --help should provide you with all sorts of options for copying files. You should be able to press cp into service for this. Using DOS utilities and DOS paths for this type of thing is putting you on the fringes of support for Cygwin. I really wouldn't recommend it. Clearly this is not such a Windows-specific problem that it outside of the capabilities of a UNIX solution. My initial attempt was with cp. But I didn't see a way of preserving the original directory structure of the source dir, What's wrong with cp -a or cp -r? -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/ -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Keyboard and Mouse Input delay/pausing using 6.8.99.901
G'day All, I just updated to 6.8.99.901 from 6.8.2.0 for Xorg under Cygwin. Logging in via XDMCP to a remote hosts is causing me some issues now. I connect okay, but as soon as I enter some text into a Konsole or the like, a variable keyboard delay of 0.5-1.0 seconds seems to start. It also seems to be affecting mouse inputs as well. You click on a button and it ma react 0.5-1.0 seconds later. If I start the session and interact only with the mouse all is fine. As soon as I enter any text anywhere, this random delay/pausing starts to happen. It is strange that at the initial XDMCP login window, entering text does not seem to have any problems. The problem only starts after getting to the desktop stage. I get the effect with either KDE or Gnome desktops. It also effects Fedora5 and Ubuntu 6.06 systems. If I run version 6.8.2.0 the problem is resolved. If I use another X server like X-Win32 the problem is also resolved. I'd rather use the Cygwin X server because it seems much more compliant than other X servers for X11 and OpenGL work. Is anyone else getting this problem? Have Fun !! Shane Hill -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://x.cygwin.com/docs/ FAQ: http://x.cygwin.com/docs/faq/
Python shell ignores keys
Greetings, I'm trying to use the Python shell from within Cygwin. The complication is that I need to use the Windows version of Python not the Cygwin version, ultimately through ssh, but atleast from a cygwin bash prompt for now. The problem is that the Windows Python shell ignores or doesn't recv important keys like ctrl- z or ctrl-c and the up/down cursor keys move the cursor not the command line history. Otherwise everything seems to work fine (the scripts themselves run without an issue), atleast for what I need. I do quite a lot of python programming, its not really feasible (or a good idea) to maintain a second cygwin python installation. I realize these guys are chalk and cheese, but is it possible to fix this ? Windows XP Cygwin is current Python 2.4.3 Thanks, Shane -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Changing root path from /ecos-c/ to /
Hi Larry, Well, I wish it were that simple. That command doesnt work - it needs the Win32 file system to basic utilities like ls, etc. When eCos is installed, it somehow makes c: mount on /ecos-c, it previously was /. We're trying to find a way to put it back. Shane. Larry Hall wrote: At 12:20 AM 2/7/2005, you wrote: Hi, I'm using Cygwin. I wish to change the default path for C: from /ecos-c/ to /. When I type mount in cygwin, I get the following: C:\PFARM\cygwin on / type system (binmode) c: on /ecos-c type user (textmode) c: on / type user (textmode) e: on /cygdrive/e type user (textmode,noumount) f: on /cygdrive/f type user (textmode,noumount) h: on /cygdrive/h type user (textmode,noumount) i: on /cygdrive/i type user (textmode,noumount) C:\PFARM\examples\EVBA7Board\GNU\Simple\RAMCode The utility I'm using fails unless C: on the Windows box is mounted as '/'. Any suggestions? Shane. p.s. Whenever Cygwin generates a POSIX path from a Win32 one, it uses the longest matching prefix in the mount table. Thus, if C: is mounted as /c and also as /, then Cygwin would translate C:/foo/bar to /c/foo/bar. (http://www.cygwin.com/cygwin-ug-net/using.html#mount-table) Seems like you're answering your own question. Try 'umount -f /ecos-c'. -- Larry Hall http://www.rfk.com RFK Partners, Inc. (508) 893-9779 - RFK Office 838 Washington Street (508) 893-9889 - FAX Holliston, MA 01746 -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Changing root path from /ecos-c/ to /
Hi Larry, Tried it again, and it worked this time. Had to restart the shell. Thanks for your help. Shane. Larry Hall wrote: At 07:13 PM 2/8/2005, you wrote: Hi Larry, Well, I wish it were that simple. That command doesnt work - it needs the Win32 file system to basic utilities like ls, etc. Huh? 'umount' is a command that doesn't rely on any other command. It's a separate executable. See 'man umount' for more details. When eCos is installed, it somehow makes c: mount on /ecos-c, it previously was /. We're trying to find a way to put it back. True enough. And technically since this is an issue with eCos, you will really need to follow-up with them if 'umount' won't work for you for some reason and the docs I pointed you to don't help you either. -- Larry Hall http://www.rfk.com RFK Partners, Inc. (508) 893-9779 - RFK Office 838 Washington Street (508) 893-9889 - FAX Holliston, MA 01746 -- Regards, Shane Tolmie (BEng. Elec. Hons.) Managing Director DesignREM Ltd. Cell: +64(21)2977741 Phone: +64(3)3793012 Fax: +64(3)3660118 [EMAIL PROTECTED] www.designrem.com This email or attachments may contain confidential or legally privileged information intended for the sole use of the addressee(s). Any use, redistribution, disclosure, or reproduction of this message, except as intended, is prohibited. If you received this email in error, please notify the sender and remove all copies of the message, including any attachments. Thanks. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Installation Freezing 90%
Hi, It has now been, for some time, a problem that I have when installing cygwin. As the installation does its thing the memory usage rises, fair enough. But every time it reaches 90% it freezes. Here are the basics of what is shown on the screen. Installing. xorg-x11-f100-6.8.1.0-3 /etc/postinstall.xorg-x11-f100.sh And again this is when it reaches 90%. I even once tried closing as much down as I could on the computer i.e. little running processes and then trying installation. The same happened. I let installation take its path to see what would happen but then stopped it when it reached using about 369,340K Memory Usage. If needs be I could host a screen dump I took of the installation at this 90% stage host it and send you a link for you to view (It shows the installation, giving details of the above e.g. the file its installing and where to and also shows the percentage bars etc. It also shows Windows Task Manager giving details about the process. I have thoroughly searched the FAQ's section and am pretty sure that this has not been answered and I hope that you can help me with my problem. Regards, Shane Bryan -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
md5.sum in root of Cygwin mirror?
G'day All, Can someone remove the md5.sum in the Cygwin root that has zero file size. It is causing problems with the lftp mirroring program. Thanks.Have Fun !! Shane Hill = Shane D. Hill (B.E. Aero. Hons.) Aircraft Vehicles Division, Platforms Sciences Laboratory, Defence Science Technology Organisation, 506 Lorimer St, Fishermens Bend, Vic. 3207, AUSTRALIA. Email: Shane.Hill at dsto aot defence dot gov dot au = -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: socket() call failing?
On 18 Aug 2004 at 15:25, Jay Fenlason wrote: I've been trying to debug why Amanda won't work under cygwin for me any more, and I've traced it down to a socket() call in the sendbackup program. strace shows the following (massively snipped) 709 258623 [main] sendbackup 628 cygwin_socket: socket (2, 1, 6) 79416 338039 [main] sendbackup 628 wsock_init: res 0 281 338320 [main] sendbackup 628 wsock_init: wVersion 514 140 338460 [main] sendbackup 628 wsock_init: wHighVersion 514 352 338812 [main] sendbackup 628 wsock_init: szDescription WinSock 2.0 147 338959 [main] sendbackup 628 wsock_init: szSystemStatus Running 134 339093 [main] sendbackup 628 wsock_init: iMaxSockets 0 135 339228 [main] sendbackup 628 wsock_init: iMaxUdpDg 0 138 339366 [main] sendbackup 628 wsock_init: lpVendorInfo 0 8525 347891 [main] sendbackup 628 __set_winsock_errno: cygwin_socket:602 - winsock error 10106 - errno 1 263 348154 [main] sendbackup 628 cygwin_socket: -1 = socket (2, 1, 6) but I've no clue how to debug further than this. The socket call socket ( PF_INET, SOCK_STREAM, IPPROTO_TCP ) looks OK to me. . . -- JF P.S. strace -o foo ./amandad.exe will silently run /bin/amanda.exe instead of the one in your current directory. Perhaps the man page needs a stronger warning about legal pathnames for executables. :-) I am wondering if the latest cygwin has real trouble somewhere in its socket creation. Yours + these two seem to suggest there is a problem. http://www.cygwin.com/ml/cygwin/2004-08/msg00452.html http://www.cygwin.com/ml/cygwin/2004-08/msg00479.html I am wondering if the fix Corinna implemented for ssh might be relevant. - http://www.cygwin.com/ml/cygwin/2004-08/msg00702.html. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Apache CGI Scripts - Network layer permission denied
On 13 Aug 2004 at 17:32, Peter Flanigan wrote: When I run CGI scripts from Apache I'm getting permission denied errors. use Net::protoent; my $p = getprotobyname(shift || 'tcp'); if (defined $p) { printf(proto for %s is %d, aliases are %s\n, $p-name, $p-proto, @{$p-aliases}); } else { printf(Bad protocol 'tcp'\n); } This correctly turns ths string 'tcp' into the number 6 via lookup in /etc/protocols when run from the pdksh prompt. Returns the bad protocol error when run as a CGI script from Apache. Apache is being started (by apachectl start) from the same shell prompt which correctly runs the script. In general I'm finding that all network related calls (creating new sockets and similar) are failing. I also note (and think that it's related) that uname -n returns the correct mixed case hostname when run from the shell prompt but returns an incorrect (all upper case) string when run from the cgi-bin/test-cgi shell script. This eliminates Perl from the equation. I think this is the same problem as posted to this mailing list by: Richard Morse [EMAIL PROTECTED] Problems with Amanda and Cygwin 1.5.10 28/07/2004 and David A. Rogers [EMAIL PROTECTED] Socket problem w/ apache perl cgi 30/07/2004 I did not notice any follow up posts to these queries (my spam filter was rejecting some Cygwin mailing list posts). I think this may be the same problem that I reported here: http://www.cygwin.com/ml/cygwin/2004-08/msg00452.html Rolling back to the previous cygwin version is the only way I have been able to get connections via apache to remote databases. Creating sockets via apache failed wheres from perl command line it worked. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
prob with 1.5.10-3 prevents MySQL connect. 1.5.9-1 is OK.
Hello, With self- built Apache, MySQL, OpenSLL and Mod_SSL I have found an obscure bug in cygwin 1.5.10-3 that is not present in 1.5.9-1. I am installing clean onto XP Pro (without the default inclusion of Apache). I am then building OpenSSL, mod_SSL, and Apache from their sources. I am installing the latest perl DBI and DBD::mysql module. Everything is building and installing with successful make test runs using the script below. With Cygwin Core 1.5.9-1 I can successfully connect to a remote mysql server from a perl script in both the shell and also from a perl script called from a web page. Great! With Cygwin Core 1.5.10-3 I can *NOT* connect to the remote mysql server from a web page called script. Apache generates the error DBI-connect(database=test;host=my hostname) failed: Can't create TCP/IP socket (1) The same script called from a command line perl CAN conenct to the db without problem (after all the make test for DBD::mysql passed 100%) I found one other reference to a similar error from an older version without any resolution http://www.cygwin.com/ml/cygwin/2001-03/msg01748.html To be clear... everything compiles, tests, and installs without error using both Cygwin Core 1.5.9-1 and Cygwin Core 1.5.10-3 however with 1.5.10-3 I can not access any remote mysql databases from a web-called perl script where a command line called script is fine...go figure. There are no errors in building this kit with either version of the Cygwin dll - just that the old version workls for me when the current version does not. If anyone wants me to test something to resolve why this happens I'm happy to help...but otherwise I hope this helps the developers. Let me know If I can help further then with this report! Details on the scripts follow. If for nothing else this report will be archived to help someone else with the install g Cheers, Shane The perl test case script is: ### #!/usr/local/bin/perl use DBI; $database = test; $host = host.somewhere.com; $username = user; $password = pass; $dbh = DBI-connect(DBI:mysql:$database:$host, $username, $password); if (not $dbh) { print Content-type: text/html \n\n you cannot connect to the database; } else { print Content-type: text/html \n\n successful connect; } exit; # The script I use to clean build apache with mod_ssl, mod_auth_dbm and mod_rewrite is... # #script to clean install to Cygwin # openssl 0.9.7d # mod_ssl 2.8.19-1.3.31 # apache 1.3.31 with staticaly linked modules # could not get successful apache DSO install to work at all :( #!/bin/sh cd /usr/src/ # delete folders and re-extract clean from tarballs rm -rf mod_ssl-2.8.19-1.3.31 rm -rf apache_1.3.31 rm -rf openssl-0.9.7d tar -xzf mod_ssl-2.8.19-1.3.31.tar.gz tar -xzf apache_1.3.31.tar.gz tar -xff openssl-0.9.7d.tar.gz # install.sh has minor tweak for cygwin to correctly add .exe # find-dbm-lib has minor tweak for cygwin to find gdbm_compat for mod_rewrite cp /usr/src/find-dbm-lib /usr/src/apache_1.3.31/src/helpers/find-dbm- lib cp /usr/src/install.sh /usr/src/apache_1.3.31/src/helpers/install.sh cd /usr/src/openssl-0.9.7d ./config make cd /usr/src/mod_ssl-2.8.19-1.3.31 ./configure --with-apache=/usr/src/apache_1.3.31 --with- ssl=/usr/src/openssl-0.9.7d --prefix=/usr/local/bin/apache --enable- module=rewrite --enable-module=auth_dbm cd /usr/src/apache_1.3.31 make make install (certificate install done separately) #copy in my apache conf file cp /usr/src/httpd.conf /usr/local/bin/apache/conf/httpd.conf # minor tweak to /apache_1.3.31/src/helpers/install.sh change if [ -f $src.exe ]; then if [ -f $src ]; then : # Cygwin [ test ] is too stupid to do [ -f $src.exe ] [ ! - f $src ] else ext=.exe fi fi to if [ -f $src.exe ]; then ext=.exe fi for some reason the original screwed up adding .exe in my early installs. # minor tweak to /apache_1.3.31/src/helpers/find-dbm-lib so gdbm is found for mod_rewrite which for some reason could not find gdbm. if ./helpers/TestCompile lib gdbm dbm_open; then DBM_LIB=-lgdbm becomes if ./helpers/TestCompile lib gdbm_compat dbm_open; then DBM_LIB=-lgdbm_compat -lgdbm # -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: prob with 1.5.10-3 prevents MySQL connect. 1.5.9-1 is OK.
On 13 Aug 2004 at 14:27, Shane wrote: Hello, With Cygwin Core 1.5.9-1 I can successfully connect to a remote mysql server from a perl script in both the shell and also from a perl script called from a web page. Great! With Cygwin Core 1.5.10-3 I can *NOT* connect to the remote mysql server from a web page called script. Apache generates the error DBI-connect(database=test;host=my hostname) failed: Can't create TCP/IP socket (1) The same script called from a command line perl CAN conenct to the db without problem (after all the make test for DBD::mysql passed 100%) [..snip..] I need to add the in switching between the two versions of the cygwin base dll using the setup utility I had to also switch versions of Perl. All other installed modules were latest versions. cygwin 1.5.9-1 used perl 5.8.2-1 cygwin 1.5.10-3 used perl 5.8.5-2 missmatched versions give errors in perl. So to be clear the two systems (one working and one not) differ in both cygwin and perl versions (nothing else). Oddly both systems can connect to a remote DB from perl - just not when a browser calls the script from apache. Lastly - it makes no difference which version of cygwin apache was compiled with. If we take a working system and then just upgrade perl and cygwin without rebuilding the error occurs. Likewise a non- working system that downgrades cygwin and perl suddenly works ok. Hope this helps. Shane -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
sec:unclassified Cygwin 1.3.21-1 problem with rsh command
Hi All, After updating to cygwin 1.3.21-1 the rsh command seems to have broken. Here is an example: rsh myhost -e ls -l rsh: select: Bad file descriptor. Running rsh on its own (i.e. rlogin) seems to work okay. Have Fun !! Shane Hill --- Shane D. Hill (B.E. Aero. Hons.) Aircraft Vehicles Division Platforms Sciences Laboratory Defence Science Technology Organisation 506 Lorimer St, Fishermens Bend, Vic. 3207 AUSTRALIA [EMAIL PROTECTED] +61 3 9626 7811 +61 3 9626 7085 --- -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
setup.exe 2.249.2.5 problems
Hi, I am trying a fresh install of a freshly downloaded Cygwin install (setup.exe version 2.249.2.5) on an WinXP box. However, when I get to the step after Select your Internet Connection the setup program pops up the Can't get a list of download sites dialog box. other relevant data: my username has no spaces in it. and I was able to reproduce this on 2 different WinXP machines...I checked the FAQ but no luck. I can load the URL (http://sources.redhat.com/cygwin/mirrors.lst) with IE. There was a new release yesterday ...so maybe I was the first to find this buglet. Here is setup.log.full (As you can see I tried IE5 and direct settings) 2003/01/24 08:55:23 Starting cygwin install, version 2.249.2.5 2003/01/24 08:55:23 Current Directory: C:\Documents and Settings\shane\Desktop\Downloads\cygwin 2003/01/24 08:55:25 source: network install 2003/01/24 08:55:26 root: C:\cygwin binary system 2003/01/24 08:55:27 Selected local directory: C:\Documents and Settings\shane\Desktop\Downloads\cygwin 2003/01/24 08:55:28 net: Direct get_url_to_membuf http://sources.redhat.com/cygwin/mirrors.lst getUrlToStream http://sources.redhat.com/cygwin/mirrors.lst getUrlToStream failed! get_url_to_membuf failed! 2003/01/24 08:55:51 net: IE5 get_url_to_membuf http://sources.redhat.com/cygwin/mirrors.lst getUrlToStream http://sources.redhat.com/cygwin/mirrors.lst getUrlToStream failed! get_url_to_membuf failed! 2003/01/24 08:56:01 source: download 2003/01/24 08:56:02 Selected local directory: C:\Documents and Settings\shane\Desktop\Downloads\cygwin 2003/01/24 08:56:05 net: Direct get_url_to_membuf http://sources.redhat.com/cygwin/mirrors.lst getUrlToStream http://sources.redhat.com/cygwin/mirrors.lst getUrlToStream failed! get_url_to_membuf failed! 2003/01/24 08:56:10 Ending cygwin install Thanks, Shane -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: [ATTN: Harold] RE: Problem with keymap under cygwin-xfree
Ok, the keys i pressed, in order: the 1 (on the top row, NOT on the number pad) the h the numlock, twice the 2 (on the top row, NOT on the number pad) the , as you can see, the 1 and the h are interpreted as the same key, as is the 2 and the , This behaviour holds true for the rest of the number keys across the top row , the dash, the equal sign, the backspace key and the tab key. To the best of my knowledge, all other keys are working correctly. Attached is the output of keytest -Original Message- From: Harold L Hunt II [mailto:huntharo;pilot.msu.edu] Sent: Friday, November 08, 2002 3:35 PM To: [EMAIL PROTECTED] Subject: RE: [ATTN: Harold] RE: Problem with keymap under cygwin-xfree Okay, run this keyboard tester, press some keys, such as H, Num Lock, Num 1, etc. and send in the results to the mailing list. http://www.msu.edu/~huntharo/xwin/keytest.exe.bz2 (6 KiB) Harold Hunt keytest.log Description: Binary data
RE: Problem with keymap under cygwin-xfree
Thank you. I have made sure NumLock And CapsLock are OFF. I have attached the output of xev when I strike the keys 0-9, backspace, -, = and tab. I included some of the standard letter keys which do, in fact, work. Also attached is /tmp/Xwin.log. One thing I did try in all of this was ensuring I had a valid 104 key us pc .Xmodmap in ~. that produced exactly the same results I am seeing now. Thanks. -Original Message- From: Thomas Chadwick [mailto:j_tetazoo;hotmail.com] Sent: Wednesday, November 06, 2002 6:12 PM To: [EMAIL PROTECTED] Subject: Re: Problem with keymap under cygwin-xfree Make sure NumLock and CapsLock are both OFF. That can sometimes cause problem with certain X clients. If that's not the culprit, try running xev and see if X is even getting the keypress events. If xev spews a bunch of messages when pressing one of the keys in questions, cut-and-paste them to the list and we'll have a look-see. One other thing, post the contents of /tmp/XWin.log. From: Harris, Shane [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: '[EMAIL PROTECTED]' [EMAIL PROTECTED] Subject: Problem with keymap under cygwin-xfree Date: Wed, 6 Nov 2002 15:35:37 -0800 I have looked all over the list archive for the answer to this, with no results. So I am going to ask. Apologies if this has been covered somewhere before. I am having a maddening problem with xfree86 not recognizing some keys on my keyboard. And what is odd about this is that they are just the number keys (not the number pad), the backspace key and the tab key that appear to not be recognized. These keys work in other win32 applications, as well as the cygwin bash shell. I am at a total loss on this. If anyone can help, it would be much appreciated. _ Add photos to your messages with MSN 8. Get 2 months FREE*. http://join.msn.com/?page=features/featuredemail XWin.log Description: Binary data xev.log Description: Binary data
RE: Problem with keymap under cygwin-xfree
Device Manager identifies it as a Standard 101/102-Key. The problem with the workaround is that pressing the '1' key reports as keycode 43 and generates an h. Pressing an 'h' also reports as keycode 43 and generates an h. Hence, if I remap keycode 43, I now have a 1, but no h. This is so freakin mystifying. Why would two keys report the same keycode when pressed?? -Original Message- From: Thomas Chadwick [mailto:j_tetazoo;hotmail.com] Sent: Thursday, November 07, 2002 12:22 PM To: [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: RE: Problem with keymap under cygwin-xfree Looking at XWin.log, I can tell you're using an XF86Config file. However, your settings seem to match mine, which is good, I think, since I don't use an XF86Config file at all (resulting in defaults). Since xev is telling us that the X server is at least getting the keystrokes, a work-around would be to systematically figure out the keycodes using xev, and then add them to your .Xmodmap file. For instance, if xev reports that pressing the 1 key gives you keycode 200, then put keycode 200 = 1 exclam in .Xmodmap. As for the root cause? I'm at a loss to explain it. However, I'm not exactly an expert on the subject, so hopefully someone else on the list may be able to figure it out. One other thought: Maybe your keyboard isn't a standard 101/102-key keyboard? What does the Windows Device Manager identify your keyboard as? From: Harris, Shane [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: '[EMAIL PROTECTED]' [EMAIL PROTECTED] Subject: RE: Problem with keymap under cygwin-xfree Date: Thu, 7 Nov 2002 08:41:22 -0800 Thank you. I have made sure NumLock And CapsLock are OFF. I have attached the output of xev when I strike the keys 0-9, backspace, -, = and tab. I included some of the standard letter keys which do, in fact, work. Also attached is /tmp/Xwin.log. One thing I did try in all of this was ensuring I had a valid 104 key us pc .Xmodmap in ~. that produced exactly the same results I am seeing now. Thanks. -Original Message- From: Thomas Chadwick [mailto:j_tetazoo;hotmail.com] Sent: Wednesday, November 06, 2002 6:12 PM To: [EMAIL PROTECTED] Subject: Re: Problem with keymap under cygwin-xfree Make sure NumLock and CapsLock are both OFF. That can sometimes cause problem with certain X clients. If that's not the culprit, try running xev and see if X is even getting the keypress events. If xev spews a bunch of messages when pressing one of the keys in questions, cut-and-paste them to the list and we'll have a look-see. One other thing, post the contents of /tmp/XWin.log. From: Harris, Shane [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: '[EMAIL PROTECTED]' [EMAIL PROTECTED] Subject: Problem with keymap under cygwin-xfree Date: Wed, 6 Nov 2002 15:35:37 -0800 I have looked all over the list archive for the answer to this, with no results. So I am going to ask. Apologies if this has been covered somewhere before. I am having a maddening problem with xfree86 not recognizing some keys on my keyboard. And what is odd about this is that they are just the number keys (not the number pad), the backspace key and the tab key that appear to not be recognized. These keys work in other win32 applications, as well as the cygwin bash shell. I am at a total loss on this. If anyone can help, it would be much appreciated. _ Add photos to your messages with MSN 8. Get 2 months FREE*. http://join.msn.com/?page=features/featuredemail XWin.log xev.log _ Protect your PC - get McAfee.com VirusScan Online http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963
Problem with keymap under cygwin-xfree
I have looked all over the list archive for the answer to this, with no results. So I am going to ask. Apologies if this has been covered somewhere before. I am having a maddening problem with xfree86 not recognizing some keys on my keyboard. And what is odd about this is that they are just the number keys (not the number pad), the backspace key and the tab key that appear to not be recognized. These keys work in other win32 applications, as well as the cygwin bash shell. I am at a total loss on this. If anyone can help, it would be much appreciated.
RE: MMap offset parameter failing with ENOMEM error
Sorry - I didn't mean to sound impatient, just wasn't sure if it was a problem or not. Thanks for the response, keep up the good work - it is appreciated. Cheers, Shane Shane Mann Software Engineer Phone: +61-7-3259-2223 LeadUp Software Pty LtdFax: +61-7-3259-2259 339 Coronation Drive, Email: [EMAIL PROTECTED] Milton, QLD, 4064 Web: http://www.leadup.com.au -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Corinna Vinschen Sent: Wednesday, 25 September 2002 0:21 To: [EMAIL PROTECTED] Subject: Re: MMap offset parameter failing with ENOMEM error On Fri, Sep 20, 2002 at 09:38:06AM +1000, Shane Mann wrote: Perhaps the wrong error is being returned for this case - I would have thought the following error code would be more appropriate: [ENXIO] Addresses in the range [off,off+len) are invalid for the object specified by fildes. The appropriate error code is EOVERFLOW. I've checked that in. Any ideas? I've checked in a fix. The evaluation of the length to map (in 64k chunks) was incorrect. Thanks for the error report, Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Developermailto:[EMAIL PROTECTED] Red Hat, Inc. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/ -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: MMap offset parameter failing with ENOMEM error
Hi, Has this issue been confirmed as a bug? Is it being looked into? Cheers, Shane Shane Mann Software Engineer Phone: +61-7-3259-2223 LeadUp Software Pty LtdFax: +61-7-3259-2259 339 Coronation Drive, Email: [EMAIL PROTECTED] Milton, QLD, 4064 Web: http://www.leadup.com.au -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Shane Mann Sent: Friday, 20 September 2002 9:38 To: [EMAIL PROTECTED] Subject: RE: MMap offset parameter failing with ENOMEM error Corinna, I noticed this while I continued testing last night. Interestingly Linux and SunOS don't seem to be bothered by this, however, if you read the IEEE standard: http://www.opengroup.org/onlinepubs/007904975/functions/mmap.html there is the following quote: The range of bytes starting at off and continuing for len bytes shall be legitimate for the possible (not necessarily current) offsets in the file, shared memory object, or [TYM] typed memory object represented by fildes. Linux and Sun seem to allow it and return '0' if you access memory beyond the EOF, go figure. Perhaps the wrong error is being returned for this case - I would have thought the following error code would be more appropriate: [ENXIO] Addresses in the range [off,off+len) are invalid for the object specified by fildes. I have since amended the code by replacing the following line: size = finfo.st_size; with size = finfo.st_size - foff; and I still get the same error. There is some example output below: Note: index.idx is a file with a size of 557592 bytes. In the output below it worked for an offset of 40 and created a map of 157592 bytes - but failed for the same file with an offset of 30 and a potential map of 257592 bytes. Run 1) ~/c./a.exe 40 index.idx Offset is 40 File index.idx opened on descriptor 3. It is 157592 bytes long --- Syspagesize 4096, pagemask F000, offmask FFF foff is 40 --- Syspagesize 4096, pagemask F000, offmask FFF pfoff is 397312 Vars just before mmap call psyz = 163840 prot = 1 fd = 3 pfoff = 397312 Mapping 163840 at 2A231000 --- 397312 (mapping fh Result from femmap is 2A231A80 Byte is 127 Run 2) ~/c./a.exe 30 index.idx Offset is 30 File index.idx opened on descriptor 3. It is 257592 bytes long --- Syspagesize 4096, pagemask F000, offmask FFF foff is 30 --- Syspagesize 4096, pagemask F000, offmask FFF pfoff is 299008 Vars just before mmap call psyz = 262144 prot = 1 fd = 3 pfoff = 299008 perror reports errno of 12 Mapping 262144 at --- 299008 (mapping fh Any ideas? Cheers, Shane Shane Mann Software Engineer Phone: +61-7-3259-2223 LeadUp Software Pty LtdFax: +61-7-3259-2259 339 Coronation Drive, Email: [EMAIL PROTECTED] Milton, QLD, 4064 Web: http://www.leadup.com.au -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Corinna Vinschen Sent: Thursday, 19 September 2002 17:33 To: Cygwin@Cygwin. Com Subject: Re: MMap offset parameter failing with ENOMEM error On Thu, Sep 19, 2002 at 12:19:06PM +1000, Shane Mann wrote: Hi, I am having trouble with mmap using the last 'offset' parameter. I have attached a small piece of code (and sample file) which demonstrates the problem. Basically if the offset parameter is passed as a variable to mmap then the call fails with an ENOMEM error. But if you pass 0 as the parameter and then access the return address from mmap + offset - it is fine. My question: is mmap handling the offset parameter correctly? Usage for the program: ./a.exe offset file eg: ./a.exe 11200 index.idx To see the address + offset work, comment out the first mmap call and return statement and uncomment the lines below each one. Any help appreciated. Your application tries to mmap over EOF. Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Developermailto:[EMAIL PROTECTED] Red Hat, Inc. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/ -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/ -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: MMap offset parameter failing with ENOMEM error
Corinna, I noticed this while I continued testing last night. Interestingly Linux and SunOS don't seem to be bothered by this, however, if you read the IEEE standard: http://www.opengroup.org/onlinepubs/007904975/functions/mmap.html there is the following quote: The range of bytes starting at off and continuing for len bytes shall be legitimate for the possible (not necessarily current) offsets in the file, shared memory object, or [TYM] typed memory object represented by fildes. Linux and Sun seem to allow it and return '0' if you access memory beyond the EOF, go figure. Perhaps the wrong error is being returned for this case - I would have thought the following error code would be more appropriate: [ENXIO] Addresses in the range [off,off+len) are invalid for the object specified by fildes. I have since amended the code by replacing the following line: size = finfo.st_size; with size = finfo.st_size - foff; and I still get the same error. There is some example output below: Note: index.idx is a file with a size of 557592 bytes. In the output below it worked for an offset of 40 and created a map of 157592 bytes - but failed for the same file with an offset of 30 and a potential map of 257592 bytes. Run 1) ~/c./a.exe 40 index.idx Offset is 40 File index.idx opened on descriptor 3. It is 157592 bytes long --- Syspagesize 4096, pagemask F000, offmask FFF foff is 40 --- Syspagesize 4096, pagemask F000, offmask FFF pfoff is 397312 Vars just before mmap call psyz = 163840 prot = 1 fd = 3 pfoff = 397312 Mapping 163840 at 2A231000 --- 397312 (mapping fh Result from femmap is 2A231A80 Byte is 127 Run 2) ~/c./a.exe 30 index.idx Offset is 30 File index.idx opened on descriptor 3. It is 257592 bytes long --- Syspagesize 4096, pagemask F000, offmask FFF foff is 30 --- Syspagesize 4096, pagemask F000, offmask FFF pfoff is 299008 Vars just before mmap call psyz = 262144 prot = 1 fd = 3 pfoff = 299008 perror reports errno of 12 Mapping 262144 at --- 299008 (mapping fh Any ideas? Cheers, Shane Shane Mann Software Engineer Phone: +61-7-3259-2223 LeadUp Software Pty LtdFax: +61-7-3259-2259 339 Coronation Drive, Email: [EMAIL PROTECTED] Milton, QLD, 4064 Web: http://www.leadup.com.au -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Corinna Vinschen Sent: Thursday, 19 September 2002 17:33 To: Cygwin@Cygwin. Com Subject: Re: MMap offset parameter failing with ENOMEM error On Thu, Sep 19, 2002 at 12:19:06PM +1000, Shane Mann wrote: Hi, I am having trouble with mmap using the last 'offset' parameter. I have attached a small piece of code (and sample file) which demonstrates the problem. Basically if the offset parameter is passed as a variable to mmap then the call fails with an ENOMEM error. But if you pass 0 as the parameter and then access the return address from mmap + offset - it is fine. My question: is mmap handling the offset parameter correctly? Usage for the program: ./a.exe offset file eg: ./a.exe 11200 index.idx To see the address + offset work, comment out the first mmap call and return statement and uncomment the lines below each one. Any help appreciated. Your application tries to mmap over EOF. Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Developermailto:[EMAIL PROTECTED] Red Hat, Inc. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/ -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/