cygbuild as build script [Was: Re: [Review - Not yet] rdesktop]
Charles Wilson wrote: Harold L Hunt II wrote: 4) The script builds in the same directory as the source code. I don't like that. I have not yet (I hope) released a package that does this and I don't give good reviews to packages that do it. If the package fails to build out of the source directory, than you can either use lndir or the similar but more portable suggestion that Chuck Wilson had a few weeks/months ago (don't remember what it was called). If they cygbuild script doesn't yet support this, then it isn't ready for prime-time and you should hack up one of the scripts from another source package to do the job in the mean time. Disagree. It is NOT a requirement that you must be able to build a package outside of the source dir. *I* find it convenient to do so, solely because it makes it easier to generate a patch without build-generated files making a mess of it. But that's an issue for the maintainer; if Jari has another system that works for him, let's not put barriers in his way. Jari's package: (1) distributes source (2) distributes a patch (3) contains a build script. Except for the naming issues and worries about clobbering that you raise, his system looks like it satisfies the spirit behind our packaging guidelines. That's enough for me. If cygbuild is going to ever be promoted as a build script, then it should continue to support building out of the source dir, whether it requires a flag (such as defaulting to not needing a hack and having a flag that tells it to us lndir or a similar hack) or not. If cygbuild isn't going to have that feature, then I will continue reviewing up to the point of everything being fixed expect for my niggle above. I won't give such a package a Good to go. That is my choice. Somebody else can pick it up and give it a Good to go at that point if they want to, but my choice is not to do so. Harold
Re: [Review - Not yet] libsmi
Abe, I have an interesting update for the libsmi package. I got it to build shared libraries and it passes 19 of the 20 tests. I am very close to having it build out of the source directory... the one remaining step is to get the mibs to install correctly. I should have an updated -src package for you soon that you can take over. Harold
Re: [Review - Not yet] libsmi
Abe, Here are some things I did: 1) Used a standard build script. 2) Passed -no-undefined on the LDFLAGS to allow the DLL to be built. 3) Ran 'make check' to see that only 1 of the 20 test failed (the last one, seems to be something trivial). 4) Relibtoolized the source to allow the DLL to be built. 5) Fixed mibs/Makefile.am and mibs/*/Makefile.am to allow the 'make install' command to properly install the mibs when srcdir != builddir. 6) Updated the libsmi.README file to list proper build requirements. 7) Updated the setup.hint file to list proper run-time requirements. 8) Rolled the libsmi, libsmi-devel, and libsmi-tools packages back into one package (this was really the default from using a generic build script). The libsmi package was only containing the readme files and not much else (it was around 30 KiB). Now that we have a DLL we could change this to be: libsmi (docs, exes, mibs, etc.), libsmi2 (contains the DLL), and libsmi-devel (includes and link libs). I have some previous packages that I have done this for and I could do it for this package as well (the process consists of copying a few lines into the build script and tweaking a few other lines). Here are my files: cut here -- #!/bin/bash wget \ http://msu.edu/~huntharo/cygwin/release/libsmi/libsmi-0.4.1-1.tar.bz2 wget \ http://msu.edu/~huntharo/cygwin/release/libsmi/libsmi-0.4.1-1-src.tar.bz2 cut here -- I think these changes really improve the package, but I admit that it is not 100% complete yet. Feel free to do with these changes as you want to. Harold
RE: [RFC] Globally creating a user and a group root
Corinna Vinschen wrote: John, On Wed, Nov 12, 2003 at 03:44:44PM +0100, Corinna Vinschen wrote: On Wed, Nov 12, 2003 at 12:29:08PM -, Morrison, John wrote: Corinna Vinschen wrote: Eh, no. I was thinking about a postinstall script, which checks for Everyone (S-1-0-0) in /etc/passwd and /etc/group and removes these entries silently. The same script could add the root entry to /etc/group. But then what happens if somebody messes up there /etc/[passwd|group] and have to create it again? Does mk[passwd|group] do the changes automatically? mkpasswd and mkgroup don't create an Everyone entry since August 2002. The remaining problem is the special handling of root. I guess it's best to create the /etc/group entry already in mkgroup. For mkpasswd it might be best, to add a special handling like this: If a Windows user root exists, give it the uid 0. any chance you can poke the base-passwd script soon, to check for a user and a group with SID S-1-1-0 in the existing /etc/passwd and /etc/group files and remove them silently? Also it would be good if the script adds the following entry to /etc/group, if possible as the first line: root:S-1-5-32-544:0: OK, just a few questions: 1) does your script do all this? 2) does the script need to be run after passwd-grp.sh? 3) will your script run OK as a postinstall? (I know it prompts for information...) 4) I get an Error in addUserRights (LsaAddAccountRights returned 0xc060=STATUS_NO_SUCH_PRIVILEGE)! on a w2k box (I have full, local, admin rights. Is this OK? (or have I lost the plot again :| :) J. == Information in this email and any attachments are confidential, and may not be copied or used by anyone other than the addressee, nor disclosed to any third party without our permission. There is no intention to create any legally binding contract or other binding commitment through the use of this electronic communication unless it is issued in accordance with the Experian Limited standard terms and conditions of purchase or other express written agreement between Experian Limited and the recipient Experian Limited (registration number 653331) Registered office: Talbot House, Talbot Street, Nottingham NG80 1TH
Re: [RFC] Globally creating a user and a group root
On Thu, Nov 27, 2003 at 08:33:24AM -, Morrison, John wrote: Corinna Vinschen wrote: any chance you can poke the base-passwd script soon, to check for a user and a group with SID S-1-1-0 in the existing /etc/passwd and /etc/group files and remove them silently? Also it would be good if the script adds the following entry to /etc/group, if possible as the first line: root:S-1-5-32-544:0: OK, just a few questions: 1) does your script do all this? No. I was asking you to add the above to the passwd related postinstall script. That has nothing to do with my create-root script. 2) does the script need to be run after passwd-grp.sh? Which script? Your's should run on postinstall. Removing S-1-1-0 just removes the Everyone entries which are old and superfluous anyway. I've changed mkgroup to add a root entry as above when called with -l but this only helps if mkgroup is called at all. Your script should make sure that such an entry exists on machines which already have an /etc/group file. 3) will your script run OK as a postinstall? (I know it prompts for information...) No. I was just asking you to do the above in a postinstall script. This is independent of my create-root script. 4) I get an Error in addUserRights (LsaAddAccountRights returned 0xc060=STATUS_NO_SUCH_PRIVILEGE)! on a w2k box (I have full, local, admin rights. Is this OK? (or have I lost the plot again :| Details? Which user right does result in that error? Does W2K not have the SeDenyXXX rights, perhaps? Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Developermailto:[EMAIL PROTECTED] Red Hat, Inc.
Re: [RFC] Globally creating a user and a group root
On Thu, Nov 27, 2003 at 11:14:29AM +0800, Dean Scarff wrote: Corinna Vinschen wrote: I currently have an entry in /etc/passwd: root:unused_by_nt/2000/xp:500:513:U-SCARFF-BOX\root,S-1-5-21-1606980848-1563985344-1060284298-500:/home/root:/bin/bash This is a side-effect of using mkpasswd with the Administrator user renamed to root on XP. For me (and others like me who renamed the Administrator user) will this cause problems? Forgive me if I've Not the above adding of the admins group as root to /etc/group. But it is also the intend to create a root user which has the necessary rights to do all the ugly stuff like LocalSystem. I'd appreciate if you would read the whole thread again and then to discuss solution of that problem. overlooked where this point was addressed before in the thread, I only followed it loosely. Too bad. Why am I trying to discuss this stuff in public? Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Developermailto:[EMAIL PROTECTED] Red Hat, Inc.
Re: HEADSUP: cygserver now has MSG, SEM and SHM support
On Wed, Nov 26, 2003 at 10:38:29PM -0500, Harold L Hunt II wrote: Christopher Faylor wrote: The gotcha here, though, is that you still have to install cygserver and get it running. Corinna has provided install scripts to make that to happen but experience shows us that there will probably be some tweaking required before everything is perfect. You just have the same gotcha with cygipc. So nothing really new with this. Ah, well, in the case of XWin.exe, it detects whether the an SHM allocation succeeds or fails and enables/disables the X SHM extension appropriately. Each X application tests for presence of the SHM extension and doesn't try to use it if it is not there. Thus, all of our apps and libs link against cygipc. We can still build with SHM support but it won't actually be used unless things are setup properly, as you described. Ok, I don't know how that's implemented in cygipc so I think it might be necessary to mention that: The implementation of the MSG, SEM and SHM functions in Cygwin are so that if the functions are not available (be it that CYGWIN doesn't contain the word server or Cygserver isn't running) not only set errno to ENOSYS, but they also raise a SIGSYS. So it's crucial that your code handles that signal. At least, it should ignore it (signal(SISYS, SIG_IGN)). Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Developermailto:[EMAIL PROTECTED] Red Hat, Inc.
RE: [RFC] Globally creating a user and a group root
Corinna Vinschen wrote: On Thu, Nov 27, 2003 at 08:33:24AM -, Morrison, John wrote: Corinna Vinschen wrote: any chance you can poke the base-passwd script soon, to check for a user and a group with SID S-1-1-0 in the existing /etc/passwd and /etc/group files and remove them silently? Also it would be good if the script adds the following entry to /etc/group, if possible as the first line: root:S-1-5-32-544:0: OK, just a few questions: 1) does your script do all this? No. I was asking you to add the above to the passwd related postinstall script. That has nothing to do with my create-root script. Sorry I thought it was your create-root.sh script we were talking about. 2) does the script need to be run after passwd-grp.sh? Which script? Your's should run on postinstall. Removing S-1-1-0 err, the one refered to in point 1? just removes the Everyone entries which are old and superfluous anyway. I've changed mkgroup to add a root entry as above when called with -l but this only helps if mkgroup is called at all. Your script should make sure that such an entry exists on machines which already have an /etc/group file. 3) will your script run OK as a postinstall? (I know it prompts for information...) No. I was just asking you to do the above in a postinstall script. This is independent of my create-root script. 4) I get an Error in addUserRights (LsaAddAccountRights returned 0xc060=STATUS_NO_SUCH_PRIVILEGE)! on a w2k box (I have full, local, admin rights. Is this OK? (or have I lost the plot again :| Details? Which user right does result in that error? Does W2K not have the SeDenyXXX rights, perhaps? Sorry, you are talking double dutch (sorry all you Dutch ;). All I did was run the create-root.sh and enter a password. I've not got the time atm to go any deeper - I've a non-flexable deadline at work that I'm up against :( Having said that, I'll try and scavenge some time at the weekend if not before and try and do the mods. J. == Information in this email and any attachments are confidential, and may not be copied or used by anyone other than the addressee, nor disclosed to any third party without our permission. There is no intention to create any legally binding contract or other binding commitment through the use of this electronic communication unless it is issued in accordance with the Experian Limited standard terms and conditions of purchase or other express written agreement between Experian Limited and the recipient Experian Limited (registration number 653331) Registered office: Talbot House, Talbot Street, Nottingham NG80 1TH
RE: [RFC] Globally creating a user and a group root
Corinna Vinschen wrote: On Thu, Nov 27, 2003 at 09:05:02AM -, Morrison, John wrote: Corinna Vinschen wrote: On Thu, Nov 27, 2003 at 08:33:24AM -, Morrison, John wrote: Corinna Vinschen wrote: any chance you can poke the base-passwd script soon, to check for a user and a group with SID S-1-1-0 in the existing /etc/passwd and /etc/group files and remove them silently? Also it would be good if the script adds the following entry to /etc/group, if possible as the first line: root:S-1-5-32-544:0: OK, just a few questions: 1) does your script do all this? No. I was asking you to add the above to the passwd related postinstall script. That has nothing to do with my create-root script. Sorry I thought it was your create-root.sh script we were talking about. No, my create-root script is an entirely different - later - step. I was just asking you this: You already have this base-passwd and other general scripts running on postinstall. What we need is, having these stone age old Everyone entries with uid and gid 0 and SID S-1-1-0 removed from /etc/passwd and /etc/group. So I'm asking you to add something to your postinstall magic, which does that trick. Just removing these entries once and for all, not even asking the user for anything. And the second wish is, to create a root:S-1-5-32-544:0: entry to /etc/group, if /etc/group already exists and doesn't contain such a root entry already. If there's no /etc/group, your script will create one anyway, and in future, already mkgroup will create the above root entry. That's it. Your script is just one step on the way to the root user and group we were talking about to get the service problems solved. The create-root script is another step, which has nothing to do with what I'm asking you for. Did I describe that clear enough this time? ... OK, chill :)) I'll try for the weekend. 4) I get an Error in addUserRights (LsaAddAccountRights returned 0xc060=STATUS_NO_SUCH_PRIVILEGE)! on a w2k box (I have full, local, admin rights. Is this OK? (or have I lost the plot again :| Details? Which user right does result in that error? Does W2K not have the SeDenyXXX rights, perhaps? Sorry, you are talking double dutch (sorry all you Dutch ;). All Just look into the create-root script. There are seven calls to the famous new editrights tool. Three of them are only called on machines with NT5 (W2K) and up. Probably that's not ok. I don't know exactly but it's possible that W2K doesn't have these SeDenyWhatever user rights. editrights -a SeDenyRemoteInteractiveLogonRight -u root was the one that failed. Would it be best (assuming these last 3 are optional) to 2 /dev/null them? Is there any other info that would help? J. == Information in this email and any attachments are confidential, and may not be copied or used by anyone other than the addressee, nor disclosed to any third party without our permission. There is no intention to create any legally binding contract or other binding commitment through the use of this electronic communication unless it is issued in accordance with the Experian Limited standard terms and conditions of purchase or other express written agreement between Experian Limited and the recipient Experian Limited (registration number 653331) Registered office: Talbot House, Talbot Street, Nottingham NG80 1TH
Re: [RFC] Globally creating a user and a group root
On Thu, Nov 27, 2003 at 09:58:52AM -, Morrison, John wrote: Corinna Vinschen wrote: I don't know exactly but it's possible that W2K doesn't have these SeDenyWhatever user rights. editrights -a SeDenyRemoteInteractiveLogonRight -u root was the one that failed. Would it be best (assuming these last 3 are optional) to 2 /dev/null them? Is there any other info that would help? Er... only this one? Hmm, that makes sense, sort of. Look like this right only exist since XP. I'm reluctant to devnull them. There might be a real error covered when doing this. I'll better add another variable to check if running under XP or newer. Thanks, Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Developermailto:[EMAIL PROTECTED] Red Hat, Inc.
Re: [RFC] Globally creating a user and a group root
On Thu, Nov 27, 2003 at 11:24:55AM +0100, Corinna Vinschen wrote: On Thu, Nov 27, 2003 at 09:58:52AM -, Morrison, John wrote: Corinna Vinschen wrote: I don't know exactly but it's possible that W2K doesn't have these SeDenyWhatever user rights. editrights -a SeDenyRemoteInteractiveLogonRight -u root was the one that failed. Would it be best (assuming these last 3 are optional) to 2 /dev/null them? Is there any other info that would help? Er... only this one? Hmm, that makes sense, sort of. Look like this right only exist since XP. I'm reluctant to devnull them. There might be a real error covered when doing this. I'll better add another variable to check if running under XP or newer. ping! ping! ping! New create-root script attached! This script should do it right now. Additionally I changed the way how the password is read to using the `read -s' option to hide what's actually typed in. Can you give this script another test? Thanks, Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Developermailto:[EMAIL PROTECTED] Red Hat, Inc. #!/bin/bash # # create-root, Copyright 2003 Red Hat Inc. # # This file is part of Cygwin. # Subdirectory where the new package is being installed PREFIX=/usr # Directory where the config files are stored SYSCONFDIR=/etc progname=$0 auto_answer= port_number=22 request() { if [ ${auto_answer} = yes ] then echo $1 (yes/no) yes return 0 elif [ ${auto_answer} = no ] then echo $1 (yes/no) no return 1 fi answer= while [ X${answer} != Xyes -a X${answer} != Xno ] do echo -n $1 (yes/no) read -e answer done if [ X${answer} = Xyes ] then return 0 else return 1 fi } # Check options while : do case $# in 0) break ;; esac option=$1 shift case ${option} in -d | --debug ) set -x ;; -y | --yes ) auto_answer=yes ;; -n | --no ) auto_answer=no ;; -c | --cygwin ) cygwin_value=$1 shift ;; -p | --port ) port_number=$1 shift ;; -w | --pwd ) password_value=$1 shift ;; *) echo usage: ${progname} [OPTION]... echo echo This script creates a root user which has appropriate privileges echo to run services which need to switch user context without password. echo echo Options: echo --debug -dEnable shell's debug output. echo --yes-yAnswer all questions with \yes\ automatically. echo --no -nAnswer all questions with \no\ automatically. echo --pwd-w passwd Use \pwd\ as password for user 'root'. echo exit 1 ;; esac done # Check if running on NT uname | grep -q CYGWIN_NT _nt=yes # If not running on NT, nothing to do if [ ${_nt} != yes ] then echo Nothing to do on 9x/Me. exit 0 fi # Check if running under NT5 or later _nt5=`uname | awk -F- '{print ( $2 = 5.0 ) ? yes : no;}'` # Check if running under NT5.1 or later _nt5_1=`uname | awk -F- '{print ( $2 5.0 ) ? yes : no;}'` # Check for ${SYSCONFDIR} directory if [ -e ${SYSCONFDIR} -a ! -d ${SYSCONFDIR} ] then echo echo ${SYSCONFDIR} is existant but not a directory. echo Cannot create global configuration files. echo exit 1 fi # Create it if necessary if [ ! -e ${SYSCONFDIR} ] then mkdir ${SYSCONFDIR} if [ ! -e ${SYSCONFDIR} ] then echo echo Creating ${SYSCONFDIR} directory failed echo exit 1 fi fi echo echo echo Warning: The following functions require administrator privileges! echo # Check if a user root is already in /etc/passwd. grep -q '^root:' ${SYSCONFDIR}/passwd root_in_passwd=yes # Drop root from passwd since it could have wrong settings. if [ ${root_in_passwd} = yes ] then grep -v '^root:' ${SYSCONFDIR}/passwd ${SYSCONFDIR}/passwd.$$ rm -f ${SYSCONFDIR}/passwd mv ${SYSCONFDIR}/passwd.$$ ${SYSCONFDIR}/passwd chmod g-w,o-w ${SYSCONFDIR}/passwd fi # Drop root from group file since it could have wrong settings. # Inject a correct entry instead. echo 'root:S-1-5-32-544:0:' ${SYSCONFDIR}/group.$$ grep -v '^root:' ${SYSCONFDIR}/group ${SYSCONFDIR}/group.$$ rm -f ${SYSCONFDIR}/group mv ${SYSCONFDIR}/group.$$ ${SYSCONFDIR}/group chmod g-w,o-w ${SYSCONFDIR}/group # Get local name of administrators group. _admingroup=`mkgroup -l | awk -F: '/:544:/{print $1;}'` # Check if a user root exists on the system. net user root /dev/null 21 root_in_sam=yes if [ ${root_in_sam} = yes ] then echo echo You already have a user 'root' on your system. Is that user supposed if ! request to be used as the required privileged user account? then echo echo In that case, this script can't do its job. Either rename the echo user called 'root' or create a 'root' entry in ${SYSCONFDIR}/passwd, echo using a user with appropriate privileges. exit 1 fi else # Create a local root user. mkdir -p
Re: Pending Packages List, 2003-11-26
On Wed, Nov 26, 2003 at 05:51:24PM -, Daniel Reed wrote: rdesktop 1.3.0-1 [2003-11-08] Description: client for Windows terminal server. Remote desktop display Proposer: Jari Aalto Proposal: mailto:[EMAIL PROTECTED] http://tierra.dyndns.org:81/cygwin/rdesktop/rdesktop-1.3.0-1.tar.bz2 http://tierra.dyndns.org:81/cygwin/rdesktop/rdesktop-1.3.0-1-src.tar.bz2 [no hint] Aye votes: Harold L Hunt II (cygwin-apps-thread.12044) [1/3] Abe Backus (cygwin-apps-thread.12051) [2/3] Status: Package available. HOLD-UPS: Not enough votes (need 1 more). No good to go review. I vote for this. rlc
RE: [RFC] Globally creating a user and a group root
Corinna Vinschen wrote: On Thu, Nov 27, 2003 at 11:24:55AM +0100, Corinna Vinschen wrote: On Thu, Nov 27, 2003 at 09:58:52AM -, Morrison, John wrote: Corinna Vinschen wrote: I don't know exactly but it's possible that W2K doesn't have these SeDenyWhatever user rights. editrights -a SeDenyRemoteInteractiveLogonRight -u root was the one that failed. Would it be best (assuming these last 3 are optional) to 2 /dev/null them? Is there any other info that would help? Er... only this one? Hmm, that makes sense, sort of. Look like this right only exist since XP. I'm reluctant to devnull them. There might be a real error covered when doing this. I'll better add another variable to check if running under XP or newer. OK :) ping! ping! ping! New create-root script attached! This script should do it right now. Additionally I changed the way how the password is read to using the `read -s' option to hide what's actually typed in. But you display it in the next echo... User 'root' has been created with password ''. Can you give this script another test? Other than that it works fine :) Shame theres no way of setting the password to never expire :( J. == Information in this email and any attachments are confidential, and may not be copied or used by anyone other than the addressee, nor disclosed to any third party without our permission. There is no intention to create any legally binding contract or other binding commitment through the use of this electronic communication unless it is issued in accordance with the Experian Limited standard terms and conditions of purchase or other express written agreement between Experian Limited and the recipient Experian Limited (registration number 653331) Registered office: Talbot House, Talbot Street, Nottingham NG80 1TH
Re: [RFC] Globally creating a user and a group root
On Thu, Nov 27, 2003 at 11:37:11AM -, Morrison, John wrote: Corinna Vinschen wrote: On Thu, Nov 27, 2003 at 11:24:55AM +0100, Corinna Vinschen wrote: On Thu, Nov 27, 2003 at 09:58:52AM -, Morrison, John wrote: Corinna Vinschen wrote: I don't know exactly but it's possible that W2K doesn't have these SeDenyWhatever user rights. editrights -a SeDenyRemoteInteractiveLogonRight -u root was the one that failed. Would it be best (assuming these last 3 are optional) to 2 /dev/null them? Is there any other info that would help? Er... only this one? Hmm, that makes sense, sort of. Look like this right only exist since XP. I'm reluctant to devnull them. There might be a real error covered when doing this. I'll better add another variable to check if running under XP or newer. OK :) ping! ping! ping! New create-root script attached! This script should do it right now. Additionally I changed the way how the password is read to using the `read -s' option to hide what's actually typed in. But you display it in the next echo... Ouch! I'll fix that. Can you give this script another test? Other than that it works fine :) Shame theres no way of setting the password to never expire :( There is! The passwd tool in the current snapshots already have the option -e to switch expiry off. If you have another look into the script, you'll see that I check the version number of passwd, so I know if it already has that functionality or not ;-) Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Developermailto:[EMAIL PROTECTED] Red Hat, Inc.
Re: [RFC] Globally creating a user and a group root
On Thu, Nov 27, 2003 at 12:53:01PM -, Morrison, John wrote: Corinna Vinschen wrote: There is! The passwd tool in the current snapshots already have the option -e to switch expiry off. If you have another look into the script, you'll see that I check the version number of passwd, so I know if it already has that functionality or not ;-) Ahh, totally missed that. Would it be best to release the new passwd tool before this script is included? I'm not planning to do this in a rush. I'm trying to get this done step by step. Your script changes are actually the first step which will hit our users. I think we will release Cygwin 1.5.6 before I release the create-root script. Then we can switch over the services as we see fit. Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Developermailto:[EMAIL PROTECTED] Red Hat, Inc.
RE: [RFC] Globally creating a user and a group root
OK, took some lunch time ;) Hows this for passwd-grp.sh - #!/bin/sh if [ ! -e /etc/passwd -a ! -L /etc/passwd ] ; then /bin/mkpasswd -l -c /etc/passwd fi if [ ! -e /etc/group -a ! -L /etc/group ] ; then /bin/mkgroup -l -c /etc/group fi cp -f /etc/passwd ~/passwd.mkpasswd \ ( [ -w /etc/passwd ] || chmod --silent a+w /etc/passwd ; ) \ sed -e '/:S-1-1-0:/d' ~/passwd.passwd /etc/passwd \ chmod --silent --reference=/etc/group /etc/passwd rm -f ~/passwd.mkpasswd cp -f /etc/group ~/group.mkgroup \ ( [ -w /etc/group ] || chmod --silent a+w /etc/group ; ) \ echo root:S-1-5-32-544:0: /etc/group \ sed -e '/:S-1-1-0:/d' ~/group.mkgroup /etc/group \ chmod --silent --reference=/etc/passwd /etc/group rm -f ~/group.mkgroup - comments? (I tried to make it play nice with file permissions) J. == Information in this email and any attachments are confidential, and may not be copied or used by anyone other than the addressee, nor disclosed to any third party without our permission. There is no intention to create any legally binding contract or other binding commitment through the use of this electronic communication unless it is issued in accordance with the Experian Limited standard terms and conditions of purchase or other express written agreement between Experian Limited and the recipient Experian Limited (registration number 653331) Registered office: Talbot House, Talbot Street, Nottingham NG80 1TH passwd-grp.sh Description: Binary data
RE: [RFC] Globally creating a user and a group root
Corinna Vinschen wrote: On Thu, Nov 27, 2003 at 01:24:15PM -, Morrison, John wrote: OK, took some lunch time ;) Hows this for passwd-grp.sh - #!/bin/sh if [ ! -e /etc/passwd -a ! -L /etc/passwd ] ; then /bin/mkpasswd -l -c /etc/passwd fi if [ ! -e /etc/group -a ! -L /etc/group ] ; then /bin/mkgroup -l -c /etc/group fi cp -f /etc/passwd ~/passwd.mkpasswd \ ( [ -w /etc/passwd ] || chmod --silent a+w /etc/passwd ; ) \ sed -e '/:S-1-1-0:/d' ~/passwd.passwd /etc/passwd \ chmod --silent --reference=/etc/group /etc/passwd rm -f ~/passwd.mkpasswd cp -f /etc/group ~/group.mkgroup \ ( [ -w /etc/group ] || chmod --silent a+w /etc/group ; ) \ echo root:S-1-5-32-544:0: /etc/group \ sed -e '/:S-1-1-0:/d' ~/group.mkgroup /etc/group \ chmod --silent --reference=/etc/passwd /etc/group rm -f ~/group.mkgroup - comments? (I tried to make it play nice with file permissions) Cool! Just the `echo root:...' should only be called if the group file doesn't already contain that entry. As I said, the upcoming mkgroup will create this entry already (with -l). You should check its existence first. Would prefer... echo root:S-1-5-32-544:0: /etc/group \ sed -e '/:S-1-1-0:/d' -e '/root:S-1-5-32-544:0:/d' ~/group.mkgroup /etc/group \ 'cause then I don't need to check anything :) If you'd prefer me to check... help! :) J. == Information in this email and any attachments are confidential, and may not be copied or used by anyone other than the addressee, nor disclosed to any third party without our permission. There is no intention to create any legally binding contract or other binding commitment through the use of this electronic communication unless it is issued in accordance with the Experian Limited standard terms and conditions of purchase or other express written agreement between Experian Limited and the recipient Experian Limited (registration number 653331) Registered office: Talbot House, Talbot Street, Nottingham NG80 1TH
AW: HEADSUP: cygserver now has MSG, SEM and SHM support
Corinna Vinschen wrote: Ok, I don't know how that's implemented in cygipc so I think it might be necessary to mention that: The implementation of the MSG, SEM and SHM functions in Cygwin are so that if the functions are not available (be it that CYGWIN doesn't contain the word server or Cygserver isn't running) not only set errno to ENOSYS, but they also raise a SIGSYS. So it's crucial that your code handles that signal. At least, it should ignore it (signal(SISYS, SIG_IGN)). Harold for the Xserver that means, that the shm initialisation code has to be changed to the detecting code used by FreeBSD and friends (CheckForShmSyscall() in Xserver/Xext/shm.c and Xserver/Xext/xf86bigfont.c) The cygipc-related stuff could then be removed in both files. Ralf
RE: distcc - Awaiting another review, or waiting for fixes from first review?
From: Harold L Hunt II John, Have you, or are you going to, addressed the isses listed in the PPL below for distcc? Once those minor issues have been addressed, I will try to review the package for you. I don't want to review it right now only to report the same problems though. Thanks Harold, it's appreciated, but I'm pressed by a deadline at work and have (even less) time than usual. I *do* intend to address these, but prob not for at least another week or so :| J.
Re: [Review - Not yet] tcm
Hi Harold, 1) yes, my changes are very likely to be integrated into the original TCM distribution (most are in fact already included there). 2) I created a build script now and rebuilt the source and binary packages, available from the usual URLs (http://home.in.tum.de/~boesswet/tcm_cygwin.html) The latest official TCM source (2.20) already contains lots of my changes (#if'ed where necessary). The reason for patch to this source is mostly to make it conform to the cygwin packaging-rules (and some general bug fixes). The reason for the huge size of the patch (400k) was, that the original source is not clean (because some yacc-generated files are contained in the archive that are removed by make clean). I changed this manually and the new patch is only 93k BTW: The binary package has also been available from the TCM download page at http://wwwhome.cs.utwente.nl/~tcm/ since January and now makes approx. 30% of all binary downloads! Best Regards, Daniel Harold L Hunt II wrote: Daniel, I just reviewed your tcm package: Package: tcm 2.20-1 [2003-01-27] Description: Toolkit for Conceptual Modeling (TCM) Proposer: Daniel Boesswetter Proposal: http://sources.redhat.com/ml/cygwin-apps/2003-01/msg00299.html http://home.in.tum.de/~boesswet/tcm-2.20-1.tar.bz2 http://home.in.tum.de/~boesswet/tcm-2.20-1-src.tar.bz2 http://home.in.tum.de/~boesswet/setup.hint Good to go: Charles Wilson (cygwin-apps-thread.11851) (once problems are addressed) Problems: So, here's the question for the list. For the cygwin-specific README in a X-related package, where should it go? (cygwin-apps-thread.11851) Status: Attained required 3 votes. Package available. Reviewed. HOLD-UPS: Unresolved minor problems. Questions/Issues 1) You have made a lot of changes. Is there any chance that these patches will be excepted into the upstream tcm package? I'm a little nervous having this many changes in place that aren't properly #if'd since the next upstream release might not have all of this stuff and the packge will become unmaintained unless you are around to get those patches adapted for the next version. 2) There is no build script. I know that packages without build scripts are still allowed, and I used to make them myself since I thought they were really tough... but now they save me so much time that I won't release a package without them and I shy away from reviewing a package that doesn't have one. If you don't want to make the script, then I could make it for you (after you tell me what you are going to do regarding all the changes in #1). Awaiting your input, Harold
Re: Pending Packages List, 2003-11-26
) Package: joe 2.9.8-1 [2003-11-11] ) Description: Fast and simple editor which emulates 5 other editors )Proposer: Jari Aalto )Proposal: mailto:[EMAIL PROTECTED] ) http://tierra.dyndns.org:81/cygwin/joe/joe-2.9.8-1.tar.bz2 joe has now received enough votes for inclusion, and is only awaiting thorough review. Right off the bat, the binary package includes files in etc/ which may need to be installed by a postinstall script instead; usr/doc needs to become usr/share/doc and usr/man needs to become usr/share/man. -- Daniel Reed [EMAIL PROTECTED] http://naim-users.org/nmlorg/ http://naim.n.ml.org/ Real computer scientists like having a computer on their desk, else how could they read their mail?
Re: Pending Packages List, 2003-11-26
Daniel Reed wrote: ) Package: joe 2.9.8-1 [2003-11-11] ) Description: Fast and simple editor which emulates 5 other editors )Proposer: Jari Aalto )Proposal: mailto:[EMAIL PROTECTED] ) http://tierra.dyndns.org:81/cygwin/joe/joe-2.9.8-1.tar.bz2 joe has now received enough votes for inclusion, and is only awaiting thorough review. Right off the bat, the binary package includes files in etc/ which may need to be installed by a postinstall script instead; usr/doc needs to become usr/share/doc and usr/man needs to become usr/share/man. Also, as I pointed out below, there are two generic problems with the cygbuild script. One causes the build script name to keep getting another copy of the version string appended each time the 'spkg' command is run, while the other problem is that the build script is called just cygbuild-x.x.x-x.sh, which will cause problems when multiple source packages are extracted to /usr/src. The first email below is my generic review for several packages that have the same two problems as the last two emails describe in their first two problem entries. http://cygwin.com/ml/cygwin-apps/2003-11/msg00374.html http://cygwin.com/ml/cygwin-apps/2003-11/msg00373.html http://cygwin.com/ml/cygwin-apps/2003-11/msg00372.html Harold
distcc - Addressed all minor issues - Should be good to go (with John M's approval)
John, John Morrison wrote: From: Harold L Hunt II John, Have you, or are you going to, addressed the isses listed in the PPL below for distcc? Once those minor issues have been addressed, I will try to review the package for you. I don't want to review it right now only to report the same problems though. Thanks Harold, it's appreciated, but I'm pressed by a deadline at work and have (even less) time than usual. I *do* intend to address these, but prob not for at least another week or so :| I took care of the minor issues for you and added a generic patch for a --with-docdir parameter for the configure script (which I have submitted to the distcc mailing list). I created a distcc.README file that specifies you as the maintainer and added 'cygwin' and 'popt' as dependencies in the setup.hint. You are still the maintainer though. If you install this and it works, lets post it and you can handle all future maintenance. Oh yeah, I grabbed the 2.11.2 release instead of the 2.11.1 release that the initial package was for. cut here -- #!/bin/bash wget \ http://msu.edu/~huntharo/cygwin/release/distcc/distcc-2.11.2-1.tar.bz2 wget \ http://msu.edu/~huntharo/cygwin/release/distcc/distcc-2.11.2-1-src.tar.bz2 cut here -- You can point setup.exe to the following address to test my package and source package: http://msu.edu/~huntharo/cygwin/ If it works, lets get this posted. Harold
Re: [Review - Very minor issues] tcm
Harold L Hunt II wrote: 3) CYGWIN-PATCHES/tcm-2.20.README should be named CYWGIN-PATCHES/tcm.README. This isn't actually installed either, which requires that the install command in the build script be changed to the following: install() { (cd ${objdir} \ make install prefix=${instdir} \ if [ ! -d ${instdir}${prefix}/X11R6/share/doc/Cygwin ]; then \ mkdir -p ${instdir}${prefix}/X11R6/share/doc/Cygwin ; \ fi \ if [ ! -d ${instdir}${prefix}/X11R6/share/doc/${PKG}-${VER} ]; then \ mkdir -p ${instdir}${prefix}/X11R6/share/doc/${PKG}-${VER} ; \ fi \ /usr/bin/install -m 644 ${srcdir}/CYGWIN-PATCHES/${PKG}.README \ ${instdir}${prefix}/X11R6/share/doc/Cygwin/${PKG}-${VER}.README ) } Oh yeah, and this bit should be removed from the Makefile (no need to have it there... should be stuck in the distribution build script (as I showed above) ). Here is the line in the Makefile that should be removed: $(TCM_INSTALL_CMD) CYGWIN-PATCHES/tcm-2.20.README /usr/share/doc/Cygwin/ Harold
Re: [distcc] Patch for --with-docdir in configure.ac and Makefile.in
Martin, Martin Pool wrote: On 27 Nov 2003, Harold L Hunt II [EMAIL PROTECTED] wrote: The attached patch adds a --with-docdir=path flag for the configure script; the default when --with-docdir is not specified remains unchanged. The patch has been build tested on Cygwin. Thanks, taken. Cool. By the way, I am working with John Morrison to package distcc for Cygwin. This patch allows us to consolidate our documentation in the directory that we specify. It was also marked as a TODO in configure.ac so it seems that this was wanted. Yes, it's fine. Great. Once these changes are released we can have an empty platform-specific patch :) Harold