Re: A new version of setup very soon?
On Fri, 2003-02-28 at 15:20, Christopher Faylor wrote: I would really appreciate it if we could release a new version of setup.exe with (at least) Pierre Humblet's ntsec changes soon. I think it is becoming very clear that Pierre's ntsec changes are required if we want sane behavior for executables after installation. Having Gary's resize option available wouldn't hurt either. Robert, I know you're busy but is there any way you could maybe offload the release onto someone else? I'm willing to give someone cvs access if that helps. We've got a set of logic bugs to fix before we can release - orthogonal to the resize and ntsec changes. cvs access isn't the bottleneck - Max has been doing a great job of commiting post-review. The bottleneck is programmer time - someone needs to fix those logic bugs and submit a patch. There was a thread on this the last time we got keen about a release :}. I am really really hoping to flush the patch queue this weekend at least as far as code reviews goes. (although I may end up asking Max to commit :}). Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Pending packages status
1. grace date : 25 Nov 2002 version: 5.1.10-1 status : not reviewed notes : http://www.cygwin.com/ml/cygwin-apps/2002-11/msg00322.html votes : 2 (Lapo and Robert) url: http://www.scytek.de/cygwin/grace-5.1.10-1.tar.bz2 http://www.scytek.de/cygwin/grace-5.1.10-1-src.tar.bz2 http://www.scytek.de/cygwin/setup.hint 2. nfs-server date : 09 Dec 2002 version: 2.2.47-1 status : reviewed; there are several pending issues (more info can be found in the nfs related threads starting after Feb 11, 2003) notes : http://www.cygwin.com/ml/cygwin-apps/2002-12/msg00113.html http://www.cygwin.com/ml/cygwin-apps/2003-01/msg00072.html http://www.cygwin.com/ml/cygwin-apps/2003-01/msg00109.html http://www.cygwin.com/ml/cygwin-apps/2003-01/msg00211.html http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00133.html reviews: http://www.cygwin.com/ml/cygwin-apps/2002-12/msg00117.html http://www.cygwin.com/ml/cygwin-apps/2002-12/msg00120.html http://www.cygwin.com/ml/cygwin-apps/2002-12/msg00161.html http://www.cygwin.com/ml/cygwin/2003-01/msg00885.html http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00135.html http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00149.html http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00153.html http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00154.html http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00173.html http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00175.html votes : 4 (Christopher, Corinna, Hack and Sergey) url: http://www.oneparticularharbor.net/cygwin/nfs-server/nfs-server-2.2.47-1.tar.bz2 http://www.oneparticularharbor.net/cygwin/nfs-server/nfs-server-2.2.47-1-src.tar.bz2 http://www.oneparticularharbor.net/cygwin/nfs-server/setup.hint 3. LPRng date : 21 Jan 2003 version: 3.8.19-1 status : reviewed; updated package available notes : http://www.cygwin.com/ml/cygwin-apps/2003-01/msg00215.html http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00172.html reviews: http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00061.html votes : 2 (Joshua and Volker) url: https://www.as.cmu.edu/~geek/LPRng/LPRng-3.8.19-1.tar.bz2 https://www.as.cmu.edu/~geek/LPRng/LPRng-3.8.19-1-src.tgz https://www.as.cmu.edu/~geek/LPRng/setup.hint 4. ifhp date : 21 Jan 2003 version: 3.5.10-1 status : not reviewed notes : http://www.cygwin.com/ml/cygwin-apps/2003-01/msg00215.html http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00172.html votes : 1 (Joshua) url: https://www.as.cmu.edu/~geek/ifhp/ifhp-3.5.10-1.tar.bz2 https://www.as.cmu.edu/~geek/ifhp/ifhp-3.5.10-1-src.tar.bz2 https://www.as.cmu.edu/~geek/ifhp/setup.hint 5. TCM date : 27 Jan 2003 version: 2.20-1 status : not reviewed notes : http://www.cygwin.com/ml/cygwin-apps/2003-01/msg00299.html http://www.cygwin.com/ml/cygwin-apps/2003-01/msg00100.html votes : 2 (Christopher and Lapo) url: 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 6. par date : 04 Feb 2003 version: 1.52-1 status : reviewed; ready to upload once it gets the votes notes : http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00040.html reviews: http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00062.html votes : 1 (Joshua) url: http://www.lapo.it/tmp/par-1.52-1.tar.bz2 http://www.lapo.it/tmp/par-1.52-1-src.tar.bz2 7. docbook-xml42 date : 13 Feb 2003 version: 4.2-1 status : not reviewed notes : http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00148.html http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00211.html votes : none url: http://telka.sk/docbook/xml42/docbook-xml42-4.2-1.tar.bz2 http://telka.sk/docbook/xml42/docbook-xml42-4.2-1-src.tar.bz2 http://telka.sk/docbook/xml42/setup.hint 8. docbook-xsl date : 13 Feb 2003 version: 1.60.1-1 status : not reviewed notes : http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00148.html votes : none url: http://telka.sk/docbook/xsl/docbook-xsl-1.60.1-1.tar.bz2 http://telka.sk/docbook/xsl/docbook-xsl-1.60.1-1-src.tar.bz2 http://telka.sk/docbook/xsl/setup.hint 9. xmlto date : 13 Feb 2003 version: 0.0.12-1 status : not reviewed notes : http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00148.html votes : none url: http://telka.sk/docbook/xmlto/xmlto-0.0.12-1.tar.bz2 http://telka.sk/docbook/xmlto/xmlto-0.0.12-1-src.tar.bz2 http://telka.sk/docbook/xmlto/setup.hint 10. pdksh date : 20 Feb 2003 version: 5.2.14-1 status : reviewed; the patch in the source package needs to be regenerated notes : http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00190.html reivews: http://www.cygwin.com/ml/cygwin-apps/2003-02/msg00194.html
Re: ntsec patch for setup
Here is the same patch modified to be applied on main.cc from
setup-200207 branch.
On Fri, 10 Jan 2003, Pierre A. Humblet wrote:
This is the patch mentioned in the previous e-mail. I seems
to work fine on NT and an early version was tested on 2000 (there
are differences).
Essentially the installed files should be in the Users or
Admins groups instead of None.
Nothing changes if the user running setup does not have the
None gid.
Pierre
2003/01/10 Pierre Humblet [EMAIL PROTECTED]
* main.cc (set_default_dacl): Replace with set_default_sec.
(set_default_sec): Attempt to change the default group to
Users or Administrators if it is None, in addition to setting
Everyone in the default DACL.
(main): Call set_default_sec instead of set_default_dacl.
Index: main.cc
===
RCS file: /cvs/cygwin-apps/setup/main.cc,v
retrieving revision 2.18
diff -u -p -r2.18 main.cc
--- main.cc 4 May 2002 12:15:56 - 2.18
+++ main.cc 28 Feb 2003 11:26:55 -
@@ -74,13 +74,15 @@ HINSTANCE hinstance;
#define iswinnt(GetVersion() 0x8000)
void
-set_default_dacl ()
+set_default_sec ()
{
/* To assure that the created files have a useful ACL, the
default DACL in the process token is set to full access to
everyone. This applies to files and subdirectories created
in directories which don't propagate permissions to child
- objects. */
+ objects.
+ To assure that the files group is meaningful, a token primary
+ group of None is changed to Users or Administrators. */
/* Create a buffer which has enough room to contain the TOKEN_DEFAULT_DACL
structure plus an ACL with one ACE. */
@@ -98,20 +100,29 @@ set_default_dacl ()
return;
}
- /* Get the SID for Everyone. */
- PSID sid;
+ PSID esid = NULL, asid = NULL, usid = NULL;
+ HANDLE token = NULL;
+ struct {
+PSID psid;
+char buf[MAX_SID_LEN];
+ } gsid;
+ char lsid[MAX_SID_LEN];
+ char compname[MAX_COMPUTERNAME_LENGTH + 1];
+ char domain[MAX_COMPUTERNAME_LENGTH + 1];
+ DWORD size;
+
SID_IDENTIFIER_AUTHORITY sid_auth = { SECURITY_WORLD_SID_AUTHORITY };
- if (!AllocateAndInitializeSid (sid_auth, 1, 0, 0, 0, 0, 0, 0, 0, 0, sid))
+ if (!AllocateAndInitializeSid (sid_auth, 1, 0, 0, 0, 0, 0, 0, 0, 0, esid))
{
log (LOG_TIMESTAMP) AllocateAndInitializeSid() failed:
GetLastError () endLog;
- return;
+ goto out;
}
/* Create the ACE which grants full access to Everyone and store it
in dacl-DefaultDacl. */
if (!AddAccessAllowedAce
- (dacl-DefaultDacl, ACL_REVISION, GENERIC_ALL, sid))
+ (dacl-DefaultDacl, ACL_REVISION, GENERIC_ALL, esid))
{
log (LOG_TIMESTAMP) AddAccessAllowedAce() failed: %lu
GetLastError () endLog;
@@ -119,7 +130,6 @@ set_default_dacl ()
}
/* Get the processes access token. */
- HANDLE token;
if (!OpenProcessToken (GetCurrentProcess (),
TOKEN_READ | TOKEN_ADJUST_DEFAULT, token))
{
@@ -132,13 +142,111 @@ set_default_dacl ()
if (!SetTokenInformation (token, TokenDefaultDacl, dacl, sizeof buf))
log (LOG_TIMESTAMP) OpenProcessToken() failed: GetLastError ()
endLog;
+ /* Get the default group */
+ if (!GetTokenInformation (token, TokenPrimaryGroup, gsid, sizeof gsid, size))
+{
+ log (LOG_TIMESTAMP) GetTokenInformation() failed:
+ GetLastError () endLog;
+ goto out;
+}
+ /* Get the computer name */
+ if (!GetComputerName (compname, (size = sizeof compname, size)))
+{
+ log (LOG_TIMESTAMP) GetComputerName() failed:
+ GetLastError () endLog;
+ goto out;
+}
+
+ /* Get the local domain SID */
+ SID_NAME_USE use;
+ DWORD sz;
+ if (!LookupAccountName (NULL, compname, lsid, (size = sizeof lsid, size),
+ domain, (sz = sizeof domain, sz), use))
+{
+ log (LOG_TIMESTAMP) LookupAccountName() failed:
+ GetLastError () endLog;
+ goto out;
+}
+
+ /* Create the None SID from the domain SID.
+ On NT the last subauthority of a domain is -1 and it is replaced by the RID.
+ On other systems the RID is appended. */
+ sz = *GetSidSubAuthorityCount (lsid);
+ if (*GetSidSubAuthority (lsid, sz -1) != (DWORD) -1)
+*GetSidSubAuthorityCount (lsid) = ++sz;
+ *GetSidSubAuthority (lsid, sz -1) = DOMAIN_GROUP_RID_USERS;
+
+ /* See if the group is None */
+ if (EqualSid (gsid.psid, lsid))
+{
+ bool isadmins = false, isusers = false;
+ sid_auth = (SID_IDENTIFIER_AUTHORITY) { SECURITY_NT_AUTHORITY };
+ /* Get the SID for Administrators S-1-5-32-544 */
+ if (!AllocateAndInitializeSid (sid_auth, 2, SECURITY_BUILTIN_DOMAIN_RID,
+DOMAIN_ALIAS_RID_ADMINS, 0, 0, 0, 0, 0, 0, asid))
+{
+ log (LOG_TIMESTAMP) AllocateAndInitializeSid()
[PATCH] libtool patch for direct-linking-to-dll
Hi all, appended is a libtool patch (based on the current official libtool release 2003-02-16 for enabling the direct-linking-to-dll functionality, which was applied a few weeks ago to the binutils cvs release and give a major linking time improvement especially for big applications and/or libraries. See http://sources.redhat.com/ml/binutils/2002-12/msg00396.html for further details. Please note that this patch currently works only with the current cvs ld release and should be used as a prelimary patch, because it does not check, if ld does support this functionality. It is enabled by default. A way for checking this should be discussed. I see two ways for dealing this with: 1. Adding an ld option like --disable-dll-linking or so, which existance could be checked in libtool, autoconf or configure (currently I don't where the best place is). This option isn't a must be and it is only used for detecting the functionality, because the necessary ld changes are backward compatible. 2. Using the ld's release time 2003 $ ld -v GNU ld version 2.xx.yy 2003... 3. Another great idea, I currently do not have Any hints or comments ? 2003-02-27 Ralf Habacker [EMAIL PROTECTED] * libtool.m4 (AC_LIBTOOL_SYS_DYNAMIC_LINKER): Removed postinstall_cmds and postuninstall_cmds, added shared library to 'library_names_spec'. (AC_LIBTOOL_LANG_CXX_CONFIG): Removed import library generation from 'archive_cmds'. * ltmain.sh: (install cygwin/mingw): added installing of shared libraries into 'bin' dir (uninstall cygwin/mingw): added uninstalling of shared libraries Ralf libtool.m4.dif Description: Binary data ltmain.sh.dif Description: Binary data
Re: ntsec patch for setup
On Fri, 2003-02-28 at 22:44, Pavel Tsekov wrote: Here is the same patch modified to be applied on main.cc from setup-200207 branch. This looks reasonable. Pavel, if you have write access, please commit to the setup-200207 branch, otherwise, Max - if you have time could you? Failing that, I'll commit on Sunday. Either way, I'll build a release of this and upload to the snapshots page sunday. Cheers, Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: Request #2 to add a new package: ccdoc
On Thu, 27 Feb 2003, Joe Linoff wrote: You can find the updated package distribution files here: http://ccdoc.sourceforge.net/cygwin_tmp/ccdoc-0.8.39-1 Ok, seems good now. There a still two problems though: 1) The files in the binary package should be installed in /usr/bin, /usr/man and /usr/doc. Now they are installed in /bin, /man, /doc. I think this was ok yesterday, but i can't verify since i removed the old package files. 2) The 'category' directive in setup.hint contains 'Dev'. This should be changed to 'Devel'.
Re: ntsec patch for setup
Robert Collins wrote: On Fri, 2003-02-28 at 22:44, Pavel Tsekov wrote: Here is the same patch modified to be applied on main.cc from setup-200207 branch. This looks reasonable. Pavel, if you have write access, please commit to the setup-200207 branch, otherwise, Max - if you have time could you? Failing that, I'll commit on Sunday. Either way, I'll build a release of this and upload to the snapshots page sunday. Will that version still generate passwd-grp.bat from internal data? If so, domain users will continue to suffer. On the other hand if it doesn't, then we MUST update passwd-grp.sh or we fly right into the mountain. Pierre
RE: ntsec patch for setup
Sorry guys, that should have been to the list. From: Pierre A. Humblet [mailto:[EMAIL PROTECTED] Robert Collins wrote: On Fri, 2003-02-28 at 22:44, Pavel Tsekov wrote: Here is the same patch modified to be applied on main.cc from setup-200207 branch. This looks reasonable. Pavel, if you have write access, please commit to the setup-200207 branch, otherwise, Max - if you have time could you? Failing that, I'll commit on Sunday. Either way, I'll build a release of this and upload to the snapshots page sunday. Will that version still generate passwd-grp.bat from internal data? If so, domain users will continue to suffer. On the other hand if it doesn't, then we MUST update passwd-grp.sh or we fly right into the mountain. If you send me the required updates I'll host a new package. I don't have time to do more :( 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 commitment through the use of this email. Experian Limited (registration number 653331). Registered office: Talbot House, Talbot Street, Nottingham NG1 5HF
Re: ntsec patch for setup
On Fri, 28 Feb 2003, Pierre A. Humblet wrote: Robert Collins wrote: On Fri, 2003-02-28 at 22:44, Pavel Tsekov wrote: Here is the same patch modified to be applied on main.cc from setup-200207 branch. This looks reasonable. Pavel, if you have write access, please commit to the setup-200207 branch, otherwise, Max - if you have time could you? Failing that, I'll commit on Sunday. Either way, I'll build a release of this and upload to the snapshots page sunday. Will that version still generate passwd-grp.bat from internal data? If so, domain users will continue to suffer. On the other hand if it doesn't, then we MUST update passwd-grp.sh or we fly right into the mountain. Pierre, please advise how to proceed. I can create the necessary patches to backport the removal of the code which creates passwd-grp.bat. I don't think it would be a big deal. Are there any other changes necessary.
Re: A new version of setup very soon?
Robert Collins wrote: On Fri, 2003-02-28 at 15:20, Christopher Faylor wrote: I would really appreciate it if we could release a new version of setup.exe with (at least) Pierre Humblet's ntsec changes soon. I think it is becoming very clear that Pierre's ntsec changes are required if we want sane behavior for executables after installation. Having Gary's resize option available wouldn't hurt either. Gary's current pending patch is a foundation, but without user-visible changes so far. Robert, I know you're busy but is there any way you could maybe offload the release onto someone else? I'm willing to give someone cvs access if that helps. We've got a set of logic bugs to fix before we can release - orthogonal to the resize and ntsec changes. Yes. But the ntsec changes *could* be applied to the 2.249 (setup-200206) branch, though - the patch applies with fuzz, but no rejects. And that could be released right now. cvs access isn't the bottleneck - Max has been doing a great job of commiting post-review. The bottleneck is programmer time - someone needs to fix those logic bugs and submit a patch. There was a thread on this the last time we got keen about a release :}. I am really really hoping to flush the patch queue this weekend at least as far as code reviews goes. (although I may end up asking Max to commit :}). I'm happy to do that. Max.
Re: ntsec patch for setup
Robert Collins wrote: On Fri, 2003-02-28 at 22:44, Pavel Tsekov wrote: Here is the same patch modified to be applied on main.cc from setup-200207 branch. This looks reasonable. Pavel, if you have write access, please commit to the setup-200207 branch, otherwise, Max - if you have time could you? Failing that, I'll commit on Sunday. Either way, I'll build a release of this and upload to the snapshots page sunday. The current release setup is the setup-200206 branch. There is presumably some reason why the setup-200207 branch fizzled into inactivity with only a snapshot, and no actual release. I can commit ASAP, once this issue is talked out, but I think it might be wise to use 200206 for an interim release now, unless Robert is very sure 200207 is OK. Max.
Re: ntsec patch for setup
On Fri, 28 Feb 2003, Max Bowsher wrote: Robert Collins wrote: On Fri, 2003-02-28 at 22:44, Pavel Tsekov wrote: Here is the same patch modified to be applied on main.cc from setup-200207 branch. This looks reasonable. Pavel, if you have write access, please commit to the setup-200207 branch, otherwise, Max - if you have time could you? Failing that, I'll commit on Sunday. Either way, I'll build a release of this and upload to the snapshots page sunday. The current release setup is the setup-200206 branch. There is presumably some reason why the setup-200207 branch fizzled into inactivity with only a snapshot, and no actual release. I can commit ASAP, once this issue is talked out, but I think it might be wise to use 200206 for an interim release now, unless Robert is very sure 200207 is OK. Ooops! Right. Should be 200206. Anyway the patch I've posted earlier is against the same version of main.cc - version 2.18. So it should apply to 200206 too.
Re: ntsec patch for setup
Morrison, John wrote: BTW, this script doesn't add any domain users... -c adds the current user, if a domain user. If you are on a domain account at work, please test it, it takes a few seconds. Just did. How does it cope if it can't access the domain? It doesn't even try. It squeezes all the information it can from the local machine, e.g. HOMEDRIVE/HOMEPATH to find the home directory. Pierre
Re: ntsec patch for setup
On Sat, 2003-03-01 at 01:32, Max Bowsher wrote: I can commit ASAP, once this issue is talked out, but I think it might be wise to use 200206 for an interim release now, unless Robert is very sure 200207 is OK. 200206. I wasn't thinking for a sec there. Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
[PATCH] setup-200206: Do not create /etc/postinstall/passwd-grp.bat
From the ChangeLog entry for version 2.34 of desktop.cc:
* desktop.cc (etc_profile): Remove.
(make_etc_profile): Remove.
(uexists): Remove.
(make_passwd_group): Remove.
(do_desktop_setup): Don't call removed functions.
Index: desktop.cc
===
RCS file: /cvs/cygwin-apps/setup/desktop.cc,v
retrieving revision 2.30
diff -u -p -r2.30 desktop.cc
--- desktop.cc 12 May 2002 11:28:22 - 2.30
+++ desktop.cc 28 Feb 2003 15:32:25 -
@@ -64,37 +64,6 @@ static OSVERSIONINFO verinfo;
@*
*/
-static const char *etc_profile[] = {
- PATH=\/usr/local/bin:/usr/bin:/bin:$PATH\,
- ,
- USER=\`id -un`\,
- ,
- # Set up USER's home directory,
- if [ -z \$HOME\ ]; then,
-HOME=\/home/$USER\,
- fi,
- ,
- if [ ! -d \$HOME\ ]; then,
-mkdir -p \$HOME\,
- fi,
- ,
- export HOME USER,
- ,
- for i in /etc/profile.d/*.sh ; do,
-if [ -f $i ]; then,
- . $i,
-fi,
- done,
- ,
- export MAKE_MODE=unix,
- export PS1='\\[\\033]0;\\w\\007,
- [EMAIL PROTECTED] \\[\\033[33m\\w\\033[0m\\],
- $ ',
- ,
- cd \$HOME\,
- 0
-};
-
#define COMMAND9XARGS String(/E:4096 /c)
#define COMMAND9XEXE String(\\command.com)
@@ -216,101 +185,6 @@ make_cygwin_bat ()
}
static void
-make_etc_profile ()
-{
- String fname = cygpath (/etc/profile);
-
- /* if the file exists, don't overwrite it */
- if (_access (fname.cstr_oneuse(), 0) == 0)
-return;
-
- char os;
- switch (verinfo.dwPlatformId)
-{
-case VER_PLATFORM_WIN32_NT:
- os = 'N';
- break;
-case VER_PLATFORM_WIN32_WINDOWS:
- if (verinfo.dwMinorVersion == 0)
- os = '5';
- else
- os = '8';
- break;
-default:
- os = '?';
- break;
-}
- msg (os is %c, os);
-
- FILE *p = fopen (fname.cstr_oneuse(), wb);
- if (!p)
-return;
-
- int i, allow = 1;
- for (i = 0; etc_profile[i]; i++)
-{
- if (etc_profile[i][0] == '@')
- {
- allow = 0;
- msg (profile: %s, etc_profile[i]);
- for (const char *cp = etc_profile[i] + 1; *cp; cp++)
- if (*cp == os || *cp == '*')
- allow = 1;
- msg (allow is %d\n, allow);
- }
- else if (allow)
- fprintf (p, %s\n, etc_profile[i]);
-}
-
- fclose (p);
-}
-
-static int
-uexists (const char *path)
-{
- String f = cygpath (path);
- int a = _access (f.cstr_oneuse(), 0);
- if (a == 0)
-return 1;
- return 0;
-}
-
-static void
-make_passwd_group ()
-{
- String fname = cygpath (/etc/postinstall/passwd-grp.bat);
- io_stream::mkpath_p (PATH_TO_FILE, String(file://) + fname);
-
- if ((uexists (/etc/passwd) || uexists (/etc/passwd.lnk))
- (uexists (/etc/group) || uexists (/etc/group.lnk)))
-return;
-
- if (verinfo.dwPlatformId != VER_PLATFORM_WIN32_NT)
-{
- packagedb db;
- packagemeta *pkg = db.packages.getbykey (cygwin);
- if (pkg pkg-installed)
- {
- /* mkpasswd and mkgroup are not working on 9x/ME up to 1.1.5-4 */
- String border_version = canonicalize_version (1.1.5-4);
- String inst_version =
- canonicalize_version (pkg-installed-Canonical_version ());
- if (inst_version.compare(border_version) = 0)
- return;
- }
-}
-
- FILE *p = fopen (fname.cstr_oneuse(), wt);
- if (!p)
-return;
- if (!(uexists (/etc/passwd) || uexists (/etc/passwd.lnk)))
-fprintf (p, bin\\mkpasswd -l etc\\passwd\n);
- if (!(uexists (/etc/group) || uexists (/etc/group.lnk)))
-fprintf (p, bin\\mkgroup -l etc\\group\n);
- fclose (p);
-}
-
-static void
save_icon ()
{
iconname = backslash (cygpath (/cygwin.ico));
@@ -338,8 +212,6 @@ do_desktop_setup ()
save_icon ();
make_cygwin_bat ();
- make_etc_profile ();
- make_passwd_group ();
if (root_menu)
{
Re: ntsec patch for setup
Robert Collins wrote:
On Sat, 2003-03-01 at 01:32, Max Bowsher wrote:
I can commit ASAP, once this issue is talked out, but I think it
might be wise to use 200206 for an interim release now, unless
Robert is very sure 200207 is OK.
200206. I wasn't thinking for a sec there.
Right.
I've built setup from 200206 with ntsec patch as backported by Pavel. I ran
into other problems:
1) I had to merge these fixes, which enable building of setup with gcc-2.
cvs upd -kk -j2.22 -j2.23 Makefile.am
cvs upd -kk -j2.84 -j2.85 Makefile.in
cvs upd -kk -j1.5 -j1.6 libgetopt++/Makefile.am
They are to do with erroneous inclusion of -lstdc++ (which is of course
gcc3's now)
2) gcc2 (not g++2) chokes on some lines in /usr/include/mingw/string.h. I
had to patch as below, adding prototypes before the functions declared
inline.
I propose to commit the backported ntsec patch and gcc-2 fixes when I
return, in about 2 hours.
I see Pavel has just sent in a passwd-grp.bat removal patch just now. I'll
apply that once it gets the OK from Robert.
Max.
Patch follows:
--- /usr/include/mingw/string.h.orig 2003-02-28 15:07:21.0 +
+++ /usr/include/mingw/string.h 2003-02-28 15:12:50.0 +
-162,12 +162,16
int strcmpi (const char*, const char*);
int stricmp (const char*, const char*);
extern __inline__ int
+strcasecmp (const char * __sz1, const char * __sz2);
+extern __inline__ int
strcasecmp (const char * __sz1, const char * __sz2)
{return _stricmp (__sz1, __sz2);}
int stricoll (const char*, const char*);
char* strlwr (char*);
int strnicmp (const char*, const char*, size_t);
extern __inline__ int
+strncasecmp (const char * __sz1, const char * __sz2, size_t
__sizeMaxCompare);
+extern __inline__ int
strncasecmp (const char * __sz1, const char * __sz2, size_t
__sizeMaxCompare)
{return _strnicmp (__sz1, __sz2, __sizeMaxCompare);}
char* strnset (char*, int, size_t);
-180,6 +184,8
/* NOTE: There is no _wcscmpi, but this is for compatibility. */
extern __inline__ int
+wcscmpi (const wchar_t * __ws1, const wchar_t * __ws2);
+extern __inline__ int
wcscmpi (const wchar_t * __ws1, const wchar_t * __ws2)
{return _wcsicmp (__ws1, __ws2);}
wchar_t* wcsdup (wchar_t*);
Re: ntsec patch for setup
On Fri, 28 Feb 2003, Pavel Tsekov wrote:
2) gcc2 (not g++2) chokes on some lines in /usr/include/mingw/string.h. I
had to patch as below, adding prototypes before the functions declared
inline.
There are also some complaints about autoload.c. Will see what I can do.
The attached patch will cure some of the warnings in autoload.c, but not
all. I don't know how to fix the others. Its too cryptic for me ;)
The patch fixes these warnings:
../setup/autoload.c:36: warning: initialization discards qualifiers from
pointer target type
../setup/autoload.c:49: warning: initialization discards qualifiers from
pointer target type
../setup/autoload.c:63: warning: function declaration isn't a prototype
These I cannot handle:
../setup/autoload.c:36: warning: `wininet_info' defined but not used
../setup/autoload.c:49: warning: `advapi32_info' defined but not used
../setup/autoload.c:67: warning: `autoload_common' defined but not used
Index: autoload.c
===
RCS file: /cvs/cygwin-apps/setup/autoload.c,v
retrieving revision 2.3
diff -u -p -r2.3 autoload.c
--- autoload.c 13 Nov 2001 01:49:31 - 2.3
+++ autoload.c 28 Feb 2003 16:31:49 -
@@ -19,7 +19,7 @@ static const char *cvsid = \n%%% $Id: a
#include win32.h
typedef struct {
- char *name;
+ const char *name;
HINSTANCE handle;
} DllInfo;
@@ -60,7 +60,7 @@ typedef struct {
char name[100];
} AutoEntry;
-static void autoload_common () __asm__ (autoload_common);
+static void autoload_common (int) __asm__ (autoload_common);
static void
autoload_common (int x)
Re: ntsec patch for setup
On Sat, 2003-03-01 at 05:23, Max Bowsher wrote: Robert Collins wrote: On Fri, 2003-02-28 at 22:44, Pavel Tsekov wrote: Here is the same patch modified to be applied on main.cc from setup-200207 branch. This looks reasonable. Pavel, if you have write access, please commit to the setup-200207 branch, otherwise, Max - if you have time could you? If this is OK for branch, is it ok for HEAD, as well? Yes - I'd meant to apply it a while back actually :]. Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: [PATCH] setup-200206: Do not create/etc/postinstall/passwd-grp.bat
On Sat, 2003-03-01 at 02:45, Pavel Tsekov wrote: From the ChangeLog entry for version 2.34 of desktop.cc: * desktop.cc (etc_profile): Remove. (make_etc_profile): Remove. (uexists): Remove. (make_passwd_group): Remove. (do_desktop_setup): Don't call removed functions. This is fine to be commited. Thanks.. Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: ntsec patch for setup
On Sat, 2003-03-01 at 02:54, Max Bowsher wrote: I've built setup from 200206 with ntsec patch as backported by Pavel. I ran into other problems: 1) I had to merge these fixes, which enable building of setup with gcc-2. cvs upd -kk -j2.22 -j2.23 Makefile.am cvs upd -kk -j2.84 -j2.85 Makefile.in cvs upd -kk -j1.5 -j1.6 libgetopt++/Makefile.am I haven't reviewed these for 200206, but as you need them for 200206 please check the updated files in, with one exception: don't alter libgetopt. It's not branched as I recall, as it's an external dependency. If it is branched, then it's fine to commit the update. If it's not branched, there should be some LDFLAGS or CXXFLAGS to get it to build and link ok. Also, we dont' need to get the fixed version into CVS as the setup source is distributed via itself :}. You can use 'make release' to create the requisite files. 2) gcc2 (not g++2) chokes on some lines in /usr/include/mingw/string.h. I had to patch as below, adding prototypes before the functions declared inline. I propose to commit the backported ntsec patch and gcc-2 fixes when I return, in about 2 hours. Please do. I see Pavel has just sent in a passwd-grp.bat removal patch just now. I'll apply that once it gets the OK from Robert. Max. Patch follows: --- /usr/include/mingw/string.h.orig 2003-02-28 15:07:21.0 + +++ /usr/include/mingw/string.h 2003-02-28 15:12:50.0 + You might want to send a patch for this to cygwin-patches :}. Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: ntsec patch for setup
On Sat, 2003-03-01 at 03:41, Pavel Tsekov wrote: On Fri, 28 Feb 2003, Pavel Tsekov wrote: 2) gcc2 (not g++2) chokes on some lines in /usr/include/mingw/string.h. I had to patch as below, adding prototypes before the functions declared inline. There are also some complaints about autoload.c. Will see what I can do. The attached patch will cure some of the warnings in autoload.c, but not all. I don't know how to fix the others. Its too cryptic for me ;) Uhmm, don't touch autoload! Just disable warnings for autoload.c. Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: ntsec patch for setup
On Tue, 2003-01-21 at 14:00, Pierre A. Humblet wrote:
At 10:24 PM 1/19/2003 -, Max Bowsher wrote:
I'd suggest something like this:
if (isusers)
{
nsid = usid;
log(LOG_TIMESTAMP) Changing gid to Users endLog;
}
else if (isadmins)
nsid = asid;
log(LOG_TIMESTAMP) Changing gid to Administrators endLog;
}
OK, I will wait for Rob comments and add this if Rob doesn't find
it easier to include it himself.
Max, while you're committing things, please throw the above into the
mix. Logging may be essential for diagnosing issues.
Rob
--
GPG key available at: http://users.bigpond.net.au/robertc/keys.txt.
signature.asc
Description: This is a digitally signed message part
Re: ntsec patch for setup
Robert Collins wrote: On Sat, 2003-03-01 at 02:54, Max Bowsher wrote: I've built setup from 200206 with ntsec patch as backported by Pavel. I ran into other problems: 1) I had to merge these fixes, which enable building of setup with gcc-2. cvs upd -kk -j2.22 -j2.23 Makefile.am cvs upd -kk -j2.84 -j2.85 Makefile.in cvs upd -kk -j1.5 -j1.6 libgetopt++/Makefile.am I haven't reviewed these for 200206, but as you need them for 200206 please check the updated files in, with one exception: don't alter libgetopt. It's not branched as I recall, as it's an external dependency. If it is branched, then it's fine to commit the update. Libgetopt++ is branched, and the stuff is committed already. --- /usr/include/mingw/string.h.orig 2003-02-28 15:07:21.0 + +++ /usr/include/mingw/string.h 2003-02-28 15:12:50.0 + You might want to send a patch for this to cygwin-patches :}. Silly me. Oh well, I've raised the issue on cygwin and [EMAIL PROTECTED], so I'll see what response it gets first. Max.
Re: ntsec patch for setup
Robert Collins wrote: On Sat, 2003-03-01 at 03:41, Pavel Tsekov wrote: On Fri, 28 Feb 2003, Pavel Tsekov wrote: 2) gcc2 (not g++2) chokes on some lines in /usr/include/mingw/string.h. I had to patch as below, adding prototypes before the functions declared inline. There are also some complaints about autoload.c. Will see what I can do. The attached patch will cure some of the warnings in autoload.c, but not all. I don't know how to fix the others. Its too cryptic for me ;) Uhmm, don't touch autoload! Just disable warnings for autoload.c. The warnings are only warnings, and are *not* treated as errors. No change needed. Max.
RE: A new version of setup very soon?
Robert Collins wrote: On Fri, 2003-02-28 at 15:20, Christopher Faylor wrote: I would really appreciate it if we could release a new version of setup.exe with (at least) Pierre Humblet's ntsec changes soon. I think it is becoming very clear that Pierre's ntsec changes are required if we want sane behavior for executables after installation. Having Gary's resize option available wouldn't hurt either. Gary's current pending patch is a foundation, but without user-visible changes so far. No, that's not true. The chooser page is significantly larger. It isn't user-resizable, but it is statically larger. -- Gary R. Van Sickle Brewer. Patriot.
Re: A new version of setup very soon?
Gary R. Van Sickle wrote: Robert Collins wrote: On Fri, 2003-02-28 at 15:20, Christopher Faylor wrote: I would really appreciate it if we could release a new version of setup.exe with (at least) Pierre Humblet's ntsec changes soon. I think it is becoming very clear that Pierre's ntsec changes are required if we want sane behavior for executables after installation. Having Gary's resize option available wouldn't hurt either. Gary's current pending patch is a foundation, but without user-visible changes so far. No, that's not true. The chooser page is significantly larger. It isn't user-resizable, but it is statically larger. Ah, ok. Thanks for the correction. Max.
