summaryrefslogtreecommitdiffstats
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
...
* NIM: New Default View and miscellaneous fixesJeffrey Altman2007-03-2044-740/+3031
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ================================ KfW 3.1 Alpha (NetIDMgr 1.1.11.0) -- nidmgr32.dll - Only one action in a menu is allowed to have KHUI_ACTIONREF_DEFAULT flag set. This marks the action as being the default action for the menu and will be rendered as such. - Newly created identities start off with the KCDB_IDENT_FLAG_EMPTY flag set. Once credentials are associated with the identity and the identity is refreshed, the flag will be cleared. - When creating actions, enforce the name length. - khm_value_exists() now handles shadowed configuration spaces. - Add new action KHUI_ACTION_LAYOUT_MINI which toggles between 'Advanced' and 'Basic' views. - Add support for F11 and F12 keys in khui_get_cmd_accel_string(). - New option for alerts to indicate that instead of just setting the response field in the alert, the UI should dispatch the command that the user has selected. -- krb5common.obj - khm_krb5_initialize() can return a handle to a krb5_ccache that has already been closed. Now it doesn't. - Also import 'krb5_string_to_deltat()'. - Work around conditioned symbol definitions in ntsecapi.h in the Vista Platform SDK that affect Win 2000. -- krb5cred.dll - Don't clear the prompts when the options for an identity changes. The prompter code relies on the prompts being around so that the values that the user has entered can be retained if the new set of prompts is the same as the old one. - Use the same code in the new credentials acquisition and the identity configuration code to obtain krb5 parameters for an identity. - Reset the 'IMPORTED' flag when we get new credentials using a password. - If the validity of a principal is not known, then we restrict the options that can be specified when calling krb5_get_init_creds_password() so that we can reliably determine if the principal is valid. If we need to get new credentials for the principal, we need to make another call using the correct options. - The return codes from the prompter need to indicate that the password read operation was cancelled instead of arbiraty non-zero values. - When reading identity settings, if a particular setting is not defined in the registry, then default to reading the settings out of krb5.ini. - Refer to credentials as 'credentials' or 'tickets' instead of 'creds'. - If an identity has imported credentials, don't import for the same identity again. - When importing an identity, create the identity configuration in the registry if we don't already have any settings there. - Work around conditioned symbol definitions in ntsecapi.h in the Vista Platform SDK that affect Win 2000. - Rearrange declarations for clarity. - Use the correct APIs to parse configuration values from krb5.ini. -- krb4cred.dll - The dialog layout was updated to accomodate a localized string that no longer fit in its control. - Remove a spurious inclusion of ntsecapi.h and work around conditioned symbol definition in the Vista Platform SDK. -- netidmgr.exe - Fix the menu creation code to correctly tag the default action so that it will be rendered properly. - Update the menu enumeration code to use documented functions instead of accessing acton lists directly. - Pool of per-identity actions now include a set of actions for obtaining credentials for specific identities. - The default action performed when the notification icon is clicked is now configurable. When displaying the context menu in the notification area, the default action is highlighted. - Remove unnecessary handlers from the notifcation event handler. - Only handle NIN_SELECT instead of both NIN_SELECT and WM_LBUTTONUP in the notification event handler. When the user clicks the notication icon, both events are generated. NIN_SELECT is canonical. - When the handling NIN_BALLOONUSERCLICK in the notification event handler, reset balloon_alert before displaying any new alerts so that we won't overwrite it later. - Reset the notification alert icon after displaying an alert. - If a renewal fails, the displayed alert contains a button that the user can click to initiate the process of acquiring new credentials for the identity. - Alerts can optionally dispatch the commands that were added to it using the KHUI_ALERT_FLAG_DISPATCH_CMD flag. - Increase the size of the About dialog. - Correct the action text for the IDS_ACTION_OPEN_APP and IDS_ACTION_CLOSE_APP to say 'Show' and 'Hide' instead of 'Open' and 'Close'. These actions only control the visible state of the NIM window. - Add additional notification which signals that the commandline has finished processing. - Add an 'acquire' action to the per-identity actions. - The per identity actions (renew, destroy, acquire) now have useful captions, names and tooltips. - Use WM_NEXTDLGCTL message when changing the focus of dialog controls. SetFocus() is insufficient. - If we get a request to show a new credential acquisition dialog and we are already showing one, bring that one to the foreground instead of trying to display a new one or waiting quietly. - New configuration schema for the UI that include definitions for the new default view. - The alerter window can now show more than one alert at once. - If we are about to show queued alerts, then check if the alerts that are waiting are related and if they can be grouped together. If so, show them in a single alert window instead of multiple ones. - If new alerts are issued while a set of alerts are being displayed and if the new alert is related to the alerts that are being displayed, then add the new alert to the list being displayed. - Make sure we have a lock on the alert when we are manipulating or accessing it. - Set the focus to the correct control when displaying an alert. - When adding alerts from the alert queue, make sure we iterate through the queue properly. - Allow keyboard navigation inside the alert window and support scroll bars. - Check if we have a valid code pointer before invoking a UI callback. - Make sure the main window is in the normal configuration before switching to a layout that rquires it. - When moving the main window around, if it comes close to an edge of the working area of the display, snap to it. - Maintain two sets of settings for the main window placement. One for the mini mode and one for the normal mode. - When processing saved window placement information from the configuration, handle docking hints which note which edges of the screen the main window should be adjacent to, if any. - Switching to the 'Basic' view disables the layout and column selection menus. - Position the new credentials dialog above the main window if the main window is visible. - The alert that is displayed to indicate that an identity has expired, now contains a command button that can be used to invoke the new credentials dialog for that identity. -- source - Update the documentation to reflect the change in behavior regarding KHUI_ACTIONREF_DEFAULT in khui_menu_insert_action() and khui_menu_insert_paction(). - Remove notes about menu access functions being not thread safe. This is no longer true. - Update the documentation for khui_alert_show() to document new behavior regarding KHUI_ALERT_FLAG_DISPATCH_CMD. - Update documentation to indicate which KHUI_ALERT_FLAG_* flags are internal and document the new KHUI_ALERT_FLAG_DISPATCH_CMD flag. - Augment the queue handling macros to support additional operations. Also add new tree data structure with an ordered list of children. - Code reorganization to reuse code for obtaining the caption and tooltip for a system defined action in netidmgr.exe. ticket: new component: windows git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19238 dc483132-0cff-0310-8789-dd5450dbe970
* Enable Vista support for MSLSAJeffrey Altman2007-03-201-9/+99
| | | | | | | | | | | | | | | | | The MSLSA: ccache type when used on Windows Vista can take advantage of an ability to write tickets to the LSA credential cache for the current logon session. This is possible due to the addition of the KERB_SUBMIT_TICKET interface. Also new to Vista is the CACHE_INFO_EX2 interface which permits a much more efficient method of enumerating the contents of the LSA credential cache. The code to take advantage of these features has been present for more than a year. However, due to the lack of a public SDK that included the necessary data structures the functionality has been disabled. As of this commit, the functionality will be enabled if the version of NTSecAPI.h includes TRUST_ATTRIBUTE_TRUST_USES_AES_KEYS. This is a preprocessor symbol that is new to the Vista SDK. In order to build with the new Vista functionality when using the XP SP2 SDK, the NTSecAPI.h file from the Vista SDK must be used in place of the version from the XP SP2 SDK. This commit also addresses the issues associated with the inability to read session keys from a UAC limited process. When UAC limitation is detected by examining the process token elevation level all access to the MSLSA contents is disabled. At some point in the future we can implement an elevated COM service in order to obtain access to the session keys. ticket: new tags: pullup git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19237 dc483132-0cff-0310-8789-dd5450dbe970
* Zero sockaddrs in fai_add_entry() so we can compare them with memcmp()Alexandra Ellwood2007-03-191-0/+2
| | | | | | | | | | | | If we don't zero the struct sockaddrs in fai_add_entry() then any sin_zero fields will be left as random memory and the memcmp() will return that the addresses are different even if they aren't. ticket: new target_version: 1.6.1 tags: pullup git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19235 dc483132-0cff-0310-8789-dd5450dbe970
* KfW automated build scripts & supporting filesKevin Koch2007-03-169-276/+899
| | | | | | | | | | Updated scripts & additional configuration files. Ticket: new Target_Version: 1.6.1 tags: pullup git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19234 dc483132-0cff-0310-8789-dd5450dbe970
* Update 3.1.0 to 3.2.0; update location of sample directoryKevin Koch2007-03-161-3/+3
| | | | | | | | Ticket: 5409 Target_Version: 1.6.1 Tags: pullup git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19233 dc483132-0cff-0310-8789-dd5450dbe970
* Update build files for new version and file locationsKevin Koch2007-03-163-16/+13
| | | | | | | Target_Version: 1.6.1 Tags: pullup git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19232 dc483132-0cff-0310-8789-dd5450dbe970
* On Linux, add linker flags -Bsymbolic (makes for smaller libraries, atKen Raeburn2007-03-151-1/+1
| | | | | | | | the cost of being able to override the symbol names we export) and --no-undefined (errors out on shared library generation with undefined symbols). git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19229 dc483132-0cff-0310-8789-dd5450dbe970
* Depend on kadm5 server library too, now that we use the config_params callsKen Raeburn2007-03-151-1/+1
| | | | git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19228 dc483132-0cff-0310-8789-dd5450dbe970
* Use "unsigned __int{16,32}" types for Windows in load_{16,32}_n, per KevinKen Raeburn2007-03-141-0/+8
| | | | | | ticket: 5425 git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19227 dc483132-0cff-0310-8789-dd5450dbe970
* * kadm5_create.c: Include fake-addrinfo.h.Ken Raeburn2007-03-141-6/+21
| | | | | | | | | | (add_admin_princs): Use getaddrinfo instead of gethostbyname. Report the correct message on getaddrinfo errors, and return EINVAL to caller. ticket: 5257 git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19226 dc483132-0cff-0310-8789-dd5450dbe970
* Fix tail portability problem by adding an expression test to theKen Raeburn2007-03-141-1/+1
| | | | | | | | existing awk invocation instead. Patch from Robert Basch at MIT. ticket: 5447 git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19225 dc483132-0cff-0310-8789-dd5450dbe970
* Remove test pass that exercises AES256 over UDP, since we also haveKen Raeburn2007-03-141-12/+0
| | | | | | one that exercises the same configuration over TCP. git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19224 dc483132-0cff-0310-8789-dd5450dbe970
* * include/k5-platform.h: Add load_{16,32,64}_n for loading values fromKen Raeburn2007-03-142-2/+39
| | | | | | | | | | | (possibly unaligned) memory in native order. * lib/krb5/krb/get_in_tkt.c (krb5_get_init_creds): Fetch four random bytes from the crypto library and generate a 31-bit (non-negative) nonce, instead of using the timestamp. ticket: 5425 git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19223 dc483132-0cff-0310-8789-dd5450dbe970
* * sendto_kdc.c (default_debug_handler): If writing to a log file, keepKen Raeburn2007-03-141-8/+22
| | | | | | | | it open and unbuffered instead of always opening and closing. When the format string has characters to be copied literally to the output, write them all at once instead of individually. git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19222 dc483132-0cff-0310-8789-dd5450dbe970
* KFW: Vista Integrated LogonJeffrey Altman2007-03-134-33/+323
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | On Windows Vista the GINA architecture was removed. As a side effect the support for the Logon Event Handlers was also removed. The KFW Integrated Logon functionality relies on the "Logon" event handler to migrate the user's tickets from a secure FILE: ccache to an API: ccache so that the tickets will be available to NetIDMgr and all other Kerberos applications. This functionality is especially important on Vista for accounts that are members of the Administrators group because the User Account Control (UAC) restricts access to the session keys of all tickets in the MSLSA ccache. The only way for tickets to be made available to MIT Kerberos applications is by obtaining them within the Network Provider and pushing them into the Logon Session. This patch replaces the missing Logon Event Handler support with a new exported function "LogonEventHandler" which adheres to the rundll32.exe specifications. The "LogonEventHandler" function accepts as input the name of a FILE ccache and moves the contents into an API: ccache and then deletes the FILE ccache. In order for this to work the FILE ccache must be owned by the account that was used to logon to the current session. The NPLogonNotify() function must therefore lookup the SID for the active account, assign an appropriate DACL to the ccache file, and change the owner. In addition, when Vista is in use a LogonScript must be constructed that will perform the call to rundll32.exe. Other changes include altering the prototype of KFW_copy_ccache_system_file to accept a filename instead of the LogonID. This improves the abstraction and allows the filename to be computed once and passed into multiple functions from NPLogonNotify(). Many debugging calls were added to assist with implementation. #define DEBUG 1 at the top of kfwcommon.c when you wish to build with debugging that generates entries in the Windows Application Event Viewer. It is important to note that Integrated Logon attempts to logon the username within the default realm within the krb5.ini file using the provided password. This is so a local machine account name matching the default realm can obtain Kerberos tickets by synchronizing the password. ticket: new component: windows git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19221 dc483132-0cff-0310-8789-dd5450dbe970
* Remove from unused_passes a configuration also listed in the default passesKen Raeburn2007-03-091-12/+0
| | | | git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19215 dc483132-0cff-0310-8789-dd5450dbe970
* Do use valgrind on /bin/sh commands, since that's how we launch kadmind ↵Ken Raeburn2007-03-061-1/+2
| | | | | | currently git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19212 dc483132-0cff-0310-8789-dd5450dbe970
* reverting commit to windows WIX installer (revision 19207)Jeffrey Altman2007-03-062-8/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | Revision 19207 committed by Kevin Koch does not document changes made to the Wix installer files. The changes that were made remove the installation of the configuration files and the inclusion of leash32.chm. When the decision is made to remove Leash from the installer, the entire Leash feature and the associated properties used to determine when it is installed must be removed. Commenting out individual file components is not sufficient. Configuration files must be installed as part of KFW. The source of the configuration files is defined by the variable "ConfigDir" within the site-local.wxi directory. Any changes made to the WIX installer must also be made to the NSIS installer. The changes to the src/windows/wix directory from 19207 are therefore reverted. ticket: new component: windows git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19211 dc483132-0cff-0310-8789-dd5450dbe970
* osf1: get proper library dependencies installedKen Raeburn2007-03-051-1/+1
| | | | | | | | | | | | | | | | | | The LDCOMBINE setting for Tru64 left out $(CFLAGS), which meant the -pthread option wasn't being passed in, so the libraries didn't indicate a dependency on the pthread library despite the fact that we need it unconditionally (if thread support is enabled). Including $(CFLAGS) should fix this. Our own binaries get linked with -pthread anyways, so they build okay, but other binaries trying to link against or load our libraries could have problems. ticket: new target_version: 1.6.1 tags: pullup git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19208 dc483132-0cff-0310-8789-dd5450dbe970
* Train build script to use repository sources, omit sample configuration. ↵Kevin Koch2007-03-055-176/+250
| | | | | | | | | | | | [wix area.] Add control of the repository access step, fetch all sources from the repository. Track in documentation. Eliminate unhelpful output during pre-package step. The next step is to fetch only krb5/src/windows/build and run the entire build. git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19207 dc483132-0cff-0310-8789-dd5450dbe970
* Move automation documentation from wiki to here.Kevin Koch2007-03-055-1/+1295
| | | | | | | | Minor tweaks to script (start adding unzip support) and config (move svn url to right place). Target_Version: 1.6.1 git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19206 dc483132-0cff-0310-8789-dd5450dbe970
* Automation for building KfWKevin Koch2007-03-056-0/+850
| | | | | | | Target_Version: 1.6.1 Component: KfW git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19205 dc483132-0cff-0310-8789-dd5450dbe970
* Back out windows specific makefile changeSam Hartman2007-03-031-1/+1
| | | | git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19204 dc483132-0cff-0310-8789-dd5450dbe970
* More existence tests; path updateKevin Koch2007-03-0211-14/+13
| | | | | | | | | | | Modify remainder of Makefiles that were sensitive to identity/obj or other cleaned files not being present. Update util/et/Makefile.in to look for com_err.h in src/include, not src/include/src. ticket: 5457 tags: pullup git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19203 dc483132-0cff-0310-8789-dd5450dbe970
* Test for existence of identity/obj before try to delete filesKevin Koch2007-03-021-1/+1
| | | | | | | ticket: new tags: pullup git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19202 dc483132-0cff-0310-8789-dd5450dbe970
* On unload, free up g_mechSet and g_mechListKen Raeburn2007-03-021-7/+33
| | | | git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19201 dc483132-0cff-0310-8789-dd5450dbe970
* update dependenciesKen Raeburn2007-03-022-3/+8
| | | | git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19199 dc483132-0cff-0310-8789-dd5450dbe970
* Delete more stuff for 'clean' and 'distclean' targetsKen Raeburn2007-03-026-6/+8
| | | | git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19198 dc483132-0cff-0310-8789-dd5450dbe970
* valgrind detects uninitialized (but really unused) bytes in 'queue'Ken Raeburn2007-03-011-0/+6
| | | | | | | | | | | | | | | | | | | | | The gsstest program exports a GSSAPI security context to a blob in memory, writes that memory to a file, and reads it back to use it. Under valgrind, the writing phase triggers a warning about uninitialized storage. The "queue" structure as implemented in generic/util_ordering.c holds an array of values, some of which may never be initialized. As far as I can tell, those uninitialized values are never used before being initialized, either, but valgrind doesn't know that. This patch zaps the structure contents (including the array) before using the queue object. ticket: new target_version: 1.6.1 tags: pullup git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19196 dc483132-0cff-0310-8789-dd5450dbe970
* krb5_get_cred_from_kdc fails to null terminate the tgt listJeffrey Altman2007-03-011-0/+1
| | | | | | | | | | | | | | | | | | | | | | if the next tgt in a cross-realm traversal cannot be obtained find_nxt_kdc() was calling krb5_free_creds() on the last tgt in the list but was failing to nullify the pointer to the cred that was just freed. if there were no additional tgts obtained, krb5_get_cred_from_kdc() would return a non-NULL terminated cred list to the caller. This would result in a crash when attempting to manipulate the non-existent cred past the end of the list. This commit nullifies the credential pointer in find_nxt_kdc() after the call to krb5_free_creds() ticket: new tags: pullup git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19195 dc483132-0cff-0310-8789-dd5450dbe970
* Add intalert.h which should have been added as part of Jeffrey Altman2007-02-281-0/+136
| | | | | | | | revision 19189 ticket: 5452 git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19191 dc483132-0cff-0310-8789-dd5450dbe970
* Windows - some apps define ssize_t as a preprocessor symbolJeffrey Altman2007-02-281-0/+3
| | | | | | | | ticket: new component: krb5-libs tags: pullup git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19190 dc483132-0cff-0310-8789-dd5450dbe970
* NIM Improved Alert Management Jeffrey Altman2007-02-2862-1582/+5069
| | | | | | | | | | | | This patch implements the new Alert Management functionality. Many improvements to avoid race conditions and improve resource tracking. ticket: new component: windows git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19189 dc483132-0cff-0310-8789-dd5450dbe970
* Unused macro NEG removedKen Raeburn2007-02-261-1/+0
| | | | git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19183 dc483132-0cff-0310-8789-dd5450dbe970
* When importing a name with a valid (non-null) handle but a zeroKen Raeburn2007-02-241-1/+4
| | | | | | | | | length, set GSS_S_BAD_NAME but not GSS_S_CALL_INACCESSIBLE_READ. ticket: 5445 status: open git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19172 dc483132-0cff-0310-8789-dd5450dbe970
* If a reflection is detected, zap the message buffer pointer outputKen Raeburn2007-02-241-1/+4
| | | | | | | | | | argument as well as actually freeing the buffer. (Found while using the gsstest option to exercise error conditions.) ticket: 5445 status: open git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19171 dc483132-0cff-0310-8789-dd5450dbe970
* Initialize "loopback" field in newly allocated name structureKen Raeburn2007-02-221-0/+1
| | | | | | | ticket: 5445 status: open git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19170 dc483132-0cff-0310-8789-dd5450dbe970
* Ensure consistancy between prototypes and functionsEzra Peisach2007-02-199-30/+23
| | | | | | | | | | | | I am using an older compiler that is complaining that prototypes do not match the functions they reference. The issue is that a number of prototypes are using "const int foo" while the function is "int foo". From a caller sense it makes no difference - but the compiler is correct they are different. All is now consistant. git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19169 dc483132-0cff-0310-8789-dd5450dbe970
* Conditionalize some ipv6 bitsKen Raeburn2007-02-161-1/+9
| | | | git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19168 dc483132-0cff-0310-8789-dd5450dbe970
* Conditionalize more bits that might not be defined by the OSKen Raeburn2007-02-161-0/+10
| | | | git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19167 dc483132-0cff-0310-8789-dd5450dbe970
* Don't use struct in6_addr if we don't have IPv6 supportKen Raeburn2007-02-161-0/+7
| | | | git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19165 dc483132-0cff-0310-8789-dd5450dbe970
* rename krb5_server_decrypt_ticket_keyblock() toTom Yu2007-02-132-5/+6
| | | | | | | | krb5int_server_decrypt_ticket_keyblock() ticket: 5349 git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19159 dc483132-0cff-0310-8789-dd5450dbe970
* hack to permit GetEnvironmentVariable usage without requiring getenv() ↵Jeffrey Altman2007-02-121-0/+9
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | conversion Windows has a major flaw when it comes to the use of getenv/putenv. getenv/putenv do not modify the actual environment of the process. Instead, they modify a copy of the environment block at the time the C Runtime Library was initialized for the current module. In other words, the C Runtime Library environment block for the executable is not the same as the C Runtime Library environment block for the krb5_32.dll library, etc. This results in problems when a process wants to set the default ccache name outside the krb5_context. The krb5_context default ccname disappears when the context is destroyed. gss_acquire_cred() suffers from the creation and destruction of krb5_contexts and therefore the krb5_context default ccname cannot be used to set a default ccname. Instead, the process environment must be used. In order to modify the process environment, SetEnvironmentVariable() must be used. However, this does not result in the C Runtime Library environment blocks being updated. putenv() does not see the definition of "KRB5CCNAME". This patch modifies get_os_ccname() for Windows to check GetEnvironmentVariable() before checking the registry. This hack will work as long as there is no "KRB5CCNAME" variable in the C Runtime Library environment block. The long term solution is to replace all calls to getenv and putenv with GetEnvironmentVariable/SetEnvironmentVariable for Windows. ticket: new tags: pullup git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19154 dc483132-0cff-0310-8789-dd5450dbe970
* In the resolve function - ensure on memory allocation error - Ezra Peisach2007-02-121-3/+6
| | | | | | | | | | deref of null pointer does not happen. Also include kt-int.h for prototypes. ticket: 5411 git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19153 dc483132-0cff-0310-8789-dd5450dbe970
* krb5_kt_default_name should take an unsized lengthEzra Peisach2007-02-052-6/+6
| | | | | | | | | | The internal code to krb5_kt_default_name was casting the length to size_t. Change prototype to take an unsigned - which makes sense in how the code uses it. ticket: new git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19144 dc483132-0cff-0310-8789-dd5450dbe970
* krb5_kt_get_type should return const char *. Ezra Peisach2007-02-054-5/+5
| | | | | | | | | | | | The code returns a pointer to static structures. Just to enforce the assumption that users can not change the returned data. Change prototype of krb5_kt_get_type to return const char *. The other changes are to clean up warnings - no change in code - usage assumed const. ticket: new git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19143 dc483132-0cff-0310-8789-dd5450dbe970
* Add a new program to perform various tests on the WRFILE: and MEMORY: keytabsEzra Peisach2007-02-042-8/+461
| | | | | | | | | | I developed this program to test functionality of the MEMORY keytab - which resulted in the numerous fixes that have been committed recently. Tests all functioanlity of keytabs except for krb5_kt_default() and krb5_kt_read_service_key() - although essential functionality tested. git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19142 dc483132-0cff-0310-8789-dd5450dbe970
* MEMORY keytab krb5_kt_get_entry sets enctype to 0 if unspecifiedEzra Peisach2007-02-041-10/+10
| | | | | | | | | | | | | If one invokes krb5_kt_get_entry() for a memory keytab w/ an unspecified enctype (i.e. 0) - the returned keytab has it's enctype set to 0 as well. Also - in copying out the found keytab_entry - when a kvno is unspecified, the last match found is used - not the one with the highest vno. This was caused by copying out the variable "entry" and not "match". Ticket: new git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19141 dc483132-0cff-0310-8789-dd5450dbe970
* MEMORY keytab does not copy keytab_entry keyblock contentsEzra Peisach2007-02-041-3/+21
| | | | | | | | | | | | | | | | | In krb5_kt_add_entry: The MEMORY keytab does not make a copy of the keytab_entry keyblock contents - but instead retains a pointer to the incomming one. In krb5_kt_get_entry and krb5_kt_get_next - a pointer to internal keyblock contents memory is returned to the caller - which is subsequently freed when tht caller invokes krb5_free_keytab_entry_contents. Solution is to use krb5_copy_keyblock_contents() instead of simply copying the structure. Ticket: new git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19140 dc483132-0cff-0310-8789-dd5450dbe970
* The default K4 compatibility mode is now none, not preauth. Also documentRuss Allbery2007-02-031-3/+7
| | | | | | | | | | | | that the valid values for v4_mode are the valid arguments to the -4 flag to krb5kdc. Ticket: 2724 Component: krb5-doc Target_Version: 1.6.1 Tags: pullup git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19139 dc483132-0cff-0310-8789-dd5450dbe970
generated by cgit (git 2.34.1) at 2026-01-23 18:23:04 +0000