naviserver-devel — Developer discussion

Re: [naviserver-devel] bad import

[Vlad S. <vl...@cr...>]

I managed to remove the wrong files but not sure if directory is there

Vlad Seryakov wrote:
> Stephen,
> 
> I was importing new module and by mistake imported it into 
> naviserver/modules repository, second time i was able to put it into 
> modules/nsaccess.
> 
> Can you ask admins to remove the wrong import?
> 
> thanks
> 

-- 
Vlad Seryakov
571 262-8608 office
vl...@cr...
http://www.crystalballinc.com/vlad/

[naviserver-devel] bad import

[Vlad S. <vl...@cr...>]

Stephen,

I was importing new module and by mistake imported it into 
naviserver/modules repository, second time i was able to put it into 
modules/nsaccess.

Can you ask admins to remove the wrong import?

thanks

-- 
Vlad Seryakov
571 262-8608 office
vl...@cr...
http://www.crystalballinc.com/vlad/

Re: [naviserver-devel] Re: [naviserver-commits] naviserver/include ns.h,1.27,1.28

[Vlad S. <vl...@cr...>]

yes, i was confued with that function as well.
if it is not used we better rename/drop it, i agree about using V but 
that was the best i was able to come up with:-)))

Stephen Deasey wrote:
> On 6/28/05, Vlad Seryakov <ser...@us...> wrote:
> 
>>Update of /cvsroot/naviserver/naviserver/include
>>In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv3191/include
>>
>>Modified Files:
>>        ns.h
>>Log Message:
>>Added Ns_ConnVSetHeaders function to add headers to the output in printf-style way
>>Added support for Range: header for returning partial content
>>
>>
>>Index: ns.h
>>===================================================================
>>RCS file: /cvsroot/naviserver/naviserver/include/ns.h,v
>>retrieving revision 1.27
>>retrieving revision 1.28
>>diff -C2 -d -r1.27 -r1.28
>>*** ns.h        13 Jun 2005 06:14:26 -0000      1.27
>>--- ns.h        28 Jun 2005 20:15:43 -0000      1.28
>>***************
>>*** 1014,1017 ****
>>--- 1014,1018 ----
>>  NS_EXTERN int Ns_ConnFlushHeaders(Ns_Conn *conn, int status);
>>  NS_EXTERN void Ns_ConnSetHeaders(Ns_Conn *conn, char *field, char *value);
>>+ NS_EXTERN void Ns_ConnVSetHeaders(Ns_Conn *conn, char *field, char *fmt,...);
>>  NS_EXTERN void Ns_ConnCondSetHeaders(Ns_Conn *conn, char *field, char *value);
>>  NS_EXTERN void Ns_ConnReplaceHeaders(Ns_Conn *conn, Ns_Set *newheaders);
> 
> 
> 
> The 'V' in Ns_ConnVSetHeaders() suggests it takes a va_list, as in
> Ns_DStringVPrintf() used in the implementation.
> 
> Unfortunately, Ns_ConnPrintfHeader() already exists.  It prints
> directly to the output stream which looks pretty silly.  The function
> is unused.
> 
> How about we rename your function Ns_ConnPrintfHeaders and drop
> Ns_ConnPrintfHeader?  Does anyone think that will be a problem?
> 
> 
> -------------------------------------------------------
> SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
> from IBM. Find simple to follow Roadmaps, straightforward articles,
> informative Webcasts and more! Get everything you need to get up to
> speed, fast. http://ads.osdn.com/?ad_idt77&alloc_id492&opÌk
> _______________________________________________
> naviserver-devel mailing list
> nav...@li...
> https://lists.sourceforge.net/lists/listinfo/naviserver-devel

-- 
Vlad Seryakov
571 262-8608 office
vl...@cr...
http://www.crystalballinc.com/vlad/

[naviserver-devel] Re: [naviserver-commits] naviserver/include ns.h,1.27,1.28

[Stephen D. <sd...@gm...>]

On 6/28/05, Vlad Seryakov <ser...@us...> wrote:
> Update of /cvsroot/naviserver/naviserver/include
> In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv3191/include
>=20
> Modified Files:
>         ns.h
> Log Message:
> Added Ns_ConnVSetHeaders function to add headers to the output in printf-=
style way
> Added support for Range: header for returning partial content
>=20
>=20
> Index: ns.h
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> RCS file: /cvsroot/naviserver/naviserver/include/ns.h,v
> retrieving revision 1.27
> retrieving revision 1.28
> diff -C2 -d -r1.27 -r1.28
> *** ns.h        13 Jun 2005 06:14:26 -0000      1.27
> --- ns.h        28 Jun 2005 20:15:43 -0000      1.28
> ***************
> *** 1014,1017 ****
> --- 1014,1018 ----
>   NS_EXTERN int Ns_ConnFlushHeaders(Ns_Conn *conn, int status);
>   NS_EXTERN void Ns_ConnSetHeaders(Ns_Conn *conn, char *field, char *valu=
e);
> + NS_EXTERN void Ns_ConnVSetHeaders(Ns_Conn *conn, char *field, char *fmt=
,...);
>   NS_EXTERN void Ns_ConnCondSetHeaders(Ns_Conn *conn, char *field, char *=
value);
>   NS_EXTERN void Ns_ConnReplaceHeaders(Ns_Conn *conn, Ns_Set *newheaders)=
;


The 'V' in Ns_ConnVSetHeaders() suggests it takes a va_list, as in
Ns_DStringVPrintf() used in the implementation.

Unfortunately, Ns_ConnPrintfHeader() already exists.  It prints
directly to the output stream which looks pretty silly.  The function
is unused.

How about we rename your function Ns_ConnPrintfHeaders and drop
Ns_ConnPrintfHeader?  Does anyone think that will be a problem?

[naviserver-devel] ns_cache

[Zoran V. <zv...@ar...>]

FYI:

> See the stand-alone nscache module.  I have NO idea why this isn't
> part
> of the AOLserver core ... unless folks really object, I think we
> should
> fold the nscache module in.
>

  Agreed -- I'll add it in.  In particular, "ns_cache eval" is smart.


This is from the AS mailing list: seems that we all need this in the
core :-)

Zoran

[naviserver-devel] Fixup completed

[Zoran V. <zv...@ar...>]

Hi!

Well, I have corrected all of the places touched by the
latest Win ckeckin objected by Stephen (perfectly valid
objections...).
It is to a great extent my fault and partially due to
the bad timing we (i.e. Ibrahim) added those changes to
the code base, as already explained.

Cheers
Zoran

Re: [naviserver-devel] Re: [naviserver-commits] naviserver/nsd return.c,1.5,1.6

[Zoran V. <zv...@ar...>]

Am 24.06.2005 um 22:41 schrieb Stephen Deasey:

>
> Finally, is this function necessary?  Ns_ConnReturnFile already exists
> in fastpath.c.
>

Absolutely. I'm removing this from the code.
Zoran

Re: [naviserver-devel] Re: [naviserver-commits] naviserver/nsd nsmain.c,1.13,1.14

[Zoran V. <zv...@ar...>]

Am 27.06.2005 um 05:48 schrieb Stephen Deasey:

>
> What does 'tcp' stand for?  How about 'cwd':
>

tcp is rubbish. I have fixed this now.
Zoran

[naviserver-devel] Re: [naviserver-commits] naviserver/nsd tclcmds.c,1.14,1.15

[Zoran V. <zv...@ar...>]

Hi!

I've seen some problems with last checkin for Win which
Ibrahim did over the weekend.

The timing was not really ideal since I was away over
the weekend and have not double-checked everything
on all OS'es promptly.

I will go now, double-check everything and rectify
some issues which popped-up by this checkin. I will
also see that Ibrahim gets on the development list
so he gets the info back from you promptly if something
is broken.

Sorry for the confusion, but this is the type of
things we'd be needing to handle every now and then
when new people are about to start working in the
project.

Cheers
Zoran

Re: [naviserver-devel] Re: [naviserver-commits] naviserver/nsd nswin32.c,1.6,1.7

[Zoran V. <zv...@ar...>]

Am 27.06.2005 um 07:06 schrieb Stephen Deasey:

> This Windows only command asks for a restart, right?  This should be
> hooked up to the -restart switch of the ns_shutdown command, which is
> essentially the same thing.  Otherwise, Tcl programmers will have to
> test the current platform to portably restart the server.
>


I believe yes. I will see to marry those two. Ibrahim was not aware
that we have the -restart option now.

Zoran 

[naviserver-devel] Re: [naviserver-commits] naviserver/nsd nswin32.c,1.6,1.7

[Stephen D. <sd...@gm...>]

On 6/24/05, abe-t <it...@us...> wrote:
> Update of /cvsroot/naviserver/naviserver/nsd
> In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv16846/nsd
>=20
> Modified Files:
>         nswin32.c
> Log Message:
>   Added provisions to fake that nsd as a service failed on Windows in ord=
er
>     for the Service Control Manager to restart it automatically.
>=20
>=20
> Index: nswin32.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> RCS file: /cvsroot/naviserver/naviserver/nsd/nswin32.c,v
> retrieving revision 1.6
> retrieving revision 1.7
> diff -C2 -d -r1.6 -r1.7
> *** nswin32.c   13 Jun 2005 10:17:27 -0000      1.6
> --- nswin32.c   24 Jun 2005 08:40:27 -0000      1.7
> ***************
> *** 51,54 ****
> --- 51,55 ----
>   static void ExitService(void);
>   static char *GetServiceName(Ns_DString *dsPtr, char *server);
> + static int ReportException(int ec, char *msg);
>   static SERVICE_STATUS_HANDLE hStatus =3D 0;
>   static SERVICE_STATUS curStatus;
> ***************
> *** 57,60 ****
> --- 58,62 ----
>   static int tick;
>   static int sigpending;
> + static int servicefailed =3D 0;
>=20
>   #define SysErrMsg() (NsWin32ErrMsg(GetLastError()))
> ***************
> *** 349,353 ****
>    */
>=20
> ! void
>   NsHandleSignals(void)
>   {
> --- 351,355 ----
>    */
>=20
> ! int
>   NsHandleSignals(void)
>   {
> ***************
> *** 388,391 ****
> --- 390,395 ----
>           StartTicker(SERVICE_STOP_PENDING);
>       }
> +
> +     return pending;
>   }
>=20
> ***************
> *** 413,416 ****
> --- 417,421 ----
>       switch (sig) {
>       case NS_SIGTERM:
> +     case NS_SIGINT:
>       case NS_SIGHUP:
>           Ns_MutexLock(&lock);
> ***************
> *** 932,948 ****
>   ServiceMain(DWORD argc, LPTSTR *argv)
>   {
> !     hStatus =3D RegisterServiceCtrlHandler(argv[0], ServiceHandler);
> !     if (hStatus =3D=3D 0) {
> !         Ns_Fatal("nswin32: RegisterServiceCtrlHandler() failed: '%s'",
> !                  SysErrMsg());
>       }
> -     curStatus.dwServiceType =3D SERVICE_WIN32_OWN_PROCESS;
> -     curStatus.dwServiceSpecificExitCode =3D 0;
> -     StartTicker(SERVICE_START_PENDING);
> -     Ns_Main(argc, argv, NULL);
> -     StopTicker();
> -     ReportStatus(SERVICE_STOP_PENDING, NO_ERROR, 100);
> -     ReportStatus(SERVICE_STOPPED, 0, 0);
> -     Ns_Log(Notice, "nswin32: service exiting");
>   }
>=20
> --- 937,964 ----
>   ServiceMain(DWORD argc, LPTSTR *argv)
>   {
> !     __try {
> !         hStatus =3D RegisterServiceCtrlHandler(argv[0], ServiceHandler)=
;
> !         if (hStatus =3D=3D 0) {
> !             Ns_Fatal("nswin32: RegisterServiceCtrlHandler() failed: '%s=
'",
> !                      SysErrMsg());
> !         }
> !         curStatus.dwServiceType =3D SERVICE_WIN32_OWN_PROCESS;
> !         curStatus.dwServiceSpecificExitCode =3D 0;
> !         StartTicker(SERVICE_START_PENDING);
> !         Ns_Main(argc, argv, NULL);
> !         StopTicker();
> !         ReportStatus(SERVICE_STOP_PENDING, NO_ERROR, 100);
> !         if (!servicefailed) {
> !             ReportStatus(SERVICE_STOPPED, 0, 0);
> !         }
> !         Ns_Log(Notice, "nswin32: service exiting");
> !
> !         if(servicefailed) {
> !             exit(-1);
> !         }
> !     }
> !     __except (ReportException(GetExceptionCode(), "ServiceMain")) {
> !         // No code; this block is never executed.
>       }
>   }
>=20
> ***************
> *** 1006,1009 ****
> --- 1022,1028 ----
>       curStatus.dwCurrentState =3D state;
>       curStatus.dwWin32ExitCode =3D code;
> +     if (code =3D=3D ERROR_SERVICE_SPECIFIC_ERROR) {
> +         curStatus.dwServiceSpecificExitCode =3D code;
> +     }
>       curStatus.dwWaitHint =3D hint;
>       if (state =3D=3D SERVICE_RUNNING || state =3D=3D SERVICE_STOPPED) {
> ***************
> *** 1090,1091 ****
> --- 1109,1163 ----
>   }
>=20
> +
> + /*
> +  *---------------------------------------------------------------------=
-
> +  *
> +  * NsTclFailServiceObjCmd --
> +  *
> +  *    Tell the server not to unregister from the Service Control Manage=
r
> +  *  (SCM) when exiting, effectively leaving SCM wondering about the
> +  *  service, thinking it failed and restarting it
> +  *  (provided restarting the serice is configured in SCM).
> +  *
> +  * Results:
> +  *    None.
> +  *
> +  *---------------------------------------------------------------------=
-
> +  */
> +
> + int
> + NsTclFailServiceObjCmd(ClientData dummy, Tcl_Interp *interp, int argc, =
char **argv)
> + {
> +     servicefailed =3D 1;
> +
> +     return TCL_OK;
> + }



This Windows only command asks for a restart, right?  This should be
hooked up to the -restart switch of the ns_shutdown command, which is
essentially the same thing.  Otherwise, Tcl programmers will have to
test the current platform to portably restart the server.




> + /*
> +  *---------------------------------------------------------------------=
-
> +  *
> +  * reportException --
> +  *
> +  *    Handle expecptions - cause the server to terminate abruptly
> +  *    and leave a log trace. It is intended to prevents the default
> +  *    handling in Windows where a Window pops up and the user has to
> +  *    confirm - which then prevents the service control manager to
> +  *    restart the service. (The idea, however, does not seem to work
> +  *    as intended - Windows still brings up the popup for reporting
> +  *    the "bug" to Microsoft).
> +  *
> +  * Results:
> +  *    None.
> +  *
> +  * Side effects:
> +  *    Exits the server
> +  *
> +  *---------------------------------------------------------------------=
-
> +  */
> +
> + static int ReportException(int ec, char *msg)
> + {
> +     fprintf(stderr, "EXCEPTION %x in %s\n", ec, msg); fflush(stderr);
> +     exit(-1);
> +     return EXCEPTION_EXECUTE_HANDLER;
> + }



If this is experimental debugging code it probably shouldn't be
committed to CVS.  It does look pretty experimental -- there's no way
this function will return a value after calling exit(), for example.

[naviserver-devel] Re: [naviserver-commits] naviserver/nsd nsmain.c,1.13,1.14

[Stephen D. <sd...@gm...>]

On 6/24/05, abe-t <it...@us...> wrote:
> Update of /cvsroot/naviserver/naviserver/nsd
> In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv16462/nsd
>=20
> Modified Files:
>         nsmain.c
> Log Message:
>   Corrected normalizing the home path on Windows.
>=20
>=20
> Index: nsmain.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> RCS file: /cvsroot/naviserver/naviserver/nsd/nsmain.c,v
> retrieving revision 1.13
> retrieving revision 1.14
> diff -C2 -d -r1.13 -r1.14
> *** nsmain.c    11 Jun 2005 10:03:56 -0000      1.13
> --- nsmain.c    24 Jun 2005 08:38:27 -0000      1.14
> ***************
> *** 111,115 ****
>       int  i, fd, sig, optind, cmdargc;
>       char **cmdargv;
> !     char *config;
>       Ns_Time timeout;
>=20
> --- 111,115 ----
>       int  i, fd, sig, optind, cmdargc;
>       char **cmdargv;
> !     char *config, *tcp;
>       Ns_Time timeout;
>=20
> ***************
> *** 592,596 ****
>        */
>=20
> !     nsconf.home =3D getcwd(buf, sizeof(buf));
>       if (nsconf.home =3D=3D NULL) {
>           Ns_Fatal("nsmain: getcwd failed: '%s'", strerror(errno));
> --- 592,597 ----
>        */
>=20
> !     tcp =3D
> !     nsconf.home =3D getcwd(NULL, _MAX_PATH);
>       if (nsconf.home =3D=3D NULL) {
>           Ns_Fatal("nsmain: getcwd failed: '%s'", strerror(errno));
> ***************
> *** 604,608 ****
>           ++nsconf.home;
>       }
> !     nsconf.home =3D buf;
>=20
>       /*
> --- 605,609 ----
>           ++nsconf.home;
>       }
> !     nsconf.home =3D tcp;
>=20
>       /*


What does 'tcp' stand for?  How about 'cwd':

    nsconf.home =3D getcwd(NULL, _MAX_PATH);
    if (nsconf.home =3D=3D NULL) {
        Ns_Fatal("nsmain: getcwd failed: '%s'", strerror(errno));
    }
    for (cwd =3D nsconf.home; *cwd !=3D '\0'; cwd++) {
        if (*cwd =3D=3D '\\') {
            *cwd =3D '/';
        } else if (isupper(*cwd)) {
            *cwd =3D tolower(*cwd);
        }
    }

[naviserver-devel] Re: [naviserver-commits] naviserver/nsd driver.c,1.14,1.15 binder.c,1.9,1.10

[Stephen D. <sd...@gm...>]

On 6/24/05, abe-t <it...@us...> wrote:
> Update of /cvsroot/naviserver/naviserver/nsd
> In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv15355/nsd
>=20
> Modified Files:
>         driver.c binder.c
> Log Message:
>   Excluded Unix calls from Windows code since some of the structs needed
>     to compile them are missing in Windows.
>=20
>=20
> Index: driver.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> RCS file: /cvsroot/naviserver/naviserver/nsd/driver.c,v
> retrieving revision 1.14
> retrieving revision 1.15
> diff -C2 -d -r1.14 -r1.15
> *** driver.c    19 Jun 2005 16:31:58 -0000      1.14
> --- driver.c    24 Jun 2005 08:36:29 -0000      1.15
> ***************
> *** 448,453 ****
> --- 448,455 ----
>           if (drvPtr->opts & NS_DRIVER_UDP) {
>               drvPtr->sock =3D Ns_SockListenUdp(drvPtr->bindaddr, drvPtr-=
>port);
> + #ifndef _WIN32
>           } else if (drvPtr->opts & NS_DRIVER_UNIX) {
>               drvPtr->sock =3D Ns_SockListenUnix(drvPtr->bindaddr);
> + #endif
>           } else {
>               drvPtr->sock =3D Ns_SockListenEx(drvPtr->bindaddr, drvPtr->=
port,
>=20
> Index: binder.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> RCS file: /cvsroot/naviserver/naviserver/nsd/binder.c,v
> retrieving revision 1.9
> retrieving revision 1.10
> diff -C2 -d -r1.9 -r1.10
> *** binder.c    15 Jun 2005 04:14:04 -0000      1.9
> --- binder.c    24 Jun 2005 08:36:29 -0000      1.10
> ***************
> *** 36,40 ****
>=20
>   #include "nsd.h"
> ! #include <sys/un.h>
>=20
>   NS_RCSID("@(#) $Header$");
> --- 36,43 ----
>=20
>   #include "nsd.h"
> !
> ! #ifndef _WIN32
> ! # include <sys/un.h>
> ! #endif
>=20
>   NS_RCSID("@(#) $Header$");
> ***************
> *** 52,56 ****
>   #ifndef _WIN32
>   static void PreBind(char *line);
> - #endif
>=20
>=20
> --- 55,58 ----
> ***************
> *** 103,107 ****
>       return (SOCKET)sock;
>   }
> !
>=20
>   /*
> --- 105,109 ----
>       return (SOCKET)sock;
>   }
> ! #endif /* _WIN32 */
>=20
>   /*
> ***************
> *** 189,192 ****
> --- 191,195 ----
>=20
>=20
> + #ifndef _WIN32
>   /*
>    *---------------------------------------------------------------------=
-
> ***************
> *** 237,241 ****
>       return (SOCKET)sock;
>   }
> !
>=20
>   /*
> --- 240,244 ----
>       return (SOCKET)sock;
>   }
> ! #endif /* _WIN32 */
>=20
>   /*
> ***************
> *** 275,278 ****
> --- 278,282 ----
>=20
>=20
> + #ifndef _WIN32
>   /*
>    *---------------------------------------------------------------------=
-
> ***************
> *** 315,319 ****
>       return (SOCKET)sock;
>   }
> !
>=20
>   /*
> --- 319,323 ----
>       return (SOCKET)sock;
>   }
> ! #endif /* _WIN32 */
>=20
>   /*


You could have the functions return an error for Windows rather than
defining them out.  Something like::


SOCKET
Ns_SockListenUnix(char *path)
{
    int sock =3D -1;
    Tcl_HashEntry *hPtr;
    Tcl_HashSearch search;

+#ifdef _WIN32
+    Ns_Log(Error, "AF_UNIX sockets not supported on Windows: %s", path);
+    return -1;
+#else
+    Ns_MutexLock(&lock);
+    ...
+#endif


This has the advantage that you don't also have to ifdef around the
use of this function, as is done for the prebind functions and use of
Ns_SockListenUnix in driver.c.  It's easier for module writers, too.

I'm not sure why Ns_SockListenEx is no longer available on Windows.=20
This is for simple TCP sockets, and it's called by Ns_SockListen in
sock.c.

[naviserver-devel] Re: [naviserver-commits] naviserver/include nscheck.h,1.2,1.3

[Stephen D. <sd...@gm...>]

On 6/24/05, abe-t <it...@us...> wrote:
> Update of /cvsroot/naviserver/naviserver/include
> In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv14150/include
>=20
> Modified Files:
>         nscheck.h
> Log Message:
>   VC for Windows cannot handle GNU preprocessor
>     directives and macros (fixed).
>=20
>=20
> Index: nscheck.h
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> RCS file: /cvsroot/naviserver/naviserver/include/nscheck.h,v
> retrieving revision 1.2
> retrieving revision 1.3
> diff -C2 -d -r1.2 -r1.3
> *** nscheck.h   11 Jun 2005 20:00:48 -0000      1.2
> --- nscheck.h   24 Jun 2005 08:35:14 -0000      1.3
> ***************
> *** 39,42 ****
> --- 39,43 ----
>   #define NSCHECK_H
>=20
> + #ifndef _WIN32
>=20
>   #undef  __GNUC_PREREQ
> ***************
> *** 109,112 ****
> --- 110,131 ----
>   #endif
>=20
> + #else /* _WIN32 */
> +
> + # define NS_GNUC_SENTINEL
> + # define NS_GNUC_NONNULL
> + # define NS_GNUC_WARN_UNUSED_RESULT
> + # define NS_GNUC_MAYALIAS
> + # define NS_GNUC_DEPRECATED
> + # define NS_GNUC_USED
> + # define NS_GNUC_FORMAT(m)
> + # define NS_GNUC_UNUSED
> + # define NS_GNUC_NORETURN
> + # define NS_GNUC_PRINTF(fmtarg, firstvararg)
> + # define NS_GNUC_SCANF(fmtarg, firstvararg)
> + # define NS_GNUC_MALLOC
> + # define NS_GNUC_PURE
> + # define NS_GNUC_CONST
> +
> + #endif /* _WIN32 */
>=20
>   /*


I think a better fix here would be something like this, right?

#undef  __GNUC_PREREQ
#if defined __GNUC__ && defined __GNUC_MINOR__
# define __GNUC_PREREQ(maj, min) \
        ((__GNUC__ << 16) + __GNUC_MINOR__ >=3D ((maj) << 16) + (min))
+#else
+# define __GNUC_PREREQ(maj, min) (0)
#endif

[naviserver-devel] Re: [naviserver-commits] naviserver/nsd return.c,1.5,1.6

[Stephen D. <sd...@gm...>]

On 6/24/05, abe-t <it...@us...> wrote:
> Update of /cvsroot/naviserver/naviserver/nsd
> In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv18008/nsd
>=20
> Modified Files:
>         return.c
> Log Message:
>   Added Ns_ConnReturnPath to open a file and send its contents
>     via the connection.
>=20
>=20
> Index: return.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> RCS file: /cvsroot/naviserver/naviserver/nsd/return.c,v
> retrieving revision 1.5
> retrieving revision 1.6
> diff -C2 -d -r1.5 -r1.6
> *** return.c    11 Jun 2005 20:04:11 -0000      1.5
> --- return.c    24 Jun 2005 08:42:02 -0000      1.6
> ***************
> *** 1362,1363 ****
> --- 1362,1397 ----
>       return result;
>   }
> +
> + /*
> +  *---------------------------------------------------------------------=
-
> +  *
> +  * Ns_ConnReturnPath --
> +  *
> +  *  Open a path and send contents out the conn.
> +  *
> +  * Results:
> +  *  See ReturnOpen.
> +  *
> +  * Side effects:
> +  *  See ReturnOpen.
> +  *
> +  * Note:
> +  *  Added by Archiware
> +  *
> +  *---------------------------------------------------------------------=
-
> +  */
> +
> + NS_EXPORT int
> + Ns_ConnReturnPath(Ns_Conn *conn, int status, char *type, char *path, in=
t len);
> +
> + int
> + Ns_ConnReturnPath(Ns_Conn *conn, int status, char *type, char *path, in=
t len)
> + {
> +     int fd =3D open(path, O_RDONLY|O_BINARY);
> +     int rv;
> +
> +     if (fd < 0) return -1;
> +     rv =3D ReturnOpen(conn, status, type, NULL, NULL, fd, len);
> +     close(fd);
> +     return rv;
> + }


I noticed a couple of problems with this change:

The function declaration should never come immediately before the
definition; you're not getting any extra checking from the compiler
this way.  Private (static) declarations go at the top of the file,
declarations private to a module go in private headers (nsd.h in this
case) and have no underscore in the prefix (NsConnReturnPath), and
public declarations go in include/ns.h which is where this one lives.

There's no 'Note:' section in the comments where you can put 'Added by
Archiware'.  Be verbose in the commit message/ChangeLog if you must.

The body of an 'if' statement is always enclosed in braces. Do only
simple initialisation of variables in the variable declaration block.=20
You can pick this up by looking at existing code, but here's our new
style guidelines doc, such as it is:

http://naviserver.sourceforge.net/wiki/index.php/Code_Standards


Finally, is this function necessary?  Ns_ConnReturnFile already exists
in fastpath.c.

Re: [naviserver-devel] uploading large files

[Zoran V. <zv...@ar...>]

Am 17.06.2005 um 22:26 schrieb Stephen Deasey:

>
> We just need to build up the test library one test at a time.
>

OK. I'm going to expand this when I come to rewriting this
upload file stuff. But one thing at a time. Currently I'm
rewriting Tcl VFS related stuff and this will take me
(given the level of work I have lately) at least one week
or more :-(

But, we are in no rush...

Zoran

Re: [naviserver-devel] New member

[Zoran V. <zv...@ar...>]

Am 17.06.2005 um 22:29 schrieb Stephen Deasey:

> Hmm, actually we'll no loger be able to use the optimised FILE_KEYS
> due to the Tcl VFS back end...  :-(
>

See: man Tcl_FSSstat()

No big deal... No need to worry.
Zoran

Re: [naviserver-devel] New member

[Stephen D. <sd...@gm...>]

On 6/17/05, Zoran Vasiljevic <zv...@ar...> wrote:
> Hi!
>=20
> I have included my collegue Ibrahim Tannir (username itannir)
> to the list of project developers. He will mainly take care
> about the Windos part of the code, improve it as the time goes
> by and hopefully arrange a decent build-system so we can build
> on Windows so nicely as we build on Unix :-)


Groovy!  So, no more Windows droppings sprinkled throughout the code:

#ifdef _WIN32
    keys =3D TCL_STRING_KEYS;
#else
    keys =3D FILE_KEYS;
#endif


Hmm, actually we'll no loger be able to use the optimised FILE_KEYS
due to the Tcl VFS back end...  :-(

Re: [naviserver-devel] uploading large files

[Stephen D. <sd...@gm...>]

On 6/17/05, Zoran Vasiljevic <zv...@ar...> wrote:
>=20
> Am 17.06.2005 um 21:23 schrieb Stephen Deasey:
>=20
> >
> > Sounds like a plan for stage 1.
> >
>=20
> Let me then try this one and see how far I'll come :-)
>=20
> >
>=20
> > If there's any part of the code base we should be testing, it's this.
>=20
> Yes. But this is also the most complex one...


Luckily, the tests are easy to write :-)

We already have some tests for this area of code in tests/http.test.=20
Here's a 3 line test for 404 pages:

test http-1.3 {HTTP/1.0 GET} -constraints serverListen -body {
    nstest_http -getbody 1 GET /noexist
} -match glob -result {404 *Not Found*}

We just need to build up the test library one test at a time.

Re: [naviserver-devel] uploading large files

[Zoran V. <zv...@ar...>]

Am 17.06.2005 um 21:23 schrieb Stephen Deasey:

>
> Sounds like a plan for stage 1.
>

Let me then try this one and see how far I'll come :-)

>

> If there's any part of the code base we should be testing, it's this.

Yes. But this is also the most complex one...

Cheers
Zoran

Re: [naviserver-devel] uploading large files

[Stephen D. <sd...@gm...>]

On 6/17/05, Zoran Vasiljevic <zv...@ar...> wrote:
>=20
> Am 17.06.2005 um 12:30 schrieb Zoran Vasiljevic:
>=20
> >>
> >> I see that quite a lot of plumbing must be done
> >> to get this working. We must keep the Sock for
> >
>=20
> Wrong. Not very many plumbings, as we have almost
> all in place.
>=20
> The chages would be mainly in the driver.c SockRead()
> to declare SOCK_READY at the point of reading maxreadahead
> even if there are still bytes left to consume from socket.
>=20
> The Ns_ConnContent has to check this fact and go and
> fetch the remaining (spooling all thats left in a
> mmaped file) before returning the pointer to the
> content.
>=20
> Ns_ConnClose should respectively do unmapping
> of files, if any mapped.  Conn structure should be expaned
> to hold mmaped file maintenance fields of course.


Sounds like a plan for stage 1.


> The NsTclConnObjCmd should get someting like
> ns_conn setmaxinput or similar (details can be cleared here).


Stage 2...


> All in all, it does not seems like too much plumbing
> as I anticipated.


It would be great to beef up the test suite in this area.  This has
historicaly been a buggy part of the code base (you found another one
just yesterday), which I guess is inevitable:  a lot of string
processing and complicted sequence of OS calls.

If there's any part of the code base we should be testing, it's this.

Re: [naviserver-devel] New member

[Bernd E. <eid...@we...>]

> Am 17.06.2005 um 15:08 schrieb Bernd Eidenschink:
> > Welcome Tannir!
>
> His first name is: Ibrahim.
> (try again... :-)

I knew, I just, eh, wanted, I, test, test, aaaaaaaaaaaaaaaaaaarrrrrrgh ;-)

Welcome, Ibrahim! :-)

Re: [naviserver-devel] New member

[Zoran V. <zv...@ar...>]

Am 17.06.2005 um 15:08 schrieb Bernd Eidenschink:

>
> Welcome Tannir!

His first name is: Ibrahim.
(try again... :-)

Zoran

Re: [naviserver-devel] New member

[Bernd E. <eid...@we...>]

Welcome Tannir!

> Hi!
>
> I have included my collegue Ibrahim Tannir (username itannir)
> to the list of project developers. He will mainly take care
> about the Windos part of the code, improve it as the time goes
> by and hopefully arrange a decent build-system so we can build
> on Windows so nicely as we build on Unix :-)

[naviserver-devel] New member

[Zoran V. <zv...@ar...>]

Hi!

I have included my collegue Ibrahim Tannir (username itannir)
to the list of project developers. He will mainly take care
about the Windos part of the code, improve it as the time goes
by and hopefully arrange a decent build-system so we can build
on Windows so nicely as we build on Unix :-)


Cheers
Zoran