(no subject)
Submitter-Id: net Originator: Andrew R. Reiter Andrew R. Reiter Organization: net Confidential: no Synopsis: remote cvs pserver core -- possible signal race? Severity: critical Priority: high Category: cvs Class: sw-bug Release: cvs-1.11 Environment: System: FreeBSD fledge.watson.org 4.4-PRERELEASE FreeBSD 4.4-PRERELEASE #0: Thu Aug 16 15:16:58 EDT 2001 [EMAIL PROTECTED]:/usr/obj/data/fbsd-stable/src/sys/FLEDGE i386 Description: The function src/server.c:server_cleanup(int), which is the signal handler for a number of signals related to the pserver code, if I understand the code correctly, contains some problems which I'm assuming led to a core I managed to cause remotely. Essentially, the handler function calls a few unsafe/non-reentrant functions, such as select(). This could possibly cause a race condition to occur. When I found the core, I gdb'd it and found it bailing in multiple calls to sigprocmask()... How-To-Repeat: I managed to cause the core to occur remotely by doing: cvs login cvs co src - large source tree (FreeBSD) ctl-C ctl-C ctl-Z killall -9 cvs - kill the process for real Fix: If this is not the code in which the bug is from, atleast make the handler reentrant. ___ Bug-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/bug-cvs
checkoutlist files in subdirectories
Hello,
I have the following problem using a NT 'pserver' CVSNT 1.11.1.2 (Build
36) and having files in subdirectories in the checkoutlist, e.g.
CVSSupport/ACL.pm. The cvs server on NT cannot rename the new temporary
file e.g. .#314 to CVSSupport/ACL.pm. Therefore the administrative
database is not updated correctly.
I think, the problem is before this step, because the server tries to
rename the existing CVSSupport/ACL.pm to .#CVSSupport/ACL.pm, which is not
possible under Windows NT, though our Linux (1.11.1pl1) server seems to
handle this rename.
IMHO, this rename is not correct, neither for Linux nor for Windows. I
guess it should be a rename from CVSSupport/ACL.pm to CVSSupport/.#ACL.pm.
To achieve this I have patched the mkmodules.c. Using this patch (see
below) the above problem is solved and the .#-files appear in the
subdirectories as they can seen in the CVSROOT directory.
BTW, I am using this patch with a Windows and Linux cvs server.
Frank
*** mkmodules.c.#.1.1.1.1 Fri Oct 26 07:45:31 2001
--- mkmodules.c Wed Oct 24 13:55:08 2001
***
*** 787,792
--- 787,793
char *bak;
struct stat statbuf;
char *rcs;
+ char *dirsep;
/* Set x bits if set in original. */
rcs = xmalloc (strlen (real) + sizeof (RCSEXT) + 10);
***
*** 800,806
if (chmod (temp, 0444 | (statbuf.st_mode 0111)) 0)
error (0, errno, warning: cannot chmod %s, temp);
bak = xmalloc (strlen (real) + sizeof (BAKPREFIX) + 10);
! (void) sprintf (bak, %s%s, BAKPREFIX, real);
/* rm .#loginfo */
if (unlink_file (bak) 0
--- 801,814
if (chmod (temp, 0444 | (statbuf.st_mode 0111)) 0)
error (0, errno, warning: cannot chmod %s, temp);
bak = xmalloc (strlen (real) + sizeof (BAKPREFIX) + 10);
! /* rename in the directory */
! dirsep = strrchr (real, '/');
! if (NULL == dirsep)
! (void) sprintf (bak, %s%s, BAKPREFIX, real);
! else
! {
! (void) sprintf (bak, %.*s%s%s, dirsep-real+1, real, BAKPREFIX,
! dirsep+1);
! }
/* rm .#loginfo */
if (unlink_file (bak) 0
===
plettac electronic security GmbH
Videosensorik
Dr. Frank Stockinger
Würzburger Straße 150
90766 Fürth
Tel +49 (0)911 75884 334
fax +49 (0)911 75884 121
email: [EMAIL PROTECTED]
[EMAIL PROTECTED]
___
Bug-cvs mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/bug-cvs
QUIENES SOMOS
PARA SATISFACER REQUERIMIENTOS: 1).- LOS QUE QUIEREN SABER QUIENES SOMOS: SCHETTINO, ADRIANA CAPITAL RUMITTI, RAFAEL V. CAPITAL UMITTI, CARLOS A.CAPITAL CARTA, GENCIANA H. CAPITAL CAVALET NORA ESTHER CAPITAL FEIJO CARLOS ALBERTO LA PLATA RIAL, JORGE DANIEL LA PLATA RODRIGUEZ JUAN CARLOS CORDOBA DIAZ EDGARDO M CORDOBA SCARDELLA JORGE H. ROSARIO BRAVO FRANCISCO PABLO ROSARIO BAHIA BLANCA SANTA FE MARTINEZ DANIEL ALBERTO BAHIA BLANCA CANGAS ALFREDO ROSARIO WIRTH, CARLOS E. ROSARIO AIELLO PEDRO OSCAR MAR DEL PLATA OVALLE DANIEL MAR DEL PLATA GUIMAREY JORGE LA MATANZA DURANY JUAN DIONISIO SAN NICOLAS CAGNAS ALFREDO SAN NICOLAS DI CLEMENTE MARIA R. AVELLANEDA GAMBINA JUAN JOSE AVELLANEDA SANTOS JULIO EDGARDO LANUS URRUTIA ALBERTO CORDOBA ALVAREZ MANUEL MENSOZA TODARO, HUGO ALBERTO POSADAS MAS 21 CHICOS DE UNA ESCUELA DE COMPUTACION QUE NOS AYUDAN EN EL PROCESO DE DATOS.- SEDE EN UNA CASA DE FAMILIA EN CAPITAL FEDERAL.- COMUNICACIONES ENTRE TODOS A TRAVES DE MAIL Y CHAT.- 2).- PARA LOS QUE NOS INJURIARON Y NOS ENSEARON ORTOGRAFIA ESTAMOS MAS HAYA DEL BIEN Y EL MAL, PUEDEN SEGUIR INJURIANDONOS, EL PAIS ESTA POR SOBRE VUESTRA "BUENA EDUCACION".- A LOS QUE NOS ENSEARON ORTOGRAFIA, GRACIAS, MUCHOS DE NOSOTROS, TRABAJAMOS, DEDICAMOS MUCHAS HORAS A ESTE PROYECTO, NO HEMOS PODIDO TERMINAR LA EDUCACION PRIMARIA PORQUE HEMOS TENIDO QUE DAR DE DOMER A NUESTRAS FAMILIAS, POR ELLO, JUNTO A LOS ERRORES QUE NOS INDICARON, TAMBIEN PONGAN EN CONOCIMIENTO DE SUS HIJOS LAS INJURIAS QUE NOS INFIRIERON.- 3).- A LOS QUE SOLICITAN LA BAJA EL PROGRAMA CONSTA DE 23 ARCHIVOS DE 32700 DIRECCIONES CADA UNO - INDISCRIMINADOS - POR ESE MOTIVO LAS BAJAS DEBEN REALZARSE UNA A UNA Y ELLO VA A LLEVAR - EN EL TIEMPO LIBRE MAS DE DOS MESES.- PEDIMOS LAS DISCULPAS DEL CASO Y SUGERIMOS HAGAN USO DE LA OPCION CORREO NO DESEADO DEL OUTOOCK.- 4).- A LOS QUE NOS HAN REQUERIDOCONSULTAS.- NO LAS HEMOS CONTESTADO POR FALTA DE TIEMPO MATERIAL.- POR RAZONES DE ORGANIZACION LAS CONSULTAS DEBEN DIRIGIRSE A: [EMAIL PROTECTED] [EMAIL PROTECTED] LAS ENCUESTAS SOLAMENTE DIRIGIRLAS A [EMAIL PROTECTED] ESTAMOS EN PLENO TRABAJO DE PROCESAMIENTO, LA PRIMER CONSULTA NO LLEVARA APROXIMADAMENTE VEINTE DIAS, A PARTIR DE ALLI RECIBIRA LA PROXIMA.-- QUIENES ENCUENTREN EN ESTE MENSAJE SATISFECHOS SUS REQUERIMIENTOS PUEDEN REMITIR LA CONSULTA A [EMAIL PROTECTED] UN ABRAZO Y HASTA PRONTO FUTURO ARGENTINO ___ Bug-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/bug-cvs
Re: New email address
Hi Brian. Brian Behlendorf tells me that Collab Net will be supporting the bandwidth and hardware for cvshome.org for an indefinite period. Are you sure of this? As of yesterday morning, I've been unable to get to cvshome.org either by web or cvs pserver, so as far as I can tell, it's already dead... The server is definitely up - we moved it to a new network location on Friday and had some resulting routing issues, causing some hefty downtime then, but it should be back up. If you get a clearer idea of in what way the system is down for you (DNS won't resolve or goes to an older IP address, IP address can't be reached, IP address is reachable but connectivity is poor, or IP address connectivity is good but nothing's listening on port 80) then let me know and I'll look at it. When I had a look for it on Sunday morning, I found the following mappings (from memory): cvs.cvshome.org 127.0.0.9 www.cvshome.org 127.0.0.15 Both were 127.0.0.* addresses, and neither was valid - the only valid 127.* addresses here are 127.0.0.1 and 127.0.0.255 !!! I've just tried again, and see the following mapping at the moment: cvs.cvshome.org 64.125.132.231 64.125.132.213 www.cvshome.org 64.125.132.231 I'm not sure whether there should be two IP's for the cvs server, but I see two in response to the `host cvs.cvshome.org` command here, although only one shows up to `host www.cvshome.org` at this time. Based on that, I've just done `cvs update -d -P` from the directory I downloaded the CVS tree to, and get... ===8=== CUT ===8=== $ cvs update -d -P cvs [update aborted]: connect to cvs.cvshome.org(64.125.132.231):2401 failed: connection refused $ ===8=== CUT ===8=== ...which is the same message, but with an IP address that is at least valid - the previous one was not. In case it helps, I've enclosed a traceroute to cvs.cvshome.org when it selected that IP... CollabNet will continue to support the CVS community and codebase by hosting cvshome.org for the indefinite future. The bandwidth load is pretty modest, we're OK on hardware allocations, so it's not a problem. Derek's indicated a willingness to continue administering it, which I definitely appreciate. That I'm certainly glad to hear. These economic times are very tough, and last week CollabNet had to let some very valuable people go, including Derek. He assisted us tremendously with some pretty deep CVS problems, much of the results of which ended up as contributions back (on CVS and other projects as well). If any of you are using CVS in demanding situations and looking to hire a guru, I would recommend him highly. If I was in a position to do so, I'm sure I would. Ufortunately, I'm also the possessor of far too much free time, for similar reasons :( Best wishes from Riley. traceroute to cvs.cvshome.org (64.125.132.231), 30 hops max, 38 byte packets 1 leopard.lns.birmingham.access.planet.net.uk (195.92.168.22) 113.715 ms * 1498.844 ms 2 * PBR-1.BRM.AS5388.NET (195.92.168.2) 111.201 ms * 3 BNR-2.BRM.AS5388.NET (195.92.169.3) 2239.281 ms 2322.246 ms 1158.991 ms 4 BNR-1.TCL.AS5388.NET (195.92.201.153) 118.801 ms 2461.639 ms * 5 BNR-2.TCL.AS5388.NET (195.92.200.195) 2739.316 ms 2634.588 ms * 6 pomegranate.AS5388.NET (195.92.201.13) 2100.165 ms 1577.329 ms * 7 linx.london.above.net (195.66.224.76) 2729.521 ms * * 8 lga1-lhr1-stm4.lga1.above.net (208.184.231.21) 196.944 ms 2450.044 ms * 9 * * iad1-lga1-oc192-2.iad1.above.net (208.184.233.61) 1339.991 ms 10 core1-core4-oc48.iad1.above.net (208.185.0.137) 2061.354 ms * 219.314 ms 11 sjc2-iad1-oc48.sjc2.above.net (216.200.127.26) 270.791 ms 1069.196 ms * 12 sfo1-sjc2-oc48-2.sfo1.above.net (208.184.232.54) 889.482 ms 2820.425 ms * 13 main1colo78-core1-oc48.sfo1.above.net (208.184.228.2) 1910.051 ms 2318.183 ms 1878.781 ms 14 above-gw.sfo1.collab.net (209.133.67.29) 2088.998 ms * 1759.875 ms 15 * * cr1-ge2-2-sfo.collab.net (64.125.132.9) 2269.453 ms 16 d2r1-ge2-1-sfo.collab.net (64.125.132.46) 2348.632 ms * 315.531 ms 17 64.125.132.231 (64.125.132.231) 284.888 ms 1204.667 ms 909.628 ms
