xtime.h fix for IRIX-5.3

2001-06-13 Thread Martin Neitzel
Here is a trivial fix to make the bleeding edge CVS compile under IRIX-5.3. Without the fix, the src/server.c compilation fails due to including lib/xtime.h twice, both times indirectly. Shields up!... Martin Neitzel Index: lib/ChangeLog =

Re: WinCVS and 'cvs' using different date formats

2001-06-13 Thread Scott Fagg
I just downloaded cvs 1.11.1p1 and WinCVS 1.3 1.11.1.p1 is still using a different format to WinCVS1.2 WinCVS 1.3 crashed when i tried a 'status' . So as far as i can tell the problem still exists. I'm tempted to patch cvs1.11.1p1 myself to fix this.. surely it cant be that much of a patch.

cvs conflict resolution bug ?

2001-06-13 Thread Jacob Burckhardt
Baptiste Lepilleur writes: > the conflict report is > missing a part of the block of text that was removed. > ... > This problem occured with cvs server version 1.10.2 and CVS 1.11 running on a > linux box, and using either WinCVS 1.2 or CVS 1.10.2 as a client on the linux > box with cvs server v

Re: Improper fixed-size buffer usage in connect_to_gserver()

2001-06-13 Thread Michael Richardson
> "Larry" == Larry Jones <[EMAIL PROTECTED]> writes: Larry> Alexey Mahotkin writes: >> + snprintf (buf, BUFSIZE, "cvs@%s", hostinfo->h_name); Larry> snprintf() isn't portable. Machines without snprintf() aren't worth using. If snprintf() isn't available, take it with you.

Re: out of memory during commit

2001-06-13 Thread Larry Jones
Dumas Patrice writes: > > I use top to monitor what happens, and it seems that cvs use more and more > memory until it uses 162Mo, then it stops growing and abort. I don't know > if this limit is a limit linked with the operating system or not. You're almost certainly running into a per-process

out of memory during commit

2001-06-13 Thread Dumas Patrice
Hi, I report something, which may be a bug of cvs or not, because it may also be related to a problem with my memory configuration. It happened with cvs 1.10 rpm binary and 1.11p1 compiled from source. I run linux on an intel platform, my distribution is a redhat 7.0. I have added a big file, (

Re: Improper fixed-size buffer usage in connect_to_gserver()

2001-06-13 Thread Derek R. Price
Larry Jones wrote: > Alexey Mahotkin writes: > > > > +snprintf (buf, BUFSIZE, "cvs@%s", hostinfo->h_name); > > snprintf() isn't portable. I switched this to simply check the length of hostinfo->h_name and error out if necessary rather than using snprintf. Checked in. Thanks. Derek -- Der

Re: Improper fixed-size buffer usage in connect_to_gserver()

2001-06-13 Thread Alexey Mahotkin
> "AM" == Alexey Mahotkin <[EMAIL PROTECTED]> writes: AM> ok, will that do? "%*s" is documented in K&R. Hell sure. --- client.cTue Apr 24 22:14:53 2001 +++ client.c-newWed Jun 13 22:21:53 2001 @@ -4189,13 +4189,15 @@ /* Connect to the server using GSSAPI authentication. */

Re: Improper fixed-size buffer usage in connect_to_gserver()

2001-06-13 Thread Alexey Mahotkin
> "LJ" == Larry Jones <[EMAIL PROTECTED]> writes: LJ> Alexey Mahotkin writes: >> + snprintf (buf, BUFSIZE, "cvs@%s", hostinfo->h_name); LJ> snprintf() isn't portable. ok, will that do? "%*s" is documented in K&R. --- client.cTue Apr 24 22:14:53 2001 +++ client.c-newWed Jun 13

Re: Improper fixed-size buffer usage in connect_to_gserver()

2001-06-13 Thread Larry Jones
Alexey Mahotkin writes: > > +snprintf (buf, BUFSIZE, "cvs@%s", hostinfo->h_name); snprintf() isn't portable. -Larry Jones I'm a genius. -- Calvin ___ Bug-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/bug-cvs

Improper fixed-size buffer usage in connect_to_gserver()

2001-06-13 Thread Alexey Mahotkin
Humm, it seems to me like there is a plain old buffer overflow in connect_to_gserver(). hostinfo->h_name is taken directly from DNS (in init_sockaddr()) which could be under control of remote host. Then it is sprintf()'ed without length checking into buf[1024]. You get an idea. 2001-06-13 A

cvs conflict resolution bug ?

2001-06-13 Thread Baptiste Lepilleur
Hi, we are considering to use CVS as our version control system. My personnal experience with CVS is good (I'm a member of an open-source project on sourceforge). We made some simple test to figure out how it compare to our current version control system. One of those tests consisted in removing

Re: CVS bug - invalid change text

2001-06-13 Thread Larry Jones
[EMAIL PROTECTED] writes: > > I was not complaining about wrong text in the sources checked out or in the > repository > (this might be caused indeed by network problems), > but I was complaining about the internal integrity of the CVS repository. > The internal syntax/structure of the repository

Re: problem in cvs

2001-06-13 Thread Derek R. Price
Pallavi Moolchandani wrote: > or the password and then after i enter my password it sayscvs > login:failed to open /root /.cvspass for reading:No such file or > directorycvs [login aborted ]: fatal Please see if this section of the manual is helpful: http://cvshome.org/docs/manual/cvs_21.html#SE

Re: patch1

2001-06-13 Thread Derek R. Price
Bear Giles wrote: > diff -Naur --recursive cvs.orig/src/CVS/Entries cvs/src/CVS/Entries > --- cvs.orig/src/CVS/EntriesTue Jun 12 17:42:11 2001 > +++ cvs/src/CVS/Entries Tue Jun 12 17:44:25 2001 > @@ -6,18 +6,18 @@ > /ChangeLog-97/1.1.1.1/Sun Jun 10 04:09:10 2001/-ko/ > /Makefile.am/1.1.1.1/