I do a pre- and post network traversing md5sum against VIFs (Very Important
Files) and maintain the outputted hashes for comparison. By this, I mean that I
maintain a checkout on the repo box itself, eliminating any network travel at
all. (Perhaps the file i/o is sent through the loopback in this case, but it is
inconsequential.) I do the md5sum hash there, and then again in the foreign
machine's newly checked-out working dir. If the hashes compare OK, then I have
confidence that the network did not interfere with my checkout.
In the example below, $target_X is a pattern defined elsewhere which maps to my
list of VIFs.
This doesn't address the exact issue the Ulrich seems to be reporting, but may be
a starting point.
...
if ( $host eq $repoMachine ) {
@hash_target = ("$repo_top/$target_1",
"$repo_top/$target_2",
"$repo_top/$target_3",
"$repo_top/$target_4");
foreach $target (@hash_target) {
system("find $target|grep -v CVS|xargs md5sum -t >
$repo_hash_result/$host-md5-hash-pass-$count.txt");
$count++;
}
} else {
@hash_target = ("$buildMachine_top/$target_1",
"$buildMachine_top/$target_2",
"$buildMachine_top/$target_3",
"$buildMachine_home/$target_4");
foreach $target (@hash_target) {
system("find $target|grep -v CVS|xargs md5sum -t >
$buildMach_result/$host-md5-hash-pass-$count.txt");
$count++;
}
[EMAIL PROTECTED] wrote:
> This is a problem. SCCS has a checksum in each file, but RCS does not
> and the RCS file format does not lend itself to adding one. And many
> CVS commands don't read the entire file and so wouldn't be able to
> verify the checksum anyway.
>
> > Additionally CVS should report the line number/section/revision number
> > where the
> > invalid syntax is detected, and it should display a short reason message.
>
> Feel free to submit patches.
_____________________________________
John Minnihan
mailto:[EMAIL PROTECTED]
http://www.freepository.com
_______________________________________________
Info-cvs mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/info-cvs