Dear Folks,
So, basically what you are saying is that xc-mail-2.0 is the last
reliable version? Or is the problem something to do with the CVS set?
Right now I have a file, "6710597 Apr 16 02:58 Sent.gz", which has
been stuffed with various editions of "Sent" at various times. Will all
the tinkering with the various compression functions now make further
additions or extractions difficult? How did you manage to successfully
operate that compression function in the past?
Just as an aside, should one expect to tinker with the contents of
machine created files with impunity? How,pray tell, does David archive
these files in the "ARCHIV" sub-directory? Is he using the xc-mail
program or directly gzipping into that directory? Curious onlookers
should like to know.
Shalom,
John B. Brown.
<[EMAIL PROTECTED]>
On Wed, 15 May 2002 20:30:57 +0200 [EMAIL PROTECTED] said:
> Hi David,
>
> so you used the CVS sources, not the tarball?
>
> I think it's not only a problem of too big mail copying into a gziped
> folder, it may be a problem of too many in a short time.
>
> But it looks like the added wait() function call to wait until gzip has
> been finished doesn't help - or maybe doesn't work.
>
> Maybe s simple sleep() call for testing if this is the right idea may help
> or find the problem.
>
> If you can test this, open src/folder_io_gzip.cxx, line 397.
>
> Remove this line and add
>
> sleep(1);
>
> This will slow down everything, so it's nothing for realy production work,
> but for testing - every mail will now take 1 seconds to be written - enough
> time for gzip to finish.
>
> J�rgen
>
>
> David Pilgram wrote:
>
>> Dear J�rgen,
>>
>> I downloaded the latest CVS last night, and tried archiving some emails
>> into ARCHIV. While many of the emails had attachments, there were no
>> large attachments. The received folder had 395 emails, and size 5.8MB.
>>
>> Method 1. Using the folder controller.
>>
>> Here I simply copied received from $user/Mail to $user/Mail/ARCHIV.
>>
>> Depending which time I did this, the archived version had 135, 102 or 395.
>> emails in $user/Mail/ARCHIV/received (file size 3.9MB).
>>
>> It seems unrepeatable in exactly how many emails are in the archived
>> folder. The folder is gzipped, but if one renamed it received.gz, then
>> try and gzip -d it, one gets "invalid compressed data-format violated".
>>
>> This was done repeatedly on the same folder, brought out of the back-up
>> every time.
>>
>> I closed down xcmail. It would not restart. The only way to restart was
>> delete all files, and copy from my back-up of the Mail directory. Clearly
>> something was set in this directory which prevented xcmail starting - it
>> was not a setting in $home/.XCmail directory.
>>
>> Method 2.
>>
>> This was to highlight a number of emails in the received folder, and then
>> Move them to an existing received.gz file.
>>
>> In small numbers, e.g. 15, this seemed reliable. To do a lot, e.g. 50 or
>> 100, then it would appear to start loosing emails. But again, not
>> predictably. After each 15, I would check that they all had been moved
>> correctly. It is a very slow way of backing up!
>>
>>
>> Maybe the problems are due to any particular attachment in some emails -
>> whether it be html, which is the most likely attachment in this folder, or
>> something else.
>>
>> I have all the folders etc archived up, so can run any particular test on
>> it.
>>
>> Regards,
>>
>> David Pilgram.
>>
>
>
