You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@maven.apache.org by "Mark R. Diggory" <md...@apache.org> on 2004/10/06 20:08:38 UTC
Re: md5's
Dion,
I really think this is more systemic in terms of the upload behavior of
Maven.
The delema is that the general policy in promotion for /dist/ right now
is that group "writable" on files are very bad. However, as Henk pointed
out, this doesn't mean you can't take ownership of the jars, it just
means that Maven needs to be a little more savvy about publishing files.
For instance the artifact publishing plugin could catch this issue: if
the file isn't group writable and the directory is, then the file should
just be moved out of the way or deleted and the new file uploaded. It
will then be owned by the individual who uploaded it. This way we are
not left guessing who in the group last altered the file which we have
no way of recording right now when maven uploads a file over an existing
group writable file.
Ultimately it would, as well, be of great benefit if the artifact
publishing worked initially in a staging area as not to confuse existing
cron jobs attempting to validate md5's on the server. This would
probably be best served in the systems default tmp directory.
-Mark
Dion Gillard wrote:
> On Tue, 05 Oct 2004 14:11:03 -0400, Mark R. Diggory <md...@apache.org> wrote:
>
>>Sigh, yes I did, but it was to allow Dion to take ownership of them, now
>>I see he has not.
>
>
> Um, if you want me to change the ownership of some files, just let me
> know. At the moment, group writability is enough from what I
> understand.
>
> [snip]
>
--
Mark Diggory
Open Source Software Developer
Apache Jakarta Project
http://jakarta.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: md5's
Posted by Brett Porter <br...@apache.org>.
http://jira.codehaus.org/browse/MPARTIFACT
Thanks
Mark R. Diggory wrote:
> Thanks Brett. I'll open a JIRA issue, which project would your
> recommend I open it under?
>
> -Mark
>
> Brett Porter wrote:
>
>> If you want to put something to this effect into JIRA I could
>> probably do this. I'm not sure how well it might work with the FTP
>> deployer, but that doesn't apply to Apache so it can be omitted there
>> if troublesome.
>>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: md5's
Posted by "Mark R. Diggory" <md...@latte.harvard.edu>.
Thanks Brett. I'll open a JIRA issue, which project would your recommend
I open it under?
-Mark
Brett Porter wrote:
> If you want to put something to this effect into JIRA I could probably
> do this. I'm not sure how well it might work with the FTP deployer, but
> that doesn't apply to Apache so it can be omitted there if troublesome.
>
> Thanks,
> Brett
>
> Henk P. Penning wrote:
>
>> On Wed, 6 Oct 2004, Mark R. Diggory wrote:
>>
>>
>>
>>> Date: Wed, 06 Oct 2004 14:08:38 -0400
>>> From: Mark R. Diggory <md...@apache.org>
>>> To: Dion Gillard <di...@gmail.com>
>>> Cc: Henk P. Penning <he...@cs.uu.nl>, apmirror@apache.org,
>>> repository@apache.org, Maven Developers List <de...@maven.apache.org>
>>> Subject: Re: md5's
>>>
>>
>>
>>
>>
>>> Ultimately it would, as well, be of great benefit if the artifact
>>> publishing worked initially in a staging area as not to confuse existing
>>> cron jobs attempting to validate md5's on the server. This would
>>> probably be best served in the systems default tmp directory.
>>>
>>
>>
>> That would be great.
>>
>> I think, the best way for adding/replace stuff is
>>
>> -- write a 'temp'
>> -- rename 'temp' to 'file'
>>
>> because a rename is truly atomic if 'temp' and 'file' are
>> in the same file system.
>>
>> If you can implement the 'temp' for 'file' to be,
>> for instance, '.tmp.file', I can easily teach the checkers
>> to ignore '.tmp.*' files. I think rsync does something
>> like that (even better .tmp.$$.file).
>>
>> Just a thought.
>>
>>
>>
>>> -Mark
>>>
>>
>>
>> HPP
>>
>> ---------------------------------------------------------------- _
>> Henk P. Penning, Computer Systems Group R Uithof CGN-A232 _/ \_
>> Dept of Computer Science, Utrecht University T +31 30 253 4106 / \_/ \
>> Padualaan 14, 3584CH Utrecht, the Netherlands F +31 30 251 3791 \_/ \_/
>> http://www.cs.uu.nl/staff/henkp.html M penning@cs.uu.nl \_/
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>> For additional commands, e-mail: dev-help@maven.apache.org
>>
>>
>>
>>
>>
>>
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
>
--
Mark Diggory
Software Developer
Harvard MIT Data Center
http://www.hmdc.harvard.edu
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: md5's
Posted by Brett Porter <br...@apache.org>.
If you want to put something to this effect into JIRA I could probably
do this. I'm not sure how well it might work with the FTP deployer, but
that doesn't apply to Apache so it can be omitted there if troublesome.
Thanks,
Brett
Henk P. Penning wrote:
>On Wed, 6 Oct 2004, Mark R. Diggory wrote:
>
>
>
>>Date: Wed, 06 Oct 2004 14:08:38 -0400
>>From: Mark R. Diggory <md...@apache.org>
>>To: Dion Gillard <di...@gmail.com>
>>Cc: Henk P. Penning <he...@cs.uu.nl>, apmirror@apache.org,
>> repository@apache.org, Maven Developers List <de...@maven.apache.org>
>>Subject: Re: md5's
>>
>>
>
>
>
>>Ultimately it would, as well, be of great benefit if the artifact
>>publishing worked initially in a staging area as not to confuse existing
>> cron jobs attempting to validate md5's on the server. This would
>>probably be best served in the systems default tmp directory.
>>
>>
>
> That would be great.
>
> I think, the best way for adding/replace stuff is
>
> -- write a 'temp'
> -- rename 'temp' to 'file'
>
> because a rename is truly atomic if 'temp' and 'file' are
> in the same file system.
>
> If you can implement the 'temp' for 'file' to be,
> for instance, '.tmp.file', I can easily teach the checkers
> to ignore '.tmp.*' files. I think rsync does something
> like that (even better .tmp.$$.file).
>
> Just a thought.
>
>
>
>>-Mark
>>
>>
>
> HPP
>
>---------------------------------------------------------------- _
>Henk P. Penning, Computer Systems Group R Uithof CGN-A232 _/ \_
>Dept of Computer Science, Utrecht University T +31 30 253 4106 / \_/ \
>Padualaan 14, 3584CH Utrecht, the Netherlands F +31 30 251 3791 \_/ \_/
>http://www.cs.uu.nl/staff/henkp.html M penning@cs.uu.nl \_/
>
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>For additional commands, e-mail: dev-help@maven.apache.org
>
>
>
>
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: md5's
Posted by "Mark R. Diggory" <md...@apache.org>.
The ball is in play.
http://jira.codehaus.org/browse/MPARTIFACT-39
-Mark
Henk P. Penning wrote:
> On Wed, 6 Oct 2004, Mark R. Diggory wrote:
>
>
>>Date: Wed, 06 Oct 2004 14:08:38 -0400
>>From: Mark R. Diggory <md...@apache.org>
>>To: Dion Gillard <di...@gmail.com>
>>Cc: Henk P. Penning <he...@cs.uu.nl>, apmirror@apache.org,
>> repository@apache.org, Maven Developers List <de...@maven.apache.org>
>>Subject: Re: md5's
>
>
>>Ultimately it would, as well, be of great benefit if the artifact
>>publishing worked initially in a staging area as not to confuse existing
>> cron jobs attempting to validate md5's on the server. This would
>>probably be best served in the systems default tmp directory.
>
>
> That would be great.
>
> I think, the best way for adding/replace stuff is
>
> -- write a 'temp'
> -- rename 'temp' to 'file'
>
> because a rename is truly atomic if 'temp' and 'file' are
> in the same file system.
>
> If you can implement the 'temp' for 'file' to be,
> for instance, '.tmp.file', I can easily teach the checkers
> to ignore '.tmp.*' files. I think rsync does something
> like that (even better .tmp.$$.file).
>
> Just a thought.
>
>
>>-Mark
>
>
> HPP
>
> ---------------------------------------------------------------- _
> Henk P. Penning, Computer Systems Group R Uithof CGN-A232 _/ \_
> Dept of Computer Science, Utrecht University T +31 30 253 4106 / \_/ \
> Padualaan 14, 3584CH Utrecht, the Netherlands F +31 30 251 3791 \_/ \_/
> http://www.cs.uu.nl/staff/henkp.html M penning@cs.uu.nl \_/
>
--
Mark Diggory
Open Source Software Developer
Apache Jakarta Project
http://jakarta.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: md5's
Posted by "Mark R. Diggory" <md...@apache.org>.
The ball is in play.
http://jira.codehaus.org/browse/MPARTIFACT-39
-Mark
Henk P. Penning wrote:
> On Wed, 6 Oct 2004, Mark R. Diggory wrote:
>
>
>>Date: Wed, 06 Oct 2004 14:08:38 -0400
>>From: Mark R. Diggory <md...@apache.org>
>>To: Dion Gillard <di...@gmail.com>
>>Cc: Henk P. Penning <he...@cs.uu.nl>, apmirror@apache.org,
>> repository@apache.org, Maven Developers List <de...@maven.apache.org>
>>Subject: Re: md5's
>
>
>>Ultimately it would, as well, be of great benefit if the artifact
>>publishing worked initially in a staging area as not to confuse existing
>> cron jobs attempting to validate md5's on the server. This would
>>probably be best served in the systems default tmp directory.
>
>
> That would be great.
>
> I think, the best way for adding/replace stuff is
>
> -- write a 'temp'
> -- rename 'temp' to 'file'
>
> because a rename is truly atomic if 'temp' and 'file' are
> in the same file system.
>
> If you can implement the 'temp' for 'file' to be,
> for instance, '.tmp.file', I can easily teach the checkers
> to ignore '.tmp.*' files. I think rsync does something
> like that (even better .tmp.$$.file).
>
> Just a thought.
>
>
>>-Mark
>
>
> HPP
>
> ---------------------------------------------------------------- _
> Henk P. Penning, Computer Systems Group R Uithof CGN-A232 _/ \_
> Dept of Computer Science, Utrecht University T +31 30 253 4106 / \_/ \
> Padualaan 14, 3584CH Utrecht, the Netherlands F +31 30 251 3791 \_/ \_/
> http://www.cs.uu.nl/staff/henkp.html M penning@cs.uu.nl \_/
>
--
Mark Diggory
Open Source Software Developer
Apache Jakarta Project
http://jakarta.apache.org
Re: md5's
Posted by "Henk P. Penning" <he...@cs.uu.nl>.
On Wed, 6 Oct 2004, Mark R. Diggory wrote:
> Date: Wed, 06 Oct 2004 14:08:38 -0400
> From: Mark R. Diggory <md...@apache.org>
> To: Dion Gillard <di...@gmail.com>
> Cc: Henk P. Penning <he...@cs.uu.nl>, apmirror@apache.org,
> repository@apache.org, Maven Developers List <de...@maven.apache.org>
> Subject: Re: md5's
> Ultimately it would, as well, be of great benefit if the artifact
> publishing worked initially in a staging area as not to confuse existing
> cron jobs attempting to validate md5's on the server. This would
> probably be best served in the systems default tmp directory.
That would be great.
I think, the best way for adding/replace stuff is
-- write a 'temp'
-- rename 'temp' to 'file'
because a rename is truly atomic if 'temp' and 'file' are
in the same file system.
If you can implement the 'temp' for 'file' to be,
for instance, '.tmp.file', I can easily teach the checkers
to ignore '.tmp.*' files. I think rsync does something
like that (even better .tmp.$$.file).
Just a thought.
> -Mark
HPP
---------------------------------------------------------------- _
Henk P. Penning, Computer Systems Group R Uithof CGN-A232 _/ \_
Dept of Computer Science, Utrecht University T +31 30 253 4106 / \_/ \
Padualaan 14, 3584CH Utrecht, the Netherlands F +31 30 251 3791 \_/ \_/
http://www.cs.uu.nl/staff/henkp.html M penning@cs.uu.nl \_/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: md5's
Posted by "Henk P. Penning" <he...@cs.uu.nl>.
On Wed, 6 Oct 2004, Mark R. Diggory wrote:
> Date: Wed, 06 Oct 2004 14:08:38 -0400
> From: Mark R. Diggory <md...@apache.org>
> To: Dion Gillard <di...@gmail.com>
> Cc: Henk P. Penning <he...@cs.uu.nl>, apmirror@apache.org,
> repository@apache.org, Maven Developers List <de...@maven.apache.org>
> Subject: Re: md5's
> Ultimately it would, as well, be of great benefit if the artifact
> publishing worked initially in a staging area as not to confuse existing
> cron jobs attempting to validate md5's on the server. This would
> probably be best served in the systems default tmp directory.
That would be great.
I think, the best way for adding/replace stuff is
-- write a 'temp'
-- rename 'temp' to 'file'
because a rename is truly atomic if 'temp' and 'file' are
in the same file system.
If you can implement the 'temp' for 'file' to be,
for instance, '.tmp.file', I can easily teach the checkers
to ignore '.tmp.*' files. I think rsync does something
like that (even better .tmp.$$.file).
Just a thought.
> -Mark
HPP
---------------------------------------------------------------- _
Henk P. Penning, Computer Systems Group R Uithof CGN-A232 _/ \_
Dept of Computer Science, Utrecht University T +31 30 253 4106 / \_/ \
Padualaan 14, 3584CH Utrecht, the Netherlands F +31 30 251 3791 \_/ \_/
http://www.cs.uu.nl/staff/henkp.html M penning@cs.uu.nl \_/