Does MQ support compressed patch files?

[Peter Williams]

On 07/04/11 18:33, timeless wrote:
> On Thu, Apr 7, 2011 at 5:15 AM, Peter Williams<pwil3058 at bigpond.net.au>  wrote:
>> Quilt (upon which MQ was originally based)
>
> I think "after which MQ was modeled"
>
>> supported compressed patches based on the extension to the patch name
>> (e.g. patch.gz would be compressed using gzip).
>>   Does MQ support this functionality?
>
>> PS Just asking NOT suggesting/requesting change.
>
> So, beyond the case of monolithic patches to the Linux Kernel or to
> Mozilla or OpenOffice, are there use cases for this feature?

What part of "NOT suggesting/requesting change" did you not understand? :-)

But, anyway, I'd say that it's a hangover from the days when disk space 
was at a premium and minimising the size of the files in the "patches" 
directory may have been useful to users.

>
> I could imagine having<qimport>  support .gz and perhaps a couple of
> others - where support means "extracts the file leaving the expanded
> human readable form".

Probably not worth it because when quilt emails compressed patches (the 
most likely way to receive a patch for import) it sends them 
uncompressed (just like MQ).  And even if you have a compressed patch 
you can easily decompress it before importing it.

>
> If you're dealing w/ a DVCS that has enough files or enough large
> files that you might want to compress patch to modify it, I'm fairly
> confident that your DVCS history (.hg/store ?) will significantly
> dwarf the size of your patches. Which means that the only reason to
> use compression should be for transmission.
>
> Note that MQ also can be manged by Hg (and I in fact sometimes have
> <project>/.hg/patches/.hg/patches/.hg/patches). If the content in an
> MQ is compressed on disk, then you break Hg managed MQs. I've seen
> hints that GIT supports some magic filters which might let you
> sometimes do interesting things w/ archives, but I'm not sure if even
> it could do something particularly useful in the same vein as an Hg
> managed MQ given the constraint that the files in the queue are
> compressed.
>

I'm not sure I understand the relevance of that paragraph but if the aim 
is to argue against including this feature in MQ I agree.

Peter