[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#715216: qa.debian.org: collab-qa/upload-history: Software trusts "Date" headers which are sometimes set wrong



On 07/07/13 at 02:10 -0400, Asheesh Laroia wrote:
> New proposed fix:
> 
> In collab-qa/upload-history/munge_ddc.py , if the Message-Date we
> were going to emit is not within the year of the envelope From,
> plus/minus one year, we drop Message-Date.
> 
> This permits the udd/upload_history_gatherer.py code to use its
> existing logic about dropping replacing "Message-Date" with "Date"
> when "Message-Date" is visible.
> 
> I've tested that on a subset of debian-devel-changes.201307 and
> confirmed the patched code has identical output, and that the crazy
> "The year is 2019" message has N/A as its Message-Date. So I believe
> the behavior is correct.
> 
> I've also tested the performance of this. I re-ran the patched +
> unpatched code until runtimes were close to consistent (in an
> attempt to remove disk or other caches as the difference) and then I
> got these numbers after one run of each:
> 
> Unpatched: 22.994sec
> 
> ➜  upload-history  time python munge_ddc.py
> /tmp/debian-devel-changes/debian-devel-changes.201307 >
> /tmp/debian-devel-changes.201307.out.without-patch
> 
> 17.93s user 4.59s system 97% cpu 22.994 total
> 
> Patched: 22.983sec
> 
> ➜  upload-history  time python munge_ddc.py
> /tmp/debian-devel-changes/debian-devel-changes.201307 >
> /tmp/debian-devel-changes.201307.out.with-patch
> 
> 18.00s user 4.52s system 97% cpu 22.983 total
> 
> 
> I'm not realistically arguing that my code runs faster than the
> original code, but I believe there is less than 1% runtime
> difference between the two, and probably way less than 1%.
> 
> Patch attached for munge_ddc.py. Please review and merge, if you like!

Hi Asheesh,

Sorry for the delayed reply. Thanks a lot for your work on this.

I've added you to collab-qa, so you should be able to push your code
yourself.

Now, one comment on your changes: it would be better if everything you
do is versioned in git (for example,
/srv/udd.debian.org/email-archives/README isn't, same for scripts in
/srv/udd.debian.org/upload-history/). You could put them under
/srv/udd.debian.org/udd/scripts/upload-history/, for example.

Lucas


Reply to: