Recommended approach for converting from CVSNT to Mercurial?

[greg fenton]

Is there at least a way to persist the conversion done to a certain
pass and redo that pass if it fails?

My timings for the passes so far on the failed module are:

Time for pass1 (CollectRevsPass): 2396 seconds.
Time for pass2 (CleanMetadataPass): 1.037 seconds.
Time for pass3 (CollateSymbolsPass): 0.397 seconds.
Time for pass4 (FilterSymbolsPass): 1150 seconds.
Time for pass5 (SortRevisionSummaryPass): 1.794 seconds.
Time for pass6 (SortSymbolSummaryPass): 66.35 seconds.
Time for pass7 (InitializeChangesetsPass): 692.9 seconds.
Time for pass8 (BreakRevisionChangesetCyclesPass): 64.46 seconds.
Time for pass9 (RevisionTopologicalSortPass): 65.21 seconds.
Time for pass10 (BreakSymbolChangesetCyclesPass): 476.2 seconds.
Time for pass11 (BreakAllChangesetCyclesPass): 14801 seconds.
Time for pass12 (TopologicalSortPass): 501.6 seconds.
Time for pass13 (CreateRevsPass): 398.9 seconds.
Time for pass14 (SortSymbolsPass): 195.5 seconds.
Time for pass15 (IndexSymbolsPass): 18.63 seconds.


So I'd like to persist those 15 passes and simply re-run pass 16.  I
see that there is a -p parameter, but I'm not sure what files (e.g.
cvs2svn-tmp) need to be cleaned up before trying the re-run of pass 16
(assuming it is possible).

Thanks,
greg.fenton
--
greg.fenton ^_at_^ gmail (yes, dot com)



On Fri, Jan 15, 2010 at 09:41, greg fenton <greg.fenton at gmail.com> wrote:
> I hopped on the #cvs2svn IRC channel last night and was told that this
> approach (limit conversion to a range of revisions/changesets) does
> not exist in that tool.
>
> Not sure what the next step is for debugging this conversion failure.
> I'm not a python guy, more of a perl guy (yea, yea...).  :)
>
> I'd appreciate any insight or educated guesses.  Thanks!
> greg.fenton
> --
> greg.fenton ^_at_^ gmail (yes, dot com)
>
>
>
> On Thu, Jan 14, 2010 at 23:08, Steve Borho <steve at borho.org> wrote:
>> On Thu, Jan 14, 2010 at 9:54 PM, greg fenton <greg.fenton at gmail.com> wrote:
>>> Hi Steve,
>>>
>>> Sorry, but it isn't clear to me what option(s) I can use to limit the
>>> conversion to a number of changesets.  I'd like to run this in a loop
>>> doing, say, 1000 at a time.
>>
>> with hg convert, it's a simple -r option.  I don't know about cvs2hg, though.
>>
>> Greg Ward?
>>
>> --
>> Steve Borho
>>
>