Issue 6996 - Fix -c option

Issue 6996 - Fix -c option

Repository: https://hg.adblockplus.org/codingtools
Base revision: a4d31ca88c6e

[Vasily Kuznetsov, 2018-10-01 15:39:39]

Hi Jon,
Here's a fix for your ticket with an awesome number ;)

Tristan, Sebastian, FYI.

Cheers,
Vasily

[hub, 2018-10-01 16:06:14]

https://codereview.adblockplus.org/29897571/diff/29897572/hgreview.py
File hgreview.py (right):

https://codereview.adblockplus.org/29897571/diff/29897572/hgreview.py#newcode58
hgreview.py:58: raise error.Abort('Argument of --change must be an integer')
I'm not sure I understand correctly. Do you bail if the parameter to -c isn't an
int?

So what if one pass `-c .` (I do that all the time). Also changset id are sha1.
So how does this work?

[Vasily Kuznetsov, 2018-10-01 16:32:52]

https://codereview.adblockplus.org/29897571/diff/29897572/hgreview.py
File hgreview.py (right):

https://codereview.adblockplus.org/29897571/diff/29897572/hgreview.py#newcode58
hgreview.py:58: raise error.Abort('Argument of --change must be an integer')
On 2018/10/01 16:06:14, hub wrote:
> I'm not sure I understand correctly. Do you bail if the parameter to -c isn't
an
> int?
> 
> So what if one pass `-c .` (I do that all the time). Also changset id are
sha1.
> So how does this work?

Yeah, it seems that repo[x] kind of expects x to be an int (a local revision
number that hg uses). I thought that's the normal use case that everyone has
based on the error message in the ticket.

Sha1 hashes  don't work out of the box (I tried), but I can try to figure out
how to make them work.

[Vasily Kuznetsov, 2018-10-02 10:06:44]

Hi Hub,

Check out the new patchset, this should work with all kinds of revision ids.

Cheers,
Vasily

https://codereview.adblockplus.org/29897571/diff/29897572/hgreview.py
File hgreview.py (right):

https://codereview.adblockplus.org/29897571/diff/29897572/hgreview.py#newcode58
hgreview.py:58: raise error.Abort('Argument of --change must be an integer')
On 2018/10/01 16:32:51, Vasily Kuznetsov wrote:
> On 2018/10/01 16:06:14, hub wrote:
> > I'm not sure I understand correctly. Do you bail if the parameter to -c
isn't
> an
> > int?
> > 
> > So what if one pass `-c .` (I do that all the time). Also changset id are
> sha1.
> > So how does this work?
> 
> Yeah, it seems that repo[x] kind of expects x to be an int (a local revision
> number that hg uses). I thought that's the normal use case that everyone has
> based on the error message in the ticket.
> 
> Sha1 hashes  don't work out of the box (I tried), but I can try to figure out
> how to make them work.

I switched to using repo.revs() API. It's not documented here:
https://www.mercurial-scm.org/wiki/MercurialApi#Repositories but it works with
all kinds of revision ids: hashes, branches, bookmarks, tags and local
sequential numbers. It also produces nice looking error messages when the
revision is not found -- without us having to do anything.

Not sure if there's a better way but if there is, it's not in the documentation.

[Vasily Kuznetsov, 2018-10-05 11:02:58]

Hi guys,

Can someone check if this patch fixes the issue for them (it does for me) and
LGTM it?

Cheers,
Vasily

[Sebastian Noack, 2018-10-05 11:04:25]

LGTM