[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: how to use one bug ID syntax for two different butracking-pools

From: Arwin Arni Nandagopal <arwin_at_collab.net>
Date: Wed, 12 Feb 2014 14:54:39 +0530

You needn't dump/load.

Just run svn propget --revprop svn:log -r$N URL > r$N.log in a for loop
running from 0 to HEAD.

Run your sed script on all the files.

Then svn propset --revprop -r$N svn:log -F r$N.log in a similar for loop.

Regards,
Arwin Arni

On Wed, Feb 12, 2014 at 2:36 AM, Stefan <luke1410_at_gmx.de> wrote:

> Hi,
>
> I'm not sure whether TSVN mailing list would be more appropriate for the
> actual use-case-related part of the question. Hence I'm trying to keep it
> more SVN-specific:
>
> Is there a way in SVN to perform some kind of bulk-changes or
> replace-operation on all committed log-messages?
> What I have in mind is replacing all occurrances in all log messages of
> the style \#(\d+) with XT\-($1) where $1 is the captured number in the
> previous string and the \d would represent a number-value.
> In other words:
> "Foo #2354."
> Should become:
> "Foo XT-2354."
>
> The only idea I get here is to perform a dump/load-cycle and before the
> load, perform a regexp-replacement on the dump-file.
> Anybody has a better solution?
>
> Underlying usecase (maybe someone else comes up with a better solution for
> the actual issue):
> We make use of TSVN's bugtraq-properties (see:
> http://tortoisesvn.net/docs/release/TortoiseSVN_en/tsvn-
> dug-bugtracker.html).
> Up to now we used the syntax: "#XXXX" which then corresponded to issue
> number XXXX and refrenced a link to: http://foo.com/bar.php?query=XT-XXXX.
> XT is a unique-ID in the bugtracker.
>
> Now we moved to use two separate projects in the bugtracker, both relying
> on the same code base. The assigned unique IDs are XT- and XR-.
> All existing IDs used XT-, while all new ones will use the XR-. The
> expectation for commits in the future are that these will only reference
> XR-prefixes.
>
> In a perfect world, I'd like developers to offer two ways to write their
> ID-references:
> - use the #XXXX-syntax, which will then default to XR-XXXX
> - use either XT-XXXX or XR-XXXX syntaxes, if they want to explicitly
> reference either an issue for XT or XR.
> All existing syntaxes of #XXXX should of course still use the
> XT-XXXX-references.
>
> My problem is that the bugtraq-properties only support a single regexp to
> extract the actual relevant bug-message and then from that extract the
> bugID which is then used to create the link to the bugtracker.
>
> So what I can do is:
> - enable the new XT-/XR-syntax OR (problem: the #-syntax won't work at all)
> - change the #-syntax and make it refer the XR-issues (problem:
> XT-/XR-syntax is not possible, old #-issues will incorrectly reference the
> XR-issues (even though these were created for the XT-issues)
> - keep the #-syntax refere to the XT-issues (problem: XR-issues cannot be
> referenced)
> - allow the #, XT- and XR-syntax, pass it on to some wrapper CGI-script
> (aka: http://foo.com/customwrapper.cgi?query=#XXX|XT-XXX|XR-XXX) and let
> that CGI-script resolve the actual reference. (problem: hard to set-up in
> our environment, does result in inconsistent issue-naming in TSVN, since in
> one case the issue is #XXXX while in the other case issues are XT-XXXX or
> XR-XXXX).
>
> Any idea/suggestions on the best practice to approach this problem?
>
> Regards,
> Stefan
>
> ---
> This email is free from viruses and malware because avast! Antivirus
> protection is active.
> http://www.avast.com
>
>
Received on 2014-02-12 10:25:14 CET

This is an archived mail posted to the Subversion Users mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.