regression in mq: directory names are uppercased since hg 2.0.2 and thg 2.2.2

Dominik Psenner dpsenner at gmail.com
Fri Feb 10 15:39:30 UTC 2012


I investigated a little further into this problem and this is the result:

The symptoms were reproducible by doing a simple "hg add" / "hg forget"
cycle with thg. While doing the equivalent in console - thus calling
mercurial directly - worked just fine. So this seems to be a problem within
thg for at least versions > 2.1.4.

Version              | thg    | console
----------------------------------------
Thg 2.2.2 (hg 2.0.2) | broken | works
Thg 2.1.4 (hg 1.9.3) | works  | works

hg, nore mq do not seem to be innocent.

Cheers

>-----Original Message-----
>From: Dominik Psenner [mailto:dpsenner at gmail.com]
>Sent: Friday, February 10, 2012 11:00 AM
>To: mercurial-devel at selenic.com; thg-dev at googlegroups.com
>Subject: regression in mq: directory names are uppercased since hg 2.0.2
>and thg 2.2.2
>
>Hi,
>
>I've ran into a nasty bug with mq on thg 2.2.2 / mercurial 2.0.2 and I'm
>unsure whether it is thg or mercurial related. Therefore I cross post it on
>both mailing lists.
>
>The problem is that when files are added to a patch, sometimes parts of
>their path become uppercased. For example while doing the equivalent of the
>following by using thg:
>
>$ mkdir Modules/API/Set
>$ touch Modules/API/Set/ASetFormula.cs
>$ hg add Modules/API/Set/ASetFormula.cs
>$ hg qnew
>
>The patch looks like:
>
>diff -r 4ca10dbdf0ba Modules/API/SET/ASetFormula.cs
>--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
>+++ b/Modules/API/SET/ASetFormula.cs	Fri Feb 10 10:44:25 2012 +0100
>
>Instead of:
>
>diff -r 4ca10dbdf0ba Modules/API/Set/ASetFormula.cs
>--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
>+++ b/Modules/API/Set/ASetFormula.cs	Fri Feb 10 10:44:25 2012 +0100
>
>The outcome can be fixed by doing:
>
>$ hg qpop
>$ edit patch manually and fix casing
>$ hg qpush
>
>But with the next qrefresh the casing is again broken. This did not happen
>before thg 2.2.2 / mercurial 2.0.2. Did someone else run into this problem
>before?
>
>Cheers,
>Dominik





More information about the Mercurial-devel mailing list