ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Peter B. West" <>
Subject Re: cvs commit: jakarta-ant/src/main/org/apache/tools/ant/taskdefs
Date Mon, 06 Aug 2001 14:27:40 GMT

Stefan Bodewig wrote:
> On Sun, 05 Aug 2001, Peter B. West <> wrote:
>>When I made the changes to accommodate the large files, I forgot to
>>tests change-in-place.  Pretty disgraceful.  Mea culpa.
> I had modified your patch to take care of this (using a temporary file
> and moving it over when done) - but I see Conor has committed your
> newer patch now, I'll have to take a second look, as it reintroduces
> some stuff that I had modified before.

There must have been some merge butchering when I did my cvs update.  I 
saw your temporary file stuff and simply added the change comparisons to 
it.  In my merged file I had no definition of the "addcr" int which was 
still being tested later in the file, and I ended up with two FixCR 
constructors.  Probably from merging against one of my changed files.

One day I'll work out how to maintain local RCS files for keeping track 
if my work in progress.

>>I have modified the code I found in the repository, which did not
>>compile for me.
> Strange - I also tested change-in-place with this (Ant's own bootstrap
> build performs such beats).
>>I will now go back to providing a minimum set of changes to the
>>HTML, as requested.
> Seen it, thanks.
> Stefan

I realized when preparing the html that this code does not strip 
trailing white space, and I put a note to that effect in the doc.  It 
may also be worthwhile mentioning this in the release notes.  Users 
post-processing their sources may be surprised by it.

I make the assertion in the docs that, where both cr and eol are 
specified, eol takes precedence.  I confess to being both a Java and an 
ant neophyte, but I don't see how the modified code guarantees this 

I appreciate your patience on this.

Peter B. West
"Lord, to whom shall we go?"

View raw message