This is the mail archive of the
cygwin-apps@cygwin.com
mailing list for the Cygwin project.
Re: rebase / STL set patch
- From: Jason Tishler <jason at tishler dot net>
- To: cygwin-apps at cygwin dot com
- Date: Mon, 16 Sep 2002 14:40:29 -0400
- Subject: Re: rebase / STL set patch
- References: <20020630134832.GA1904@tishler.net><023001c22078$d26c9150$1800a8c0@LAPTOP> <20020701123502.GB1904@tishler.net><009001c220fd$4ff90610$1800a8c0@LAPTOP> <20020702212203.GC1776@tishler.net><039b01c22219$f796fef0$1800a8c0@LAPTOP> <20020731124402.GB1444@tishler.net><1028131377.2788.32.camel@lifelesswks> <20020827171043.GA508@tishler.net><1030522625.13255.440.camel@lifelesswks>
Rob,
Before I went on vacation, I had a chance to accomplish the following:
On Wed, Aug 28, 2002 at 06:17:04PM +1000, Robert Collins wrote:
> A bit of feedback... I had to violate the Free/UsedList abstraction
> layer to do the memory dumping. This is because they aren't exposing
> iterators themselves.
I added the iterators to the above.
> I suggest that you make a RebaseState class that contains both the free
> list and the used list as containers:
I followed the above suggestion. Besides having the benefits you
mentioned, it will allowing me to easily add a skip list in the future.
The skip list would be a list of DLLs not to rebase (e.g., cygwin1.dll).
> As I mentioned in my earlier emails though, it actually makes sense to
> have RebaseState inherit from RebaseBuilder, and implement the building
> interface itself - because it's already decoupled from the storage
> mechanism.
I followed the above suggestion too. Additionally, I added a dump
method to RebaseState so it can direct RebaseConfigBuilder to save the
its state back to a file.
Well, I can read and write configuration files now -- not too bad for 9
months! :,)
Jason