Important: RC_1_33_0 -> HEAD merge
Now that 1.33.1 is out and the RC_1_33_0 branch is most likely closed for good, I've merged many of the changes from the RC_1_33_0 branch back to HEAD so that we don't lose all of the minor tweaks we made at the end. For many changes, the merge was so trivial that I've just committed the results. However, some libraries seemed to get quite out-of-sync. This happens when changes were made on both HEAD and RC_1_33_0, and CVS can't figure out how to merge those changes. These libraries are: Date-Time Iostreams Parameter Pointer containers Range Regex Serialization uBLAS The authors of these libraries should look at the merged results to determine what changes need to be made in HEAD. Otherwise, fixes that we have done for 1.33.1 will be forgotten and lost. To do so, have a clean checkout of CVS HEAD available and run something like this: cvs -z6 update -j Version_1_33_0 -j Version_1_33_1 boost/iostreams libs/iostreams (You may want to redirect the output of that command to a file; there's a lot of information in there) Look for "conflicts" in the output to see went wrong, then dig through conflicting files looking for sections between <<<<< and ==== (HEAD code) and ===== and >>>>> (RC_1_33_0 code). Once you've cleaned up everything, commit back to head and you'll be done. Doug
Doug Gregor wrote:
Now that 1.33.1 is out and the RC_1_33_0 branch is most likely closed for good, I've merged many of the changes from the RC_1_33_0 branch back to HEAD so that we don't lose all of the minor tweaks we made at the end. For many changes, the merge was so trivial that I've just committed the results. However, some libraries seemed to get quite out-of-sync. This happens when changes were made on both HEAD and RC_1_33_0, and CVS can't figure out how to merge those changes. These libraries are:
Date-Time Iostreams
It doesn't look too bad, but I won't have time to fix it for a day or two. -- Jonathan Turkanis www.kangaroologic.com
Doug Gregor <doug.gregor <at> gmail.com> writes:
Now that 1.33.1 is out and the RC_1_33_0 branch is most likely closed for good, I've merged many of the changes from the RC_1_33_0 branch back to HEAD so that we don't lose all of the minor tweaks we made at the end...
I've just checked it out, and there're some recently modified files in libs/multi_index/ that didn't get merged from RC_1_33_0 to HEAD (for an example, see http://tinyurl.com/cznm2) Is this intentional or just an overlook? I can do the merge myself, but I'm afraid there could be some more unintended leftovers in other libs and nobody would notice. Joaquín M López Muñoz Telefónica, Investigación y Desarrollo
On Dec 9, 2005, at 1:37 PM, Joaquin M Lopez Munoz wrote:
Doug Gregor <doug.gregor <at> gmail.com> writes:
Now that 1.33.1 is out and the RC_1_33_0 branch is most likely closed for good, I've merged many of the changes from the RC_1_33_0 branch back to HEAD so that we don't lose all of the minor tweaks we made at the end...
I've just checked it out, and there're some recently modified files in libs/multi_index/ that didn't get merged from RC_1_33_0 to HEAD (for an example, see http://tinyurl.com/cznm2) Is this intentional or just an overlook? I can do the merge myself, but I'm afraid there could be some more unintended leftovers in other libs and nobody would notice.
I tried to point out all of the things that I wasn't able to merge (because of conflicts), but obviously I missed something. Sorry! Can you handle the merge for multi_index? Doug
Douglas Gregor ha escrito:
On Dec 9, 2005, at 1:37 PM, Joaquin M Lopez Munoz wrote:
Doug Gregor <doug.gregor <at> gmail.com> writes:
Now that 1.33.1 is out and the RC_1_33_0 branch is most likely closed for good, I've merged many of the changes from the RC_1_33_0 branch back to HEAD so that we don't lose all of the minor tweaks we made at the end...
I've just checked it out, and there're some recently modified files in libs/multi_index/ that didn't get merged from RC_1_33_0 to HEAD (for an example, see http://tinyurl.com/cznm2) Is this intentional or just an overlook? I can do the merge myself, but I'm afraid there could be some more unintended leftovers in other libs and nobody would notice.
I tried to point out all of the things that I wasn't able to merge (because of conflicts), but obviously I missed something. Sorry!
Nothing to be sorry about. Thank you for managing such a neverending release!
Can you handle the merge for multi_index?
It's already done. FYI, I didn't have any conflicts, I cannot guess what stopped you from merging on your first try. Are we sure everything else is already merged? Best, Joaquín M López Muñoz Telefónica, Investigación y Desarrollo
Doug Gregor wrote:
Now that 1.33.1 is out and the RC_1_33_0 branch is most likely closed for good, I've merged many of the changes from the RC_1_33_0 branch back to HEAD so that we don't lose all of the minor tweaks we made at the end. For many changes, the merge was so trivial that I've just committed the results. However, some libraries seemed to get quite out-of-sync. This happens when changes were made on both HEAD and RC_1_33_0, and CVS can't figure out how to merge those changes. These libraries are:
Date-Time Iostreams Parameter Pointer containers Range Regex Serialization uBLAS
The authors of these libraries should look at the merged results to determine what changes need to be made in HEAD. Otherwise, fixes that we have done for 1.33.1 will be forgotten and lost. To do so, have a clean checkout of CVS HEAD available and run something like this:
cvs -z6 update -j Version_1_33_0 -j Version_1_33_1 boost/iostreams libs/iostreams
Done for Range and Pointer Containers. There was no merge conflict...so it was quite easy....let's see if the regression test stays stable :-) Thanks for managing the release, Doug. -Thorsten
On Fri, 09 Dec 2005 23:36:42 +0100, Thorsten Ottosen wrote
Doug Gregor wrote:
However, some libraries seemed to get quite out-of-sync. This happens when changes were made on both HEAD and RC_1_33_0, and CVS can't figure out how to merge those changes. These libraries are:
Date-Time
I've already pulled all changes into the head for date-time. Jeff
Now that 1.33.1 is out and the RC_1_33_0 branch is most likely closed for good, I've merged many of the changes from the RC_1_33_0 branch back to HEAD so that we don't lose all of the minor tweaks we made at the end. For many changes, the merge was so trivial that I've just committed the results. However, some libraries seemed to get quite out-of-sync. This happens when changes were made on both HEAD and RC_1_33_0, and CVS can't figure out how to merge those changes. These libraries are:
Date-Time Iostreams Parameter Pointer containers Range Regex Serialization uBLAS
The authors of these libraries should look at the merged results to determine what changes need to be made in HEAD. Otherwise, fixes that we have done for 1.33.1 will be forgotten and lost. To do so, have a clean checkout of CVS HEAD available and run something like this:
I've checked the diff's and all the changes for regex are in the HEAD branch already. John.
participants (8)
-
Doug Gregor -
Douglas Gregor
-
Jeff Garland -
Joaquin M Lopez Munoz -
Joaquín Mª López Muñoz
-
John Maddock -
Jonathan Turkanis -
Thorsten Ottosen