2 - lots of fixes to get the thing working sufficienly for cvs2git
4 - optionally detect and display ancestor branch information for
6 - never allow more than one revision of a file to be in a single
7 patchset (this may disturb the patchset numbers for historical
9 - use file revisions to help sort, so that initial imports
10 won't be listed as patchset 1, with patchset 2 as the actual
11 creation of the files.
12 - reports log times in localtime (use TZ to override)
13 - many downstream patches from Debian, special thanks to Marcus
14 Crafter for maintaining. From the Debian changelog:
15 - patch from Kim Hansen to fix time calculation bug
16 - documentation patch from J. Bruce Fields
17 - directory name handling patch from David Kilzer
18 - others I'm not sure of, I'd happily update the credit here
21 - resolve -r symbols during global symbol resolution
22 - using two -r arguments implicitly sets -b with branch
23 from second -r (if you can speak cvsps arguments, you'll
25 - fix using cvsps against sourceforge where the actual
26 server repository path has many instances of the project
27 name in it and the 'strstr' was finding the wrong one.
28 instead use the equiv. of 'strrstr'
31 - fix bug where filenames with 'bad' characters (such as ' ')
32 were being sent unescaped to a shell
33 - add --no-cvs-direct and -Z 0 to disable cvs-direct and compression
34 - extend the 'fuzz-factor' concept by keeping a min/max date
35 encountered for each patchset and fuzz on this interval,
36 rather than fuzzing on the date of the first encountered
37 member only, helps for handling REALLY slow commits where a
38 HUGE number of individual files were modified at a slow trickle.
39 - it turns out that old versions of cvs have two problems
40 1) they don't support 'rlog' (handled)
41 2) they don't support 'version', which caused the workaround
43 Now cvsps detects cvs versions too old to support version and/or
44 rlog. Additionally, if server is too old, cvs-direct is
46 - parse the rc file first so that the command line will override
47 - add -q to shut-up the warnings
50 - remove Mbinary from valid-response list - can't handle it
51 - much more (pedantic?) clean-up/close path in cvs_direct
52 to fix all of the lost data and hanging problems at close
53 - added rlog capability to cvs_direct, switched main code
55 - added version capability to cvs_direct, now cvsps is
56 completely independant of cvs binary for all operations!
57 - added --root and repository command-line arguments,
58 now cvsps is completely indepentant of having working
62 - fix the cvs_rupdate calls to use the real repository
63 instead of the mangled use_rep_path. This broke
64 cvs_direct generation of add/remove diffs
65 - fix the sed expression to account for the difference
66 between +++ and --- in the first and second lines.
67 This broke remove diffs for direct and external.
68 - attempt to fix a hang-at-flush bug in cvs-direct
72 - implement compression in cvs_direct
73 - add command line -Z to specify compression level.
74 applies to cvs-direct and external cvs
75 - implement 'cvs_diff' in cvs-direct
76 - use cvs_diff instead of cvs_rdiff because it allows
77 us to specify all of the lovely diff options
78 - change system to my_system so ctrl-c will get
79 through. indirectly, this is necessary because
80 cvs has loused up the exit codes with the 'cvs diff'
81 command, making it impossible to determine whether
82 the program exited abnormally
83 - add parsing of ~/.cvsps/cvspsrc file at startup
84 where command line arguments can be listed, one
85 per line (parameters to arguments on the same line
86 as the argument). --norc prevents this
89 - fix bug where if -r was specified along with other
90 filter options, the other filter options were
91 applied first, and the start/end tags were
93 - change -f from strstr to regex
94 - change regex to EXTENDED regex
95 - add the --cvs-direct option
98 - add a cvs_check_cap for capability checking.
99 - use cvs_check_cap to check whether rlog is
100 supported. it's not supported in 1.11 and
101 before. warn if non-supported version is
103 - add --no-rcmds to manually disable rlog and rdiff
104 - add a fix for some bizarre inconsistencies found
105 in some random repositories, such as tagged
106 revisions that don't exist.
107 - remove the adaptive crap filter and bk_log_border
108 logic now that Larry has cleaned up the exported
112 - add the bk_log_border parsing logic to handle the
113 case where 'cvs log' text was committed into the log
114 - add 'adaptive crap filter' logic to handle all
115 of the different corruptions of the cvs log file
116 - switch to rdiff instead of diff to not require working
117 files to be checked out. Note: when --diff-opts are
118 specified, diff must be used (not rdiff) because
119 rdiff doesn't support options(?!!)
122 - add the --diff-opts for explicitly setting the diff
123 options. important for setting increased context for
125 - use 'cvs rlog' instead of 'cvs log'. important if
126 there are empty directory pruned with cvs update -P
127 for getting consistent patchset numbering
128 - clean up patch_set_affects_branch for handling INVALID
129 tags and the -r especially around a branch point
130 - add the bkcvs proof-of-concept hack
134 - change the way INVALID tags are reported - show the exact
135 patchsets and revisions that conflict, and don't print
136 anything for 'funky' tags/revisions
137 - add psid to the patchset structure and pass through tree
138 once just to assign the numbers. this allows reporting
139 psid when problems are found (above)
140 - more code restructuring - move stats to own module
141 - commenting cleanups
142 - include the very rudimentary merge_utils.sh functions
143 I use to assist merging stuff
144 - change the way diffs are generated, handle individual
145 members that are before/after the -r tag restrictions
146 - change the semantics of the -s argument. it now only
147 restricts the patch sets the same as the other filtration
148 arguments. new argument -g turns on diff generation
149 - add manpage section discussing tag handling semantics
152 - major restructuring of source code. single cvsps.c split into
154 - major rework of the data structures in order to support the most
155 frequently requested, and most desired feature: -r. cvsps
156 can now associated symbolic tags with specific patchsets,
157 branches are now associated with the patchset instead of
158 the individual revision. you can view patchsets committed
159 after a given symbolic tag, or between two tags.
160 - changed the handling of same author, same log message commits
161 on different branches. this now creates multiple patchsets
162 - changed the handling of 'file xyz initially added on branch xyz'
163 log messages, they now create hidden patchsets.
164 - these last two items have made the patchset numbering scheme
165 incompatible with the old scheme. patchsets are going
168 1.99.1 (not released)
169 - redid the data structures a bit to 'normalize' the references to
170 a file revision. this allows the program to detect the case
171 when a file is added on a branch, and generate the right output
172 for 'cvsps -s' in this case (frequent bug report)
173 - added the '-l' option to restrict patchsets based on log descr
174 content. (patch from Geoff Soutter). added man page for it.
175 - added the --summary-first option suggested by Andi Kleen to
176 have the PatchSet summary information for all patchsets at
177 the head of the output, when multiple patchsets are given
179 - added the all: target to Makefile. (patch from Ben Elliston)
180 - use a single centralized cvsps.cache file per root/repository
181 this obviates the need to 'cvsps -u' in multiple checked out
182 trees of the same repository (patch adopted from Baruch Even)
183 - change the magic name TRUNK (for -b option) to HEAD to be
184 consistent with cvs. (patch from Henrik Nordstrom)
185 - when parsing 'cvs log' output, look for the exact strings
186 that separate the log entries and the files, instead of
187 just the first 8 characters. this allows a wider variety
188 of stuff to be in the log format (including actual patchsets).
189 Inspired by Andy Isaacson.
190 - added the -p option allowing patchset output to go into
191 individual files in a named directory. Based on a patch
192 by Henrik Nordstrom. This is a great idea. Thanks Henrik.
194 1.3.3 (why isn't this 1.4 again?)
195 - added the 'memory reduction' patch from Baruch Even. This patch
196 greatly reduces the memory footprint of CVSps against a large
197 repository by using dynamically allocated buffers, and by using
198 a tree to hold "common" strings (to avoid having a thousand copies
199 of the string "1.1"). This patch also adds a new option '-t' which
200 displays some memory usage statistics at the top.
202 - added the 00-strip-revision.patch from Steven Tweedie which fixes
203 a parsing problem when there are locked files in the repository
205 - added the 'Tweedie Tweenie' patch from Steven Tweedie (01-stable-tree)
206 which fixes a bug where interspersed commits could cause unstable
207 tree behavior. This would happen especially when multiple users commit
208 at the same time to different parts of a large tree over a slow link.
210 - added the --norc option (based on a suggestion by Soren S. Jorvang)
211 to handle cases where people have bad stuff in their .cvsrc which makes
212 cvsps fail. I could have added the '-f' unconditionally to the command
213 line for cvs when run under cvsps, but somehow that seemed dangerous.
215 1.3.2 (small feature release)
216 - added the 'multi-patchset' feature to the -s option, provided in its
217 entirety by Daiki Ueno <ueno@unixuser.org>
219 1.3.1 (bugfix release mostly)
220 - fix bug with updating cache (-u) having to do with matching new and old
223 - fix timestamp_fuzz_factor bug where the fuzz was applied to loading
224 from cache by mistake.
226 - add a spec file (from Jan IVEN <Jan.Iven@cern.ch>)
228 - improve the parameterization of Makefile, and make things relocatable
229 (from many people, esp. Amitai Schlair <schmonz@schmonz.com>)
231 - fix strip_path_len calculation (again - I had munged the prior fix)
232 Jeffrey Ebert <ebert@sonicsinc.com>
236 - fix case where CVSROOT is initial substring of Repository when
237 creating strip_path variable.
239 Thanks to Jean-Michel Rouet <jean-michel.rouet@philips.com> for
242 - fix case where a subdirectory has a different CVSROOT or Repository
243 path from the toplevel directory (happens when one project is checked
244 out inside another projects subdirectory).
246 Thanks to Philippe M. Chiasson <gozer@cpan.org> for the fix.
251 - Moved cache file location to CVS/ subdirectory. This is where CVS keeps
252 it's meta-data, and putting the cache there keeps it out of the way of
253 other CVS operations. (thanks to Amitai Schlair for suggestion)
255 - Added manual page. (thanks to Amitai Schlair for suggestion)
257 - Fixed the '-b' option to take a special branch name 'TRUNK' which
258 will restrict the output to patchsets made on the main trunk.
261 1.1 Initial public release.
262 -------- end of changelog ---------